CHANGES revision 295016
1 2 OpenSSL CHANGES 3 _______________ 4 5 Changes between 1.0.1q and 1.0.1r [28 Jan 2016] 6 7 *) Protection for DH small subgroup attacks 8 9 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been 10 switched on by default and cannot be disabled. This could have some 11 performance impact. 12 [Matt Caswell] 13 14 *) SSLv2 doesn't block disabled ciphers 15 16 A malicious client can negotiate SSLv2 ciphers that have been disabled on 17 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have 18 been disabled, provided that the SSLv2 protocol was not also disabled via 19 SSL_OP_NO_SSLv2. 20 21 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram 22 and Sebastian Schinzel. 23 (CVE-2015-3197) 24 [Viktor Dukhovni] 25 26 *) Reject DH handshakes with parameters shorter than 1024 bits. 27 [Kurt Roeckx] 28 29 Changes between 1.0.1p and 1.0.1q [3 Dec 2015] 30 31 *) Certificate verify crash with missing PSS parameter 32 33 The signature verification routines will crash with a NULL pointer 34 dereference if presented with an ASN.1 signature using the RSA PSS 35 algorithm and absent mask generation function parameter. Since these 36 routines are used to verify certificate signature algorithms this can be 37 used to crash any certificate verification operation and exploited in a 38 DoS attack. Any application which performs certificate verification is 39 vulnerable including OpenSSL clients and servers which enable client 40 authentication. 41 42 This issue was reported to OpenSSL by Lo��c Jonas Etienne (Qnective AG). 43 (CVE-2015-3194) 44 [Stephen Henson] 45 46 *) X509_ATTRIBUTE memory leak 47 48 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak 49 memory. This structure is used by the PKCS#7 and CMS routines so any 50 application which reads PKCS#7 or CMS data from untrusted sources is 51 affected. SSL/TLS is not affected. 52 53 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using 54 libFuzzer. 55 (CVE-2015-3195) 56 [Stephen Henson] 57 58 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs. 59 This changes the decoding behaviour for some invalid messages, 60 though the change is mostly in the more lenient direction, and 61 legacy behaviour is preserved as much as possible. 62 [Emilia K��sper] 63 64 *) In DSA_generate_parameters_ex, if the provided seed is too short, 65 return an error 66 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>] 67 68 Changes between 1.0.1o and 1.0.1p [9 Jul 2015] 69 70 *) Alternate chains certificate forgery 71 72 During certificate verfification, OpenSSL will attempt to find an 73 alternative certificate chain if the first attempt to build such a chain 74 fails. An error in the implementation of this logic can mean that an 75 attacker could cause certain checks on untrusted certificates to be 76 bypassed, such as the CA flag, enabling them to use a valid leaf 77 certificate to act as a CA and "issue" an invalid certificate. 78 79 This issue was reported to OpenSSL by Adam Langley/David Benjamin 80 (Google/BoringSSL). 81 (CVE-2015-1793) 82 [Matt Caswell] 83 84 *) Race condition handling PSK identify hint 85 86 If PSK identity hints are received by a multi-threaded client then 87 the values are wrongly updated in the parent SSL_CTX structure. This can 88 result in a race condition potentially leading to a double free of the 89 identify hint data. 90 (CVE-2015-3196) 91 [Stephen Henson] 92 93 Changes between 1.0.1n and 1.0.1o [12 Jun 2015] 94 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI 95 incompatibility in the handling of HMAC. The previous ABI has now been 96 restored. 97 98 Changes between 1.0.1m and 1.0.1n [11 Jun 2015] 99 100 *) Malformed ECParameters causes infinite loop 101 102 When processing an ECParameters structure OpenSSL enters an infinite loop 103 if the curve specified is over a specially malformed binary polynomial 104 field. 105 106 This can be used to perform denial of service against any 107 system which processes public keys, certificate requests or 108 certificates. This includes TLS clients and TLS servers with 109 client authentication enabled. 110 111 This issue was reported to OpenSSL by Joseph Barr-Pixton. 112 (CVE-2015-1788) 113 [Andy Polyakov] 114 115 *) Exploitable out-of-bounds read in X509_cmp_time 116 117 X509_cmp_time does not properly check the length of the ASN1_TIME 118 string and can read a few bytes out of bounds. In addition, 119 X509_cmp_time accepts an arbitrary number of fractional seconds in the 120 time string. 121 122 An attacker can use this to craft malformed certificates and CRLs of 123 various sizes and potentially cause a segmentation fault, resulting in 124 a DoS on applications that verify certificates or CRLs. TLS clients 125 that verify CRLs are affected. TLS clients and servers with client 126 authentication enabled may be affected if they use custom verification 127 callbacks. 128 129 This issue was reported to OpenSSL by Robert Swiecki (Google), and 130 independently by Hanno B��ck. 131 (CVE-2015-1789) 132 [Emilia K��sper] 133 134 *) PKCS7 crash with missing EnvelopedContent 135 136 The PKCS#7 parsing code does not handle missing inner EncryptedContent 137 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs 138 with missing content and trigger a NULL pointer dereference on parsing. 139 140 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7 141 structures from untrusted sources are affected. OpenSSL clients and 142 servers are not affected. 143 144 This issue was reported to OpenSSL by Michal Zalewski (Google). 145 (CVE-2015-1790) 146 [Emilia K��sper] 147 148 *) CMS verify infinite loop with unknown hash function 149 150 When verifying a signedData message the CMS code can enter an infinite loop 151 if presented with an unknown hash function OID. This can be used to perform 152 denial of service against any system which verifies signedData messages using 153 the CMS code. 154 This issue was reported to OpenSSL by Johannes Bauer. 155 (CVE-2015-1792) 156 [Stephen Henson] 157 158 *) Race condition handling NewSessionTicket 159 160 If a NewSessionTicket is received by a multi-threaded client when attempting to 161 reuse a previous ticket then a race condition can occur potentially leading to 162 a double free of the ticket data. 163 (CVE-2015-1791) 164 [Matt Caswell] 165 166 *) Reject DH handshakes with parameters shorter than 768 bits. 167 [Kurt Roeckx and Emilia Kasper] 168 169 *) dhparam: generate 2048-bit parameters by default. 170 [Kurt Roeckx and Emilia Kasper] 171 172 Changes between 1.0.1l and 1.0.1m [19 Mar 2015] 173 174 *) Segmentation fault in ASN1_TYPE_cmp fix 175 176 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is 177 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check 178 certificate signature algorithm consistency this can be used to crash any 179 certificate verification operation and exploited in a DoS attack. Any 180 application which performs certificate verification is vulnerable including 181 OpenSSL clients and servers which enable client authentication. 182 (CVE-2015-0286) 183 [Stephen Henson] 184 185 *) ASN.1 structure reuse memory corruption fix 186 187 Reusing a structure in ASN.1 parsing may allow an attacker to cause 188 memory corruption via an invalid write. Such reuse is and has been 189 strongly discouraged and is believed to be rare. 190 191 Applications that parse structures containing CHOICE or ANY DEFINED BY 192 components may be affected. Certificate parsing (d2i_X509 and related 193 functions) are however not affected. OpenSSL clients and servers are 194 not affected. 195 (CVE-2015-0287) 196 [Stephen Henson] 197 198 *) PKCS7 NULL pointer dereferences fix 199 200 The PKCS#7 parsing code does not handle missing outer ContentInfo 201 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with 202 missing content and trigger a NULL pointer dereference on parsing. 203 204 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or 205 otherwise parse PKCS#7 structures from untrusted sources are 206 affected. OpenSSL clients and servers are not affected. 207 208 This issue was reported to OpenSSL by Michal Zalewski (Google). 209 (CVE-2015-0289) 210 [Emilia K��sper] 211 212 *) DoS via reachable assert in SSLv2 servers fix 213 214 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in 215 servers that both support SSLv2 and enable export cipher suites by sending 216 a specially crafted SSLv2 CLIENT-MASTER-KEY message. 217 218 This issue was discovered by Sean Burford (Google) and Emilia K��sper 219 (OpenSSL development team). 220 (CVE-2015-0293) 221 [Emilia K��sper] 222 223 *) Use After Free following d2i_ECPrivatekey error fix 224 225 A malformed EC private key file consumed via the d2i_ECPrivateKey function 226 could cause a use after free condition. This, in turn, could cause a double 227 free in several private key parsing functions (such as d2i_PrivateKey 228 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption 229 for applications that receive EC private keys from untrusted 230 sources. This scenario is considered rare. 231 232 This issue was discovered by the BoringSSL project and fixed in their 233 commit 517073cd4b. 234 (CVE-2015-0209) 235 [Matt Caswell] 236 237 *) X509_to_X509_REQ NULL pointer deref fix 238 239 The function X509_to_X509_REQ will crash with a NULL pointer dereference if 240 the certificate key is invalid. This function is rarely used in practice. 241 242 This issue was discovered by Brian Carpenter. 243 (CVE-2015-0288) 244 [Stephen Henson] 245 246 *) Removed the export ciphers from the DEFAULT ciphers 247 [Kurt Roeckx] 248 249 Changes between 1.0.1k and 1.0.1l [15 Jan 2015] 250 251 *) Build fixes for the Windows and OpenVMS platforms 252 [Matt Caswell and Richard Levitte] 253 254 Changes between 1.0.1j and 1.0.1k [8 Jan 2015] 255 256 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS 257 message can cause a segmentation fault in OpenSSL due to a NULL pointer 258 dereference. This could lead to a Denial Of Service attack. Thanks to 259 Markus Stenberg of Cisco Systems, Inc. for reporting this issue. 260 (CVE-2014-3571) 261 [Steve Henson] 262 263 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the 264 dtls1_buffer_record function under certain conditions. In particular this 265 could occur if an attacker sent repeated DTLS records with the same 266 sequence number but for the next epoch. The memory leak could be exploited 267 by an attacker in a Denial of Service attack through memory exhaustion. 268 Thanks to Chris Mueller for reporting this issue. 269 (CVE-2015-0206) 270 [Matt Caswell] 271 272 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is 273 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl 274 method would be set to NULL which could later result in a NULL pointer 275 dereference. Thanks to Frank Schmirler for reporting this issue. 276 (CVE-2014-3569) 277 [Kurt Roeckx] 278 279 *) Abort handshake if server key exchange message is omitted for ephemeral 280 ECDH ciphersuites. 281 282 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for 283 reporting this issue. 284 (CVE-2014-3572) 285 [Steve Henson] 286 287 *) Remove non-export ephemeral RSA code on client and server. This code 288 violated the TLS standard by allowing the use of temporary RSA keys in 289 non-export ciphersuites and could be used by a server to effectively 290 downgrade the RSA key length used to a value smaller than the server 291 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at 292 INRIA or reporting this issue. 293 (CVE-2015-0204) 294 [Steve Henson] 295 296 *) Fixed issue where DH client certificates are accepted without verification. 297 An OpenSSL server will accept a DH certificate for client authentication 298 without the certificate verify message. This effectively allows a client to 299 authenticate without the use of a private key. This only affects servers 300 which trust a client certificate authority which issues certificates 301 containing DH keys: these are extremely rare and hardly ever encountered. 302 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting 303 this issue. 304 (CVE-2015-0205) 305 [Steve Henson] 306 307 *) Ensure that the session ID context of an SSL is updated when its 308 SSL_CTX is updated via SSL_set_SSL_CTX. 309 310 The session ID context is typically set from the parent SSL_CTX, 311 and can vary with the CTX. 312 [Adam Langley] 313 314 *) Fix various certificate fingerprint issues. 315 316 By using non-DER or invalid encodings outside the signed portion of a 317 certificate the fingerprint can be changed without breaking the signature. 318 Although no details of the signed portion of the certificate can be changed 319 this can cause problems with some applications: e.g. those using the 320 certificate fingerprint for blacklists. 321 322 1. Reject signatures with non zero unused bits. 323 324 If the BIT STRING containing the signature has non zero unused bits reject 325 the signature. All current signature algorithms require zero unused bits. 326 327 2. Check certificate algorithm consistency. 328 329 Check the AlgorithmIdentifier inside TBS matches the one in the 330 certificate signature. NB: this will result in signature failure 331 errors for some broken certificates. 332 333 Thanks to Konrad Kraszewski from Google for reporting this issue. 334 335 3. Check DSA/ECDSA signatures use DER. 336 337 Reencode DSA/ECDSA signatures and compare with the original received 338 signature. Return an error if there is a mismatch. 339 340 This will reject various cases including garbage after signature 341 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS 342 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs 343 (negative or with leading zeroes). 344 345 Further analysis was conducted and fixes were developed by Stephen Henson 346 of the OpenSSL core team. 347 348 (CVE-2014-8275) 349 [Steve Henson] 350 351 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect 352 results on some platforms, including x86_64. This bug occurs at random 353 with a very low probability, and is not known to be exploitable in any 354 way, though its exact impact is difficult to determine. Thanks to Pieter 355 Wuille (Blockstream) who reported this issue and also suggested an initial 356 fix. Further analysis was conducted by the OpenSSL development team and 357 Adam Langley of Google. The final fix was developed by Andy Polyakov of 358 the OpenSSL core team. 359 (CVE-2014-3570) 360 [Andy Polyakov] 361 362 *) Do not resume sessions on the server if the negotiated protocol 363 version does not match the session's version. Resuming with a different 364 version, while not strictly forbidden by the RFC, is of questionable 365 sanity and breaks all known clients. 366 [David Benjamin, Emilia K��sper] 367 368 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject 369 early CCS messages during renegotiation. (Note that because 370 renegotiation is encrypted, this early CCS was not exploitable.) 371 [Emilia K��sper] 372 373 *) Tighten client-side session ticket handling during renegotiation: 374 ensure that the client only accepts a session ticket if the server sends 375 the extension anew in the ServerHello. Previously, a TLS client would 376 reuse the old extension state and thus accept a session ticket if one was 377 announced in the initial ServerHello. 378 379 Similarly, ensure that the client requires a session ticket if one 380 was advertised in the ServerHello. Previously, a TLS client would 381 ignore a missing NewSessionTicket message. 382 [Emilia K��sper] 383 384 Changes between 1.0.1i and 1.0.1j [15 Oct 2014] 385 386 *) SRTP Memory Leak. 387 388 A flaw in the DTLS SRTP extension parsing code allows an attacker, who 389 sends a carefully crafted handshake message, to cause OpenSSL to fail 390 to free up to 64k of memory causing a memory leak. This could be 391 exploited in a Denial Of Service attack. This issue affects OpenSSL 392 1.0.1 server implementations for both SSL/TLS and DTLS regardless of 393 whether SRTP is used or configured. Implementations of OpenSSL that 394 have been compiled with OPENSSL_NO_SRTP defined are not affected. 395 396 The fix was developed by the OpenSSL team. 397 (CVE-2014-3513) 398 [OpenSSL team] 399 400 *) Session Ticket Memory Leak. 401 402 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the 403 integrity of that ticket is first verified. In the event of a session 404 ticket integrity check failing, OpenSSL will fail to free memory 405 causing a memory leak. By sending a large number of invalid session 406 tickets an attacker could exploit this issue in a Denial Of Service 407 attack. 408 (CVE-2014-3567) 409 [Steve Henson] 410 411 *) Build option no-ssl3 is incomplete. 412 413 When OpenSSL is configured with "no-ssl3" as a build option, servers 414 could accept and complete a SSL 3.0 handshake, and clients could be 415 configured to send them. 416 (CVE-2014-3568) 417 [Akamai and the OpenSSL team] 418 419 *) Add support for TLS_FALLBACK_SCSV. 420 Client applications doing fallback retries should call 421 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV). 422 (CVE-2014-3566) 423 [Adam Langley, Bodo Moeller] 424 425 *) Add additional DigestInfo checks. 426 427 Reencode DigestInto in DER and check against the original when 428 verifying RSA signature: this will reject any improperly encoded 429 DigestInfo structures. 430 431 Note: this is a precautionary measure and no attacks are currently known. 432 433 [Steve Henson] 434 435 Changes between 1.0.1h and 1.0.1i [6 Aug 2014] 436 437 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the 438 SRP code can be overrun an internal buffer. Add sanity check that 439 g, A, B < N to SRP code. 440 441 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC 442 Group for discovering this issue. 443 (CVE-2014-3512) 444 [Steve Henson] 445 446 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate 447 TLS 1.0 instead of higher protocol versions when the ClientHello message 448 is badly fragmented. This allows a man-in-the-middle attacker to force a 449 downgrade to TLS 1.0 even if both the server and the client support a 450 higher protocol version, by modifying the client's TLS records. 451 452 Thanks to David Benjamin and Adam Langley (Google) for discovering and 453 researching this issue. 454 (CVE-2014-3511) 455 [David Benjamin] 456 457 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject 458 to a denial of service attack. A malicious server can crash the client 459 with a null pointer dereference (read) by specifying an anonymous (EC)DH 460 ciphersuite and sending carefully crafted handshake messages. 461 462 Thanks to Felix Gr��bert (Google) for discovering and researching this 463 issue. 464 (CVE-2014-3510) 465 [Emilia K��sper] 466 467 *) By sending carefully crafted DTLS packets an attacker could cause openssl 468 to leak memory. This can be exploited through a Denial of Service attack. 469 Thanks to Adam Langley for discovering and researching this issue. 470 (CVE-2014-3507) 471 [Adam Langley] 472 473 *) An attacker can force openssl to consume large amounts of memory whilst 474 processing DTLS handshake messages. This can be exploited through a 475 Denial of Service attack. 476 Thanks to Adam Langley for discovering and researching this issue. 477 (CVE-2014-3506) 478 [Adam Langley] 479 480 *) An attacker can force an error condition which causes openssl to crash 481 whilst processing DTLS packets due to memory being freed twice. This 482 can be exploited through a Denial of Service attack. 483 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching 484 this issue. 485 (CVE-2014-3505) 486 [Adam Langley] 487 488 *) If a multithreaded client connects to a malicious server using a resumed 489 session and the server sends an ec point format extension it could write 490 up to 255 bytes to freed memory. 491 492 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this 493 issue. 494 (CVE-2014-3509) 495 [Gabor Tyukasz] 496 497 *) A malicious server can crash an OpenSSL client with a null pointer 498 dereference (read) by specifying an SRP ciphersuite even though it was not 499 properly negotiated with the client. This can be exploited through a 500 Denial of Service attack. 501 502 Thanks to Joonas Kuorilehto and Riku Hietam��ki (Codenomicon) for 503 discovering and researching this issue. 504 (CVE-2014-5139) 505 [Steve Henson] 506 507 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as 508 X509_name_oneline, X509_name_print_ex et al. to leak some information 509 from the stack. Applications may be affected if they echo pretty printing 510 output to the attacker. 511 512 Thanks to Ivan Fratric (Google) for discovering this issue. 513 (CVE-2014-3508) 514 [Emilia K��sper, and Steve Henson] 515 516 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) 517 for corner cases. (Certain input points at infinity could lead to 518 bogus results, with non-infinity inputs mapped to infinity too.) 519 [Bodo Moeller] 520 521 Changes between 1.0.1g and 1.0.1h [5 Jun 2014] 522 523 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted 524 handshake can force the use of weak keying material in OpenSSL 525 SSL/TLS clients and servers. 526 527 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and 528 researching this issue. (CVE-2014-0224) 529 [KIKUCHI Masashi, Steve Henson] 530 531 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an 532 OpenSSL DTLS client the code can be made to recurse eventually crashing 533 in a DoS attack. 534 535 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue. 536 (CVE-2014-0221) 537 [Imre Rad, Steve Henson] 538 539 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can 540 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS 541 client or server. This is potentially exploitable to run arbitrary 542 code on a vulnerable client or server. 543 544 Thanks to J��ri Aedla for reporting this issue. (CVE-2014-0195) 545 [J��ri Aedla, Steve Henson] 546 547 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites 548 are subject to a denial of service attack. 549 550 Thanks to Felix Gr��bert and Ivan Fratric at Google for discovering 551 this issue. (CVE-2014-3470) 552 [Felix Gr��bert, Ivan Fratric, Steve Henson] 553 554 *) Harmonize version and its documentation. -f flag is used to display 555 compilation flags. 556 [mancha <mancha1@zoho.com>] 557 558 *) Fix eckey_priv_encode so it immediately returns an error upon a failure 559 in i2d_ECPrivateKey. 560 [mancha <mancha1@zoho.com>] 561 562 *) Fix some double frees. These are not thought to be exploitable. 563 [mancha <mancha1@zoho.com>] 564 565 Changes between 1.0.1f and 1.0.1g [7 Apr 2014] 566 567 *) A missing bounds check in the handling of the TLS heartbeat extension 568 can be used to reveal up to 64k of memory to a connected client or 569 server. 570 571 Thanks for Neel Mehta of Google Security for discovering this bug and to 572 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for 573 preparing the fix (CVE-2014-0160) 574 [Adam Langley, Bodo Moeller] 575 576 *) Fix for the attack described in the paper "Recovering OpenSSL 577 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack" 578 by Yuval Yarom and Naomi Benger. Details can be obtained from: 579 http://eprint.iacr.org/2014/140 580 581 Thanks to Yuval Yarom and Naomi Benger for discovering this 582 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076) 583 [Yuval Yarom and Naomi Benger] 584 585 *) TLS pad extension: draft-agl-tls-padding-03 586 587 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the 588 TLS client Hello record length value would otherwise be > 255 and 589 less that 512 pad with a dummy extension containing zeroes so it 590 is at least 512 bytes long. 591 592 [Adam Langley, Steve Henson] 593 594 Changes between 1.0.1e and 1.0.1f [6 Jan 2014] 595 596 *) Fix for TLS record tampering bug. A carefully crafted invalid 597 handshake could crash OpenSSL with a NULL pointer exception. 598 Thanks to Anton Johansson for reporting this issues. 599 (CVE-2013-4353) 600 601 *) Keep original DTLS digest and encryption contexts in retransmission 602 structures so we can use the previous session parameters if they need 603 to be resent. (CVE-2013-6450) 604 [Steve Henson] 605 606 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which 607 avoids preferring ECDHE-ECDSA ciphers when the client appears to be 608 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for 609 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug 610 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing 611 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer. 612 [Rob Stradling, Adam Langley] 613 614 Changes between 1.0.1d and 1.0.1e [11 Feb 2013] 615 616 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI 617 supporting platforms or when small records were transferred. 618 [Andy Polyakov, Steve Henson] 619 620 Changes between 1.0.1c and 1.0.1d [5 Feb 2013] 621 622 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 623 624 This addresses the flaw in CBC record processing discovered by 625 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 626 at: http://www.isg.rhul.ac.uk/tls/ 627 628 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 629 Security Group at Royal Holloway, University of London 630 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 631 Emilia K��sper for the initial patch. 632 (CVE-2013-0169) 633 [Emilia K��sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 634 635 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode 636 ciphersuites which can be exploited in a denial of service attack. 637 Thanks go to and to Adam Langley <agl@chromium.org> for discovering 638 and detecting this bug and to Wolfgang Ettlinger 639 <wolfgang.ettlinger@gmail.com> for independently discovering this issue. 640 (CVE-2012-2686) 641 [Adam Langley] 642 643 *) Return an error when checking OCSP signatures when key is NULL. 644 This fixes a DoS attack. (CVE-2013-0166) 645 [Steve Henson] 646 647 *) Make openssl verify return errors. 648 [Chris Palmer <palmer@google.com> and Ben Laurie] 649 650 *) Call OCSP Stapling callback after ciphersuite has been chosen, so 651 the right response is stapled. Also change SSL_get_certificate() 652 so it returns the certificate actually sent. 653 See http://rt.openssl.org/Ticket/Display.html?id=2836. 654 [Rob Stradling <rob.stradling@comodo.com>] 655 656 *) Fix possible deadlock when decoding public keys. 657 [Steve Henson] 658 659 *) Don't use TLS 1.0 record version number in initial client hello 660 if renegotiating. 661 [Steve Henson] 662 663 Changes between 1.0.1b and 1.0.1c [10 May 2012] 664 665 *) Sanity check record length before skipping explicit IV in TLS 666 1.2, 1.1 and DTLS to fix DoS attack. 667 668 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 669 fuzzing as a service testing platform. 670 (CVE-2012-2333) 671 [Steve Henson] 672 673 *) Initialise tkeylen properly when encrypting CMS messages. 674 Thanks to Solar Designer of Openwall for reporting this issue. 675 [Steve Henson] 676 677 *) In FIPS mode don't try to use composite ciphers as they are not 678 approved. 679 [Steve Henson] 680 681 Changes between 1.0.1a and 1.0.1b [26 Apr 2012] 682 683 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and 684 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately 685 mean any application compiled against OpenSSL 1.0.0 headers setting 686 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng 687 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to 688 0x10000000L Any application which was previously compiled against 689 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1 690 will need to be recompiled as a result. Letting be results in 691 inability to disable specifically TLS 1.1 and in client context, 692 in unlike event, limit maximum offered version to TLS 1.0 [see below]. 693 [Steve Henson] 694 695 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not 696 disable just protocol X, but all protocols above X *if* there are 697 protocols *below* X still enabled. In more practical terms it means 698 that if application wants to disable TLS1.0 in favor of TLS1.1 and 699 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass 700 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to 701 client side. 702 [Andy Polyakov] 703 704 Changes between 1.0.1 and 1.0.1a [19 Apr 2012] 705 706 *) Check for potentially exploitable overflows in asn1_d2i_read_bio 707 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 708 in CRYPTO_realloc_clean. 709 710 Thanks to Tavis Ormandy, Google Security Team, for discovering this 711 issue and to Adam Langley <agl@chromium.org> for fixing it. 712 (CVE-2012-2110) 713 [Adam Langley (Google), Tavis Ormandy, Google Security Team] 714 715 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections. 716 [Adam Langley] 717 718 *) Workarounds for some broken servers that "hang" if a client hello 719 record length exceeds 255 bytes. 720 721 1. Do not use record version number > TLS 1.0 in initial client 722 hello: some (but not all) hanging servers will now work. 723 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate 724 the number of ciphers sent in the client hello. This should be 725 set to an even number, such as 50, for example by passing: 726 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure. 727 Most broken servers should now work. 728 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable 729 TLS 1.2 client support entirely. 730 [Steve Henson] 731 732 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH. 733 [Andy Polyakov] 734 735 Changes between 1.0.0h and 1.0.1 [14 Mar 2012] 736 737 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET 738 STRING form instead of a DigestInfo. 739 [Steve Henson] 740 741 *) The format used for MDC2 RSA signatures is inconsistent between EVP 742 and the RSA_sign/RSA_verify functions. This was made more apparent when 743 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular 744 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect 745 the correct format in RSA_verify so both forms transparently work. 746 [Steve Henson] 747 748 *) Some servers which support TLS 1.0 can choke if we initially indicate 749 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA 750 encrypted premaster secret. As a workaround use the maximum pemitted 751 client version in client hello, this should keep such servers happy 752 and still work with previous versions of OpenSSL. 753 [Steve Henson] 754 755 *) Add support for TLS/DTLS heartbeats. 756 [Robin Seggelmann <seggelmann@fh-muenster.de>] 757 758 *) Add support for SCTP. 759 [Robin Seggelmann <seggelmann@fh-muenster.de>] 760 761 *) Improved PRNG seeding for VOS. 762 [Paul Green <Paul.Green@stratus.com>] 763 764 *) Extensive assembler packs updates, most notably: 765 766 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support; 767 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES); 768 - x86_64: bit-sliced AES implementation; 769 - ARM: NEON support, contemporary platforms optimizations; 770 - s390x: z196 support; 771 - *: GHASH and GF(2^m) multiplication implementations; 772 773 [Andy Polyakov] 774 775 *) Make TLS-SRP code conformant with RFC 5054 API cleanup 776 (removal of unnecessary code) 777 [Peter Sylvester <peter.sylvester@edelweb.fr>] 778 779 *) Add TLS key material exporter from RFC 5705. 780 [Eric Rescorla] 781 782 *) Add DTLS-SRTP negotiation from RFC 5764. 783 [Eric Rescorla] 784 785 *) Add Next Protocol Negotiation, 786 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be 787 disabled with a no-npn flag to config or Configure. Code donated 788 by Google. 789 [Adam Langley <agl@google.com> and Ben Laurie] 790 791 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224, 792 NIST-P256, NIST-P521, with constant-time single point multiplication on 793 typical inputs. Compiler support for the nonstandard type __uint128_t is 794 required to use this (present in gcc 4.4 and later, for 64-bit builds). 795 Code made available under Apache License version 2.0. 796 797 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command 798 line to include this in your build of OpenSSL, and run "make depend" (or 799 "make update"). This enables the following EC_METHODs: 800 801 EC_GFp_nistp224_method() 802 EC_GFp_nistp256_method() 803 EC_GFp_nistp521_method() 804 805 EC_GROUP_new_by_curve_name() will automatically use these (while 806 EC_GROUP_new_curve_GFp() currently prefers the more flexible 807 implementations). 808 [Emilia K��sper, Adam Langley, Bodo Moeller (Google)] 809 810 *) Use type ossl_ssize_t instad of ssize_t which isn't available on 811 all platforms. Move ssize_t definition from e_os.h to the public 812 header file e_os2.h as it now appears in public header file cms.h 813 [Steve Henson] 814 815 *) New -sigopt option to the ca, req and x509 utilities. Additional 816 signature parameters can be passed using this option and in 817 particular PSS. 818 [Steve Henson] 819 820 *) Add RSA PSS signing function. This will generate and set the 821 appropriate AlgorithmIdentifiers for PSS based on those in the 822 corresponding EVP_MD_CTX structure. No application support yet. 823 [Steve Henson] 824 825 *) Support for companion algorithm specific ASN1 signing routines. 826 New function ASN1_item_sign_ctx() signs a pre-initialised 827 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on 828 the appropriate parameters. 829 [Steve Henson] 830 831 *) Add new algorithm specific ASN1 verification initialisation function 832 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1 833 handling will be the same no matter what EVP_PKEY_METHOD is used. 834 Add a PSS handler to support verification of PSS signatures: checked 835 against a number of sample certificates. 836 [Steve Henson] 837 838 *) Add signature printing for PSS. Add PSS OIDs. 839 [Steve Henson, Martin Kaiser <lists@kaiser.cx>] 840 841 *) Add algorithm specific signature printing. An individual ASN1 method 842 can now print out signatures instead of the standard hex dump. 843 844 More complex signatures (e.g. PSS) can print out more meaningful 845 information. Include DSA version that prints out the signature 846 parameters r, s. 847 [Steve Henson] 848 849 *) Password based recipient info support for CMS library: implementing 850 RFC3211. 851 [Steve Henson] 852 853 *) Split password based encryption into PBES2 and PBKDF2 functions. This 854 neatly separates the code into cipher and PBE sections and is required 855 for some algorithms that split PBES2 into separate pieces (such as 856 password based CMS). 857 [Steve Henson] 858 859 *) Session-handling fixes: 860 - Fix handling of connections that are resuming with a session ID, 861 but also support Session Tickets. 862 - Fix a bug that suppressed issuing of a new ticket if the client 863 presented a ticket with an expired session. 864 - Try to set the ticket lifetime hint to something reasonable. 865 - Make tickets shorter by excluding irrelevant information. 866 - On the client side, don't ignore renewed tickets. 867 [Adam Langley, Bodo Moeller (Google)] 868 869 *) Fix PSK session representation. 870 [Bodo Moeller] 871 872 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations. 873 874 This work was sponsored by Intel. 875 [Andy Polyakov] 876 877 *) Add GCM support to TLS library. Some custom code is needed to split 878 the IV between the fixed (from PRF) and explicit (from TLS record) 879 portions. This adds all GCM ciphersuites supported by RFC5288 and 880 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and 881 add a special AESGCM string for GCM only. 882 [Steve Henson] 883 884 *) Expand range of ctrls for AES GCM. Permit setting invocation 885 field on decrypt and retrieval of invocation field only on encrypt. 886 [Steve Henson] 887 888 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support. 889 As required by RFC5289 these ciphersuites cannot be used if for 890 versions of TLS earlier than 1.2. 891 [Steve Henson] 892 893 *) For FIPS capable OpenSSL interpret a NULL default public key method 894 as unset and return the appopriate default but do *not* set the default. 895 This means we can return the appopriate method in applications that 896 swicth between FIPS and non-FIPS modes. 897 [Steve Henson] 898 899 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an 900 ENGINE is used then we cannot handle that in the FIPS module so we 901 keep original code iff non-FIPS operations are allowed. 902 [Steve Henson] 903 904 *) Add -attime option to openssl utilities. 905 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson] 906 907 *) Redirect DSA and DH operations to FIPS module in FIPS mode. 908 [Steve Henson] 909 910 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use 911 FIPS EC methods unconditionally for now. 912 [Steve Henson] 913 914 *) New build option no-ec2m to disable characteristic 2 code. 915 [Steve Henson] 916 917 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not 918 all cases can be covered as some introduce binary incompatibilities. 919 [Steve Henson] 920 921 *) Redirect RSA operations to FIPS module including keygen, 922 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods. 923 [Steve Henson] 924 925 *) Add similar low level API blocking to ciphers. 926 [Steve Henson] 927 928 *) Low level digest APIs are not approved in FIPS mode: any attempt 929 to use these will cause a fatal error. Applications that *really* want 930 to use them can use the private_* version instead. 931 [Steve Henson] 932 933 *) Redirect cipher operations to FIPS module for FIPS builds. 934 [Steve Henson] 935 936 *) Redirect digest operations to FIPS module for FIPS builds. 937 [Steve Henson] 938 939 *) Update build system to add "fips" flag which will link in fipscanister.o 940 for static and shared library builds embedding a signature if needed. 941 [Steve Henson] 942 943 *) Output TLS supported curves in preference order instead of numerical 944 order. This is currently hardcoded for the highest order curves first. 945 This should be configurable so applications can judge speed vs strength. 946 [Steve Henson] 947 948 *) Add TLS v1.2 server support for client authentication. 949 [Steve Henson] 950 951 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers 952 and enable MD5. 953 [Steve Henson] 954 955 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying 956 FIPS modules versions. 957 [Steve Henson] 958 959 *) Add TLS v1.2 client side support for client authentication. Keep cache 960 of handshake records longer as we don't know the hash algorithm to use 961 until after the certificate request message is received. 962 [Steve Henson] 963 964 *) Initial TLS v1.2 client support. Add a default signature algorithms 965 extension including all the algorithms we support. Parse new signature 966 format in client key exchange. Relax some ECC signing restrictions for 967 TLS v1.2 as indicated in RFC5246. 968 [Steve Henson] 969 970 *) Add server support for TLS v1.2 signature algorithms extension. Switch 971 to new signature format when needed using client digest preference. 972 All server ciphersuites should now work correctly in TLS v1.2. No client 973 support yet and no support for client certificates. 974 [Steve Henson] 975 976 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch 977 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based 978 ciphersuites. At present only RSA key exchange ciphersuites work with 979 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete 980 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods 981 and version checking. 982 [Steve Henson] 983 984 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled 985 with this defined it will not be affected by any changes to ssl internal 986 structures. Add several utility functions to allow openssl application 987 to work with OPENSSL_NO_SSL_INTERN defined. 988 [Steve Henson] 989 990 *) Add SRP support. 991 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie] 992 993 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id. 994 [Steve Henson] 995 996 *) Permit abbreviated handshakes when renegotiating using the function 997 SSL_renegotiate_abbreviated(). 998 [Robin Seggelmann <seggelmann@fh-muenster.de>] 999 1000 *) Add call to ENGINE_register_all_complete() to 1001 ENGINE_load_builtin_engines(), so some implementations get used 1002 automatically instead of needing explicit application support. 1003 [Steve Henson] 1004 1005 *) Add support for TLS key exporter as described in RFC5705. 1006 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson] 1007 1008 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only 1009 a few changes are required: 1010 1011 Add SSL_OP_NO_TLSv1_1 flag. 1012 Add TLSv1_1 methods. 1013 Update version checking logic to handle version 1.1. 1014 Add explicit IV handling (ported from DTLS code). 1015 Add command line options to s_client/s_server. 1016 [Steve Henson] 1017 1018 Changes between 1.0.0g and 1.0.0h [12 Mar 2012] 1019 1020 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 1021 in CMS and PKCS7 code. When RSA decryption fails use a random key for 1022 content decryption and always return the same error. Note: this attack 1023 needs on average 2^20 messages so it only affects automated senders. The 1024 old behaviour can be reenabled in the CMS code by setting the 1025 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 1026 an MMA defence is not necessary. 1027 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 1028 this issue. (CVE-2012-0884) 1029 [Steve Henson] 1030 1031 *) Fix CVE-2011-4619: make sure we really are receiving a 1032 client hello before rejecting multiple SGC restarts. Thanks to 1033 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 1034 [Steve Henson] 1035 1036 Changes between 1.0.0f and 1.0.0g [18 Jan 2012] 1037 1038 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 1039 Thanks to Antonio Martin, Enterprise Secure Access Research and 1040 Development, Cisco Systems, Inc. for discovering this bug and 1041 preparing a fix. (CVE-2012-0050) 1042 [Antonio Martin] 1043 1044 Changes between 1.0.0e and 1.0.0f [4 Jan 2012] 1045 1046 *) Nadhem Alfardan and Kenny Paterson have discovered an extension 1047 of the Vaudenay padding oracle attack on CBC mode encryption 1048 which enables an efficient plaintext recovery attack against 1049 the OpenSSL implementation of DTLS. Their attack exploits timing 1050 differences arising during decryption processing. A research 1051 paper describing this attack can be found at: 1052 http://www.isg.rhul.ac.uk/~kp/dtls.pdf 1053 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 1054 Security Group at Royal Holloway, University of London 1055 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 1056 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 1057 for preparing the fix. (CVE-2011-4108) 1058 [Robin Seggelmann, Michael Tuexen] 1059 1060 *) Clear bytes used for block padding of SSL 3.0 records. 1061 (CVE-2011-4576) 1062 [Adam Langley (Google)] 1063 1064 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 1065 Kadianakis <desnacked@gmail.com> for discovering this issue and 1066 Adam Langley for preparing the fix. (CVE-2011-4619) 1067 [Adam Langley (Google)] 1068 1069 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027) 1070 [Andrey Kulikov <amdeich@gmail.com>] 1071 1072 *) Prevent malformed RFC3779 data triggering an assertion failure. 1073 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 1074 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 1075 [Rob Austein <sra@hactrn.net>] 1076 1077 *) Improved PRNG seeding for VOS. 1078 [Paul Green <Paul.Green@stratus.com>] 1079 1080 *) Fix ssl_ciph.c set-up race. 1081 [Adam Langley (Google)] 1082 1083 *) Fix spurious failures in ecdsatest.c. 1084 [Emilia K��sper (Google)] 1085 1086 *) Fix the BIO_f_buffer() implementation (which was mixing different 1087 interpretations of the '..._len' fields). 1088 [Adam Langley (Google)] 1089 1090 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 1091 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 1092 threads won't reuse the same blinding coefficients. 1093 1094 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 1095 lock to call BN_BLINDING_invert_ex, and avoids one use of 1096 BN_BLINDING_update for each BN_BLINDING structure (previously, 1097 the last update always remained unused). 1098 [Emilia K��sper (Google)] 1099 1100 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf. 1101 [Bob Buckholz (Google)] 1102 1103 Changes between 1.0.0d and 1.0.0e [6 Sep 2011] 1104 1105 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted 1106 by initialising X509_STORE_CTX properly. (CVE-2011-3207) 1107 [Kaspar Brand <ossl@velox.ch>] 1108 1109 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 1110 for multi-threaded use of ECDH. (CVE-2011-3210) 1111 [Adam Langley (Google)] 1112 1113 *) Fix x509_name_ex_d2i memory leak on bad inputs. 1114 [Bodo Moeller] 1115 1116 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check 1117 signature public key algorithm by using OID xref utilities instead. 1118 Before this you could only use some ECC ciphersuites with SHA1 only. 1119 [Steve Henson] 1120 1121 *) Add protection against ECDSA timing attacks as mentioned in the paper 1122 by Billy Bob Brumley and Nicola Tuveri, see: 1123 1124 http://eprint.iacr.org/2011/232.pdf 1125 1126 [Billy Bob Brumley and Nicola Tuveri] 1127 1128 Changes between 1.0.0c and 1.0.0d [8 Feb 2011] 1129 1130 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 1131 [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 1132 1133 *) Fix bug in string printing code: if *any* escaping is enabled we must 1134 escape the escape character (backslash) or the resulting string is 1135 ambiguous. 1136 [Steve Henson] 1137 1138 Changes between 1.0.0b and 1.0.0c [2 Dec 2010] 1139 1140 *) Disable code workaround for ancient and obsolete Netscape browsers 1141 and servers: an attacker can use it in a ciphersuite downgrade attack. 1142 Thanks to Martin Rex for discovering this bug. CVE-2010-4180 1143 [Steve Henson] 1144 1145 *) Fixed J-PAKE implementation error, originally discovered by 1146 Sebastien Martini, further info and confirmation from Stefan 1147 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 1148 [Ben Laurie] 1149 1150 Changes between 1.0.0a and 1.0.0b [16 Nov 2010] 1151 1152 *) Fix extension code to avoid race conditions which can result in a buffer 1153 overrun vulnerability: resumed sessions must not be modified as they can 1154 be shared by multiple threads. CVE-2010-3864 1155 [Steve Henson] 1156 1157 *) Fix WIN32 build system to correctly link an ENGINE directory into 1158 a DLL. 1159 [Steve Henson] 1160 1161 Changes between 1.0.0 and 1.0.0a [01 Jun 2010] 1162 1163 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover 1164 (CVE-2010-1633) 1165 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>] 1166 1167 Changes between 0.9.8n and 1.0.0 [29 Mar 2010] 1168 1169 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher 1170 context. The operation can be customised via the ctrl mechanism in 1171 case ENGINEs want to include additional functionality. 1172 [Steve Henson] 1173 1174 *) Tolerate yet another broken PKCS#8 key format: private key value negative. 1175 [Steve Henson] 1176 1177 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to 1178 output hashes compatible with older versions of OpenSSL. 1179 [Willy Weisz <weisz@vcpc.univie.ac.at>] 1180 1181 *) Fix compression algorithm handling: if resuming a session use the 1182 compression algorithm of the resumed session instead of determining 1183 it from client hello again. Don't allow server to change algorithm. 1184 [Steve Henson] 1185 1186 *) Add load_crls() function to apps tidying load_certs() too. Add option 1187 to verify utility to allow additional CRLs to be included. 1188 [Steve Henson] 1189 1190 *) Update OCSP request code to permit adding custom headers to the request: 1191 some responders need this. 1192 [Steve Henson] 1193 1194 *) The function EVP_PKEY_sign() returns <=0 on error: check return code 1195 correctly. 1196 [Julia Lawall <julia@diku.dk>] 1197 1198 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it 1199 needlessly dereferenced structures, used obsolete functions and 1200 didn't handle all updated verify codes correctly. 1201 [Steve Henson] 1202 1203 *) Disable MD2 in the default configuration. 1204 [Steve Henson] 1205 1206 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to 1207 indicate the initial BIO being pushed or popped. This makes it possible 1208 to determine whether the BIO is the one explicitly called or as a result 1209 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so 1210 it handles reference counts correctly and doesn't zero out the I/O bio 1211 when it is not being explicitly popped. WARNING: applications which 1212 included workarounds for the old buggy behaviour will need to be modified 1213 or they could free up already freed BIOs. 1214 [Steve Henson] 1215 1216 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni 1217 renaming to all platforms (within the 0.9.8 branch, this was 1218 done conditionally on Netware platforms to avoid a name clash). 1219 [Guenter <lists@gknw.net>] 1220 1221 *) Add ECDHE and PSK support to DTLS. 1222 [Michael Tuexen <tuexen@fh-muenster.de>] 1223 1224 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't 1225 be used on C++. 1226 [Steve Henson] 1227 1228 *) Add "missing" function EVP_MD_flags() (without this the only way to 1229 retrieve a digest flags is by accessing the structure directly. Update 1230 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest 1231 or cipher is registered as in the "from" argument. Print out all 1232 registered digests in the dgst usage message instead of manually 1233 attempting to work them out. 1234 [Steve Henson] 1235 1236 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello: 1237 this allows the use of compression and extensions. Change default cipher 1238 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2 1239 by default unless an application cipher string requests it. 1240 [Steve Henson] 1241 1242 *) Alter match criteria in PKCS12_parse(). It used to try to use local 1243 key ids to find matching certificates and keys but some PKCS#12 files 1244 don't follow the (somewhat unwritten) rules and this strategy fails. 1245 Now just gather all certificates together and the first private key 1246 then look for the first certificate that matches the key. 1247 [Steve Henson] 1248 1249 *) Support use of registered digest and cipher names for dgst and cipher 1250 commands instead of having to add each one as a special case. So now 1251 you can do: 1252 1253 openssl sha256 foo 1254 1255 as well as: 1256 1257 openssl dgst -sha256 foo 1258 1259 and this works for ENGINE based algorithms too. 1260 1261 [Steve Henson] 1262 1263 *) Update Gost ENGINE to support parameter files. 1264 [Victor B. Wagner <vitus@cryptocom.ru>] 1265 1266 *) Support GeneralizedTime in ca utility. 1267 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson] 1268 1269 *) Enhance the hash format used for certificate directory links. The new 1270 form uses the canonical encoding (meaning equivalent names will work 1271 even if they aren't identical) and uses SHA1 instead of MD5. This form 1272 is incompatible with the older format and as a result c_rehash should 1273 be used to rebuild symbolic links. 1274 [Steve Henson] 1275 1276 *) Make PKCS#8 the default write format for private keys, replacing the 1277 traditional format. This form is standardised, more secure and doesn't 1278 include an implicit MD5 dependency. 1279 [Steve Henson] 1280 1281 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code 1282 committed to OpenSSL should pass this lot as a minimum. 1283 [Steve Henson] 1284 1285 *) Add session ticket override functionality for use by EAP-FAST. 1286 [Jouni Malinen <j@w1.fi>] 1287 1288 *) Modify HMAC functions to return a value. Since these can be implemented 1289 in an ENGINE errors can occur. 1290 [Steve Henson] 1291 1292 *) Type-checked OBJ_bsearch_ex. 1293 [Ben Laurie] 1294 1295 *) Type-checked OBJ_bsearch. Also some constification necessitated 1296 by type-checking. Still to come: TXT_DB, bsearch(?), 1297 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING, 1298 CONF_VALUE. 1299 [Ben Laurie] 1300 1301 *) New function OPENSSL_gmtime_adj() to add a specific number of days and 1302 seconds to a tm structure directly, instead of going through OS 1303 specific date routines. This avoids any issues with OS routines such 1304 as the year 2038 bug. New *_adj() functions for ASN1 time structures 1305 and X509_time_adj_ex() to cover the extended range. The existing 1306 X509_time_adj() is still usable and will no longer have any date issues. 1307 [Steve Henson] 1308 1309 *) Delta CRL support. New use deltas option which will attempt to locate 1310 and search any appropriate delta CRLs available. 1311 1312 This work was sponsored by Google. 1313 [Steve Henson] 1314 1315 *) Support for CRLs partitioned by reason code. Reorganise CRL processing 1316 code and add additional score elements. Validate alternate CRL paths 1317 as part of the CRL checking and indicate a new error "CRL path validation 1318 error" in this case. Applications wanting additional details can use 1319 the verify callback and check the new "parent" field. If this is not 1320 NULL CRL path validation is taking place. Existing applications wont 1321 see this because it requires extended CRL support which is off by 1322 default. 1323 1324 This work was sponsored by Google. 1325 [Steve Henson] 1326 1327 *) Support for freshest CRL extension. 1328 1329 This work was sponsored by Google. 1330 [Steve Henson] 1331 1332 *) Initial indirect CRL support. Currently only supported in the CRLs 1333 passed directly and not via lookup. Process certificate issuer 1334 CRL entry extension and lookup CRL entries by bother issuer name 1335 and serial number. Check and process CRL issuer entry in IDP extension. 1336 1337 This work was sponsored by Google. 1338 [Steve Henson] 1339 1340 *) Add support for distinct certificate and CRL paths. The CRL issuer 1341 certificate is validated separately in this case. Only enabled if 1342 an extended CRL support flag is set: this flag will enable additional 1343 CRL functionality in future. 1344 1345 This work was sponsored by Google. 1346 [Steve Henson] 1347 1348 *) Add support for policy mappings extension. 1349 1350 This work was sponsored by Google. 1351 [Steve Henson] 1352 1353 *) Fixes to pathlength constraint, self issued certificate handling, 1354 policy processing to align with RFC3280 and PKITS tests. 1355 1356 This work was sponsored by Google. 1357 [Steve Henson] 1358 1359 *) Support for name constraints certificate extension. DN, email, DNS 1360 and URI types are currently supported. 1361 1362 This work was sponsored by Google. 1363 [Steve Henson] 1364 1365 *) To cater for systems that provide a pointer-based thread ID rather 1366 than numeric, deprecate the current numeric thread ID mechanism and 1367 replace it with a structure and associated callback type. This 1368 mechanism allows a numeric "hash" to be extracted from a thread ID in 1369 either case, and on platforms where pointers are larger than 'long', 1370 mixing is done to help ensure the numeric 'hash' is usable even if it 1371 can't be guaranteed unique. The default mechanism is to use "&errno" 1372 as a pointer-based thread ID to distinguish between threads. 1373 1374 Applications that want to provide their own thread IDs should now use 1375 CRYPTO_THREADID_set_callback() to register a callback that will call 1376 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer(). 1377 1378 Note that ERR_remove_state() is now deprecated, because it is tied 1379 to the assumption that thread IDs are numeric. ERR_remove_state(0) 1380 to free the current thread's error state should be replaced by 1381 ERR_remove_thread_state(NULL). 1382 1383 (This new approach replaces the functions CRYPTO_set_idptr_callback(), 1384 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in 1385 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an 1386 application was previously providing a numeric thread callback that 1387 was inappropriate for distinguishing threads, then uniqueness might 1388 have been obtained with &errno that happened immediately in the 1389 intermediate development versions of OpenSSL; this is no longer the 1390 case, the numeric thread callback will now override the automatic use 1391 of &errno.) 1392 [Geoff Thorpe, with help from Bodo Moeller] 1393 1394 *) Initial support for different CRL issuing certificates. This covers a 1395 simple case where the self issued certificates in the chain exist and 1396 the real CRL issuer is higher in the existing chain. 1397 1398 This work was sponsored by Google. 1399 [Steve Henson] 1400 1401 *) Removed effectively defunct crypto/store from the build. 1402 [Ben Laurie] 1403 1404 *) Revamp of STACK to provide stronger type-checking. Still to come: 1405 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE, 1406 ASN1_STRING, CONF_VALUE. 1407 [Ben Laurie] 1408 1409 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer 1410 RAM on SSL connections. This option can save about 34k per idle SSL. 1411 [Nick Mathewson] 1412 1413 *) Revamp of LHASH to provide stronger type-checking. Still to come: 1414 STACK, TXT_DB, bsearch, qsort. 1415 [Ben Laurie] 1416 1417 *) Initial support for Cryptographic Message Syntax (aka CMS) based 1418 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility, 1419 support for data, signedData, compressedData, digestedData and 1420 encryptedData, envelopedData types included. Scripts to check against 1421 RFC4134 examples draft and interop and consistency checks of many 1422 content types and variants. 1423 [Steve Henson] 1424 1425 *) Add options to enc utility to support use of zlib compression BIO. 1426 [Steve Henson] 1427 1428 *) Extend mk1mf to support importing of options and assembly language 1429 files from Configure script, currently only included in VC-WIN32. 1430 The assembly language rules can now optionally generate the source 1431 files from the associated perl scripts. 1432 [Steve Henson] 1433 1434 *) Implement remaining functionality needed to support GOST ciphersuites. 1435 Interop testing has been performed using CryptoPro implementations. 1436 [Victor B. Wagner <vitus@cryptocom.ru>] 1437 1438 *) s390x assembler pack. 1439 [Andy Polyakov] 1440 1441 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU 1442 "family." 1443 [Andy Polyakov] 1444 1445 *) Implement Opaque PRF Input TLS extension as specified in 1446 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an 1447 official specification yet and no extension type assignment by 1448 IANA exists, this extension (for now) will have to be explicitly 1449 enabled when building OpenSSL by providing the extension number 1450 to use. For example, specify an option 1451 1452 -DTLSEXT_TYPE_opaque_prf_input=0x9527 1453 1454 to the "config" or "Configure" script to enable the extension, 1455 assuming extension number 0x9527 (which is a completely arbitrary 1456 and unofficial assignment based on the MD5 hash of the Internet 1457 Draft). Note that by doing so, you potentially lose 1458 interoperability with other TLS implementations since these might 1459 be using the same extension number for other purposes. 1460 1461 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the 1462 opaque PRF input value to use in the handshake. This will create 1463 an interal copy of the length-'len' string at 'src', and will 1464 return non-zero for success. 1465 1466 To get more control and flexibility, provide a callback function 1467 by using 1468 1469 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb) 1470 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg) 1471 1472 where 1473 1474 int (*cb)(SSL *, void *peerinput, size_t len, void *arg); 1475 void *arg; 1476 1477 Callback function 'cb' will be called in handshakes, and is 1478 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate. 1479 Argument 'arg' is for application purposes (the value as given to 1480 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly 1481 be provided to the callback function). The callback function 1482 has to return non-zero to report success: usually 1 to use opaque 1483 PRF input just if possible, or 2 to enforce use of the opaque PRF 1484 input. In the latter case, the library will abort the handshake 1485 if opaque PRF input is not successfully negotiated. 1486 1487 Arguments 'peerinput' and 'len' given to the callback function 1488 will always be NULL and 0 in the case of a client. A server will 1489 see the client's opaque PRF input through these variables if 1490 available (NULL and 0 otherwise). Note that if the server 1491 provides an opaque PRF input, the length must be the same as the 1492 length of the client's opaque PRF input. 1493 1494 Note that the callback function will only be called when creating 1495 a new session (session resumption can resume whatever was 1496 previously negotiated), and will not be called in SSL 2.0 1497 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or 1498 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended 1499 for applications that need to enforce opaque PRF input. 1500 1501 [Bodo Moeller] 1502 1503 *) Update ssl code to support digests other than SHA1+MD5 for handshake 1504 MAC. 1505 1506 [Victor B. Wagner <vitus@cryptocom.ru>] 1507 1508 *) Add RFC4507 support to OpenSSL. This includes the corrections in 1509 RFC4507bis. The encrypted ticket format is an encrypted encoded 1510 SSL_SESSION structure, that way new session features are automatically 1511 supported. 1512 1513 If a client application caches session in an SSL_SESSION structure 1514 support is transparent because tickets are now stored in the encoded 1515 SSL_SESSION. 1516 1517 The SSL_CTX structure automatically generates keys for ticket 1518 protection in servers so again support should be possible 1519 with no application modification. 1520 1521 If a client or server wishes to disable RFC4507 support then the option 1522 SSL_OP_NO_TICKET can be set. 1523 1524 Add a TLS extension debugging callback to allow the contents of any client 1525 or server extensions to be examined. 1526 1527 This work was sponsored by Google. 1528 [Steve Henson] 1529 1530 *) Final changes to avoid use of pointer pointer casts in OpenSSL. 1531 OpenSSL should now compile cleanly on gcc 4.2 1532 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson] 1533 1534 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC 1535 support including streaming MAC support: this is required for GOST 1536 ciphersuite support. 1537 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson] 1538 1539 *) Add option -stream to use PKCS#7 streaming in smime utility. New 1540 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream() 1541 to output in BER and PEM format. 1542 [Steve Henson] 1543 1544 *) Experimental support for use of HMAC via EVP_PKEY interface. This 1545 allows HMAC to be handled via the EVP_DigestSign*() interface. The 1546 EVP_PKEY "key" in this case is the HMAC key, potentially allowing 1547 ENGINE support for HMAC keys which are unextractable. New -mac and 1548 -macopt options to dgst utility. 1549 [Steve Henson] 1550 1551 *) New option -sigopt to dgst utility. Update dgst to use 1552 EVP_Digest{Sign,Verify}*. These two changes make it possible to use 1553 alternative signing paramaters such as X9.31 or PSS in the dgst 1554 utility. 1555 [Steve Henson] 1556 1557 *) Change ssl_cipher_apply_rule(), the internal function that does 1558 the work each time a ciphersuite string requests enabling 1559 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or 1560 removing ("!foo+bar") a class of ciphersuites: Now it maintains 1561 the order of disabled ciphersuites such that those ciphersuites 1562 that most recently went from enabled to disabled not only stay 1563 in order with respect to each other, but also have higher priority 1564 than other disabled ciphersuites the next time ciphersuites are 1565 enabled again. 1566 1567 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable 1568 the same ciphersuites as with "HIGH" alone, but in a specific 1569 order where the PSK ciphersuites come first (since they are the 1570 most recently disabled ciphersuites when "HIGH" is parsed). 1571 1572 Also, change ssl_create_cipher_list() (using this new 1573 funcionality) such that between otherwise identical 1574 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in 1575 the default order. 1576 [Bodo Moeller] 1577 1578 *) Change ssl_create_cipher_list() so that it automatically 1579 arranges the ciphersuites in reasonable order before starting 1580 to process the rule string. Thus, the definition for "DEFAULT" 1581 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but 1582 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH". 1583 This makes it much easier to arrive at a reasonable default order 1584 in applications for which anonymous ciphers are OK (meaning 1585 that you can't actually use DEFAULT). 1586 [Bodo Moeller; suggested by Victor Duchovni] 1587 1588 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string 1589 processing) into multiple integers instead of setting 1590 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK", 1591 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer. 1592 (These masks as well as the individual bit definitions are hidden 1593 away into the non-exported interface ssl/ssl_locl.h, so this 1594 change to the definition of the SSL_CIPHER structure shouldn't 1595 affect applications.) This give us more bits for each of these 1596 categories, so there is no longer a need to coagulate AES128 and 1597 AES256 into a single algorithm bit, and to coagulate Camellia128 1598 and Camellia256 into a single algorithm bit, which has led to all 1599 kinds of kludges. 1600 1601 Thus, among other things, the kludge introduced in 0.9.7m and 1602 0.9.8e for masking out AES256 independently of AES128 or masking 1603 out Camellia256 independently of AES256 is not needed here in 0.9.9. 1604 1605 With the change, we also introduce new ciphersuite aliases that 1606 so far were missing: "AES128", "AES256", "CAMELLIA128", and 1607 "CAMELLIA256". 1608 [Bodo Moeller] 1609 1610 *) Add support for dsa-with-SHA224 and dsa-with-SHA256. 1611 Use the leftmost N bytes of the signature input if the input is 1612 larger than the prime q (with N being the size in bytes of q). 1613 [Nils Larsch] 1614 1615 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses 1616 it yet and it is largely untested. 1617 [Steve Henson] 1618 1619 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types. 1620 [Nils Larsch] 1621 1622 *) Initial incomplete changes to avoid need for function casts in OpenSSL 1623 some compilers (gcc 4.2 and later) reject their use. Safestack is 1624 reimplemented. Update ASN1 to avoid use of legacy functions. 1625 [Steve Henson] 1626 1627 *) Win32/64 targets are linked with Winsock2. 1628 [Andy Polyakov] 1629 1630 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected 1631 to external functions. This can be used to increase CRL handling 1632 efficiency especially when CRLs are very large by (for example) storing 1633 the CRL revoked certificates in a database. 1634 [Steve Henson] 1635 1636 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so 1637 new CRLs added to a directory can be used. New command line option 1638 -verify_return_error to s_client and s_server. This causes real errors 1639 to be returned by the verify callback instead of carrying on no matter 1640 what. This reflects the way a "real world" verify callback would behave. 1641 [Steve Henson] 1642 1643 *) GOST engine, supporting several GOST algorithms and public key formats. 1644 Kindly donated by Cryptocom. 1645 [Cryptocom] 1646 1647 *) Partial support for Issuing Distribution Point CRL extension. CRLs 1648 partitioned by DP are handled but no indirect CRL or reason partitioning 1649 (yet). Complete overhaul of CRL handling: now the most suitable CRL is 1650 selected via a scoring technique which handles IDP and AKID in CRLs. 1651 [Steve Henson] 1652 1653 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which 1654 will ultimately be used for all verify operations: this will remove the 1655 X509_STORE dependency on certificate verification and allow alternative 1656 lookup methods. X509_STORE based implementations of these two callbacks. 1657 [Steve Henson] 1658 1659 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names. 1660 Modify get_crl() to find a valid (unexpired) CRL if possible. 1661 [Steve Henson] 1662 1663 *) New function X509_CRL_match() to check if two CRLs are identical. Normally 1664 this would be called X509_CRL_cmp() but that name is already used by 1665 a function that just compares CRL issuer names. Cache several CRL 1666 extensions in X509_CRL structure and cache CRLDP in X509. 1667 [Steve Henson] 1668 1669 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name) 1670 this maps equivalent X509_NAME structures into a consistent structure. 1671 Name comparison can then be performed rapidly using memcmp(). 1672 [Steve Henson] 1673 1674 *) Non-blocking OCSP request processing. Add -timeout option to ocsp 1675 utility. 1676 [Steve Henson] 1677 1678 *) Allow digests to supply their own micalg string for S/MIME type using 1679 the ctrl EVP_MD_CTRL_MICALG. 1680 [Steve Henson] 1681 1682 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the 1683 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN 1684 ctrl. It can then customise the structure before and/or after signing 1685 if necessary. 1686 [Steve Henson] 1687 1688 *) New function OBJ_add_sigid() to allow application defined signature OIDs 1689 to be added to OpenSSLs internal tables. New function OBJ_sigid_free() 1690 to free up any added signature OIDs. 1691 [Steve Henson] 1692 1693 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(), 1694 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal 1695 digest and cipher tables. New options added to openssl utility: 1696 list-message-digest-algorithms and list-cipher-algorithms. 1697 [Steve Henson] 1698 1699 *) Change the array representation of binary polynomials: the list 1700 of degrees of non-zero coefficients is now terminated with -1. 1701 Previously it was terminated with 0, which was also part of the 1702 value; thus, the array representation was not applicable to 1703 polynomials where t^0 has coefficient zero. This change makes 1704 the array representation useful in a more general context. 1705 [Douglas Stebila] 1706 1707 *) Various modifications and fixes to SSL/TLS cipher string 1708 handling. For ECC, the code now distinguishes between fixed ECDH 1709 with RSA certificates on the one hand and with ECDSA certificates 1710 on the other hand, since these are separate ciphersuites. The 1711 unused code for Fortezza ciphersuites has been removed. 1712 1713 For consistency with EDH, ephemeral ECDH is now called "EECDH" 1714 (not "ECDHE"). For consistency with the code for DH 1715 certificates, use of ECDH certificates is now considered ECDH 1716 authentication, not RSA or ECDSA authentication (the latter is 1717 merely the CA's signing algorithm and not actively used in the 1718 protocol). 1719 1720 The temporary ciphersuite alias "ECCdraft" is no longer 1721 available, and ECC ciphersuites are no longer excluded from "ALL" 1722 and "DEFAULT". The following aliases now exist for RFC 4492 1723 ciphersuites, most of these by analogy with the DH case: 1724 1725 kECDHr - ECDH cert, signed with RSA 1726 kECDHe - ECDH cert, signed with ECDSA 1727 kECDH - ECDH cert (signed with either RSA or ECDSA) 1728 kEECDH - ephemeral ECDH 1729 ECDH - ECDH cert or ephemeral ECDH 1730 1731 aECDH - ECDH cert 1732 aECDSA - ECDSA cert 1733 ECDSA - ECDSA cert 1734 1735 AECDH - anonymous ECDH 1736 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH") 1737 1738 [Bodo Moeller] 1739 1740 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported. 1741 Use correct micalg parameters depending on digest(s) in signed message. 1742 [Steve Henson] 1743 1744 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process 1745 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code. 1746 [Steve Henson] 1747 1748 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit 1749 an engine to register a method. Add ENGINE lookups for methods and 1750 functional reference processing. 1751 [Steve Henson] 1752 1753 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of 1754 EVP_{Sign,Verify}* which allow an application to customise the signature 1755 process. 1756 [Steve Henson] 1757 1758 *) New -resign option to smime utility. This adds one or more signers 1759 to an existing PKCS#7 signedData structure. Also -md option to use an 1760 alternative message digest algorithm for signing. 1761 [Steve Henson] 1762 1763 *) Tidy up PKCS#7 routines and add new functions to make it easier to 1764 create PKCS7 structures containing multiple signers. Update smime 1765 application to support multiple signers. 1766 [Steve Henson] 1767 1768 *) New -macalg option to pkcs12 utility to allow setting of an alternative 1769 digest MAC. 1770 [Steve Henson] 1771 1772 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC. 1773 Reorganize PBE internals to lookup from a static table using NIDs, 1774 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl: 1775 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative 1776 PRF which will be automatically used with PBES2. 1777 [Steve Henson] 1778 1779 *) Replace the algorithm specific calls to generate keys in "req" with the 1780 new API. 1781 [Steve Henson] 1782 1783 *) Update PKCS#7 enveloped data routines to use new API. This is now 1784 supported by any public key method supporting the encrypt operation. A 1785 ctrl is added to allow the public key algorithm to examine or modify 1786 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is 1787 a no op. 1788 [Steve Henson] 1789 1790 *) Add a ctrl to asn1 method to allow a public key algorithm to express 1791 a default digest type to use. In most cases this will be SHA1 but some 1792 algorithms (such as GOST) need to specify an alternative digest. The 1793 return value indicates how strong the prefernce is 1 means optional and 1794 2 is mandatory (that is it is the only supported type). Modify 1795 ASN1_item_sign() to accept a NULL digest argument to indicate it should 1796 use the default md. Update openssl utilities to use the default digest 1797 type for signing if it is not explicitly indicated. 1798 [Steve Henson] 1799 1800 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New 1801 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant 1802 signing method from the key type. This effectively removes the link 1803 between digests and public key types. 1804 [Steve Henson] 1805 1806 *) Add an OID cross reference table and utility functions. Its purpose is to 1807 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1, 1808 rsaEncryption. This will allow some of the algorithm specific hackery 1809 needed to use the correct OID to be removed. 1810 [Steve Henson] 1811 1812 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO 1813 structures for PKCS7_sign(). They are now set up by the relevant public 1814 key ASN1 method. 1815 [Steve Henson] 1816 1817 *) Add provisional EC pkey method with support for ECDSA and ECDH. 1818 [Steve Henson] 1819 1820 *) Add support for key derivation (agreement) in the API, DH method and 1821 pkeyutl. 1822 [Steve Henson] 1823 1824 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support 1825 public and private key formats. As a side effect these add additional 1826 command line functionality not previously available: DSA signatures can be 1827 generated and verified using pkeyutl and DH key support and generation in 1828 pkey, genpkey. 1829 [Steve Henson] 1830 1831 *) BeOS support. 1832 [Oliver Tappe <zooey@hirschkaefer.de>] 1833 1834 *) New make target "install_html_docs" installs HTML renditions of the 1835 manual pages. 1836 [Oliver Tappe <zooey@hirschkaefer.de>] 1837 1838 *) New utility "genpkey" this is analagous to "genrsa" etc except it can 1839 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to 1840 support key and parameter generation and add initial key generation 1841 functionality for RSA. 1842 [Steve Henson] 1843 1844 *) Add functions for main EVP_PKEY_method operations. The undocumented 1845 functions EVP_PKEY_{encrypt,decrypt} have been renamed to 1846 EVP_PKEY_{encrypt,decrypt}_old. 1847 [Steve Henson] 1848 1849 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public 1850 key API, doesn't do much yet. 1851 [Steve Henson] 1852 1853 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about 1854 public key algorithms. New option to openssl utility: 1855 "list-public-key-algorithms" to print out info. 1856 [Steve Henson] 1857 1858 *) Implement the Supported Elliptic Curves Extension for 1859 ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1860 [Douglas Stebila] 1861 1862 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or 1863 EVP_CIPHER structures to avoid later problems in EVP_cleanup(). 1864 [Steve Henson] 1865 1866 *) New utilities pkey and pkeyparam. These are similar to algorithm specific 1867 utilities such as rsa, dsa, dsaparam etc except they process any key 1868 type. 1869 [Steve Henson] 1870 1871 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New 1872 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(), 1873 EVP_PKEY_print_param() to print public key data from an EVP_PKEY 1874 structure. 1875 [Steve Henson] 1876 1877 *) Initial support for pluggable public key ASN1. 1878 De-spaghettify the public key ASN1 handling. Move public and private 1879 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate 1880 algorithm specific handling to a single module within the relevant 1881 algorithm directory. Add functions to allow (near) opaque processing 1882 of public and private key structures. 1883 [Steve Henson] 1884 1885 *) Implement the Supported Point Formats Extension for 1886 ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1887 [Douglas Stebila] 1888 1889 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members 1890 for the psk identity [hint] and the psk callback functions to the 1891 SSL_SESSION, SSL and SSL_CTX structure. 1892 1893 New ciphersuites: 1894 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA, 1895 PSK-AES256-CBC-SHA 1896 1897 New functions: 1898 SSL_CTX_use_psk_identity_hint 1899 SSL_get_psk_identity_hint 1900 SSL_get_psk_identity 1901 SSL_use_psk_identity_hint 1902 1903 [Mika Kousa and Pasi Eronen of Nokia Corporation] 1904 1905 *) Add RFC 3161 compliant time stamp request creation, response generation 1906 and response verification functionality. 1907 [Zolt��n Gl��zik <zglozik@opentsa.org>, The OpenTSA Project] 1908 1909 *) Add initial support for TLS extensions, specifically for the server_name 1910 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 1911 have new members for a host name. The SSL data structure has an 1912 additional member SSL_CTX *initial_ctx so that new sessions can be 1913 stored in that context to allow for session resumption, even after the 1914 SSL has been switched to a new SSL_CTX in reaction to a client's 1915 server_name extension. 1916 1917 New functions (subject to change): 1918 1919 SSL_get_servername() 1920 SSL_get_servername_type() 1921 SSL_set_SSL_CTX() 1922 1923 New CTRL codes and macros (subject to change): 1924 1925 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 1926 - SSL_CTX_set_tlsext_servername_callback() 1927 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 1928 - SSL_CTX_set_tlsext_servername_arg() 1929 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 1930 1931 openssl s_client has a new '-servername ...' option. 1932 1933 openssl s_server has new options '-servername_host ...', '-cert2 ...', 1934 '-key2 ...', '-servername_fatal' (subject to change). This allows 1935 testing the HostName extension for a specific single host name ('-cert' 1936 and '-key' remain fallbacks for handshakes without HostName 1937 negotiation). If the unrecogninzed_name alert has to be sent, this by 1938 default is a warning; it becomes fatal with the '-servername_fatal' 1939 option. 1940 1941 [Peter Sylvester, Remy Allais, Christophe Renou] 1942 1943 *) Whirlpool hash implementation is added. 1944 [Andy Polyakov] 1945 1946 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to 1947 bn(64,32). Because of instruction set limitations it doesn't have 1948 any negative impact on performance. This was done mostly in order 1949 to make it possible to share assembler modules, such as bn_mul_mont 1950 implementations, between 32- and 64-bit builds without hassle. 1951 [Andy Polyakov] 1952 1953 *) Move code previously exiled into file crypto/ec/ec2_smpt.c 1954 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP 1955 macro. 1956 [Bodo Moeller] 1957 1958 *) New candidate for BIGNUM assembler implementation, bn_mul_mont, 1959 dedicated Montgomery multiplication procedure, is introduced. 1960 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher 1961 "64-bit" performance on certain 32-bit targets. 1962 [Andy Polyakov] 1963 1964 *) New option SSL_OP_NO_COMP to disable use of compression selectively 1965 in SSL structures. New SSL ctrl to set maximum send fragment size. 1966 Save memory by seeting the I/O buffer sizes dynamically instead of 1967 using the maximum available value. 1968 [Steve Henson] 1969 1970 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code 1971 in addition to the text details. 1972 [Bodo Moeller] 1973 1974 *) Very, very preliminary EXPERIMENTAL support for printing of general 1975 ASN1 structures. This currently produces rather ugly output and doesn't 1976 handle several customised structures at all. 1977 [Steve Henson] 1978 1979 *) Integrated support for PVK file format and some related formats such 1980 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support 1981 these in the 'rsa' and 'dsa' utilities. 1982 [Steve Henson] 1983 1984 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line. 1985 [Steve Henson] 1986 1987 *) Remove the ancient ASN1_METHOD code. This was only ever used in one 1988 place for the (very old) "NETSCAPE" format certificates which are now 1989 handled using new ASN1 code equivalents. 1990 [Steve Henson] 1991 1992 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD 1993 pointer and make the SSL_METHOD parameter in SSL_CTX_new, 1994 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'. 1995 [Nils Larsch] 1996 1997 *) Modify CRL distribution points extension code to print out previously 1998 unsupported fields. Enhance extension setting code to allow setting of 1999 all fields. 2000 [Steve Henson] 2001 2002 *) Add print and set support for Issuing Distribution Point CRL extension. 2003 [Steve Henson] 2004 2005 *) Change 'Configure' script to enable Camellia by default. 2006 [NTT] 2007 2008 Changes between 0.9.8m and 0.9.8n [24 Mar 2010] 2009 2010 *) When rejecting SSL/TLS records due to an incorrect version number, never 2011 update s->server with a new major version number. As of 2012 - OpenSSL 0.9.8m if 'short' is a 16-bit type, 2013 - OpenSSL 0.9.8f if 'short' is longer than 16 bits, 2014 the previous behavior could result in a read attempt at NULL when 2015 receiving specific incorrect SSL/TLS records once record payload 2016 protection is active. (CVE-2010-0740) 2017 [Bodo Moeller, Adam Langley <agl@chromium.org>] 2018 2019 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL 2020 could be crashed if the relevant tables were not present (e.g. chrooted). 2021 [Tomas Hoger <thoger@redhat.com>] 2022 2023 Changes between 0.9.8l and 0.9.8m [25 Feb 2010] 2024 2025 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245) 2026 [Martin Olsson, Neel Mehta] 2027 2028 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to 2029 accommodate for stack sorting, always a write lock!). 2030 [Bodo Moeller] 2031 2032 *) On some versions of WIN32 Heap32Next is very slow. This can cause 2033 excessive delays in the RAND_poll(): over a minute. As a workaround 2034 include a time check in the inner Heap32Next loop too. 2035 [Steve Henson] 2036 2037 *) The code that handled flushing of data in SSL/TLS originally used the 2038 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused 2039 the problem outlined in PR#1949. The fix suggested there however can 2040 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions 2041 of Apache). So instead simplify the code to flush unconditionally. 2042 This should be fine since flushing with no data to flush is a no op. 2043 [Steve Henson] 2044 2045 *) Handle TLS versions 2.0 and later properly and correctly use the 2046 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way 2047 off ancient servers have a habit of sticking around for a while... 2048 [Steve Henson] 2049 2050 *) Modify compression code so it frees up structures without using the 2051 ex_data callbacks. This works around a problem where some applications 2052 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when 2053 restarting) then use compression (e.g. SSL with compression) later. 2054 This results in significant per-connection memory leaks and 2055 has caused some security issues including CVE-2008-1678 and 2056 CVE-2009-4355. 2057 [Steve Henson] 2058 2059 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't 2060 change when encrypting or decrypting. 2061 [Bodo Moeller] 2062 2063 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to 2064 connect and renegotiate with servers which do not support RI. 2065 Until RI is more widely deployed this option is enabled by default. 2066 [Steve Henson] 2067 2068 *) Add "missing" ssl ctrls to clear options and mode. 2069 [Steve Henson] 2070 2071 *) If client attempts to renegotiate and doesn't support RI respond with 2072 a no_renegotiation alert as required by RFC5746. Some renegotiating 2073 TLS clients will continue a connection gracefully when they receive 2074 the alert. Unfortunately OpenSSL mishandled this alert and would hang 2075 waiting for a server hello which it will never receive. Now we treat a 2076 received no_renegotiation alert as a fatal error. This is because 2077 applications requesting a renegotiation might well expect it to succeed 2078 and would have no code in place to handle the server denying it so the 2079 only safe thing to do is to terminate the connection. 2080 [Steve Henson] 2081 2082 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if 2083 peer supports secure renegotiation and 0 otherwise. Print out peer 2084 renegotiation support in s_client/s_server. 2085 [Steve Henson] 2086 2087 *) Replace the highly broken and deprecated SPKAC certification method with 2088 the updated NID creation version. This should correctly handle UTF8. 2089 [Steve Henson] 2090 2091 *) Implement RFC5746. Re-enable renegotiation but require the extension 2092 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION 2093 turns out to be a bad idea. It has been replaced by 2094 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with 2095 SSL_CTX_set_options(). This is really not recommended unless you 2096 know what you are doing. 2097 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson] 2098 2099 *) Fixes to stateless session resumption handling. Use initial_ctx when 2100 issuing and attempting to decrypt tickets in case it has changed during 2101 servername handling. Use a non-zero length session ID when attempting 2102 stateless session resumption: this makes it possible to determine if 2103 a resumption has occurred immediately after receiving server hello 2104 (several places in OpenSSL subtly assume this) instead of later in 2105 the handshake. 2106 [Steve Henson] 2107 2108 *) The functions ENGINE_ctrl(), OPENSSL_isservice(), 2109 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error 2110 fixes for a few places where the return code is not checked 2111 correctly. 2112 [Julia Lawall <julia@diku.dk>] 2113 2114 *) Add --strict-warnings option to Configure script to include devteam 2115 warnings in other configurations. 2116 [Steve Henson] 2117 2118 *) Add support for --libdir option and LIBDIR variable in makefiles. This 2119 makes it possible to install openssl libraries in locations which 2120 have names other than "lib", for example "/usr/lib64" which some 2121 systems need. 2122 [Steve Henson, based on patch from Jeremy Utley] 2123 2124 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of 2125 X690 8.9.12 and can produce some misleading textual output of OIDs. 2126 [Steve Henson, reported by Dan Kaminsky] 2127 2128 *) Delete MD2 from algorithm tables. This follows the recommendation in 2129 several standards that it is not used in new applications due to 2130 several cryptographic weaknesses. For binary compatibility reasons 2131 the MD2 API is still compiled in by default. 2132 [Steve Henson] 2133 2134 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved 2135 and restored. 2136 [Steve Henson] 2137 2138 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and 2139 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name 2140 clash. 2141 [Guenter <lists@gknw.net>] 2142 2143 *) Fix the server certificate chain building code to use X509_verify_cert(), 2144 it used to have an ad-hoc builder which was unable to cope with anything 2145 other than a simple chain. 2146 [David Woodhouse <dwmw2@infradead.org>, Steve Henson] 2147 2148 *) Don't check self signed certificate signatures in X509_verify_cert() 2149 by default (a flag can override this): it just wastes time without 2150 adding any security. As a useful side effect self signed root CAs 2151 with non-FIPS digests are now usable in FIPS mode. 2152 [Steve Henson] 2153 2154 *) In dtls1_process_out_of_seq_message() the check if the current message 2155 is already buffered was missing. For every new message was memory 2156 allocated, allowing an attacker to perform an denial of service attack 2157 with sending out of seq handshake messages until there is no memory 2158 left. Additionally every future messege was buffered, even if the 2159 sequence number made no sense and would be part of another handshake. 2160 So only messages with sequence numbers less than 10 in advance will be 2161 buffered. (CVE-2009-1378) 2162 [Robin Seggelmann, discovered by Daniel Mentz] 2163 2164 *) Records are buffered if they arrive with a future epoch to be 2165 processed after finishing the corresponding handshake. There is 2166 currently no limitation to this buffer allowing an attacker to perform 2167 a DOS attack with sending records with future epochs until there is no 2168 memory left. This patch adds the pqueue_size() function to detemine 2169 the size of a buffer and limits the record buffer to 100 entries. 2170 (CVE-2009-1377) 2171 [Robin Seggelmann, discovered by Daniel Mentz] 2172 2173 *) Keep a copy of frag->msg_header.frag_len so it can be used after the 2174 parent structure is freed. (CVE-2009-1379) 2175 [Daniel Mentz] 2176 2177 *) Handle non-blocking I/O properly in SSL_shutdown() call. 2178 [Darryl Miles <darryl-mailinglists@netbauds.net>] 2179 2180 *) Add 2.5.4.* OIDs 2181 [Ilya O. <vrghost@gmail.com>] 2182 2183 Changes between 0.9.8k and 0.9.8l [5 Nov 2009] 2184 2185 *) Disable renegotiation completely - this fixes a severe security 2186 problem (CVE-2009-3555) at the cost of breaking all 2187 renegotiation. Renegotiation can be re-enabled by setting 2188 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at 2189 run-time. This is really not recommended unless you know what 2190 you're doing. 2191 [Ben Laurie] 2192 2193 Changes between 0.9.8j and 0.9.8k [25 Mar 2009] 2194 2195 *) Don't set val to NULL when freeing up structures, it is freed up by 2196 underlying code. If sizeof(void *) > sizeof(long) this can result in 2197 zeroing past the valid field. (CVE-2009-0789) 2198 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>] 2199 2200 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not 2201 checked correctly. This would allow some invalid signed attributes to 2202 appear to verify correctly. (CVE-2009-0591) 2203 [Ivan Nestlerode <inestlerode@us.ibm.com>] 2204 2205 *) Reject UniversalString and BMPString types with invalid lengths. This 2206 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have 2207 a legal length. (CVE-2009-0590) 2208 [Steve Henson] 2209 2210 *) Set S/MIME signing as the default purpose rather than setting it 2211 unconditionally. This allows applications to override it at the store 2212 level. 2213 [Steve Henson] 2214 2215 *) Permit restricted recursion of ASN1 strings. This is needed in practice 2216 to handle some structures. 2217 [Steve Henson] 2218 2219 *) Improve efficiency of mem_gets: don't search whole buffer each time 2220 for a '\n' 2221 [Jeremy Shapiro <jnshapir@us.ibm.com>] 2222 2223 *) New -hex option for openssl rand. 2224 [Matthieu Herrb] 2225 2226 *) Print out UTF8String and NumericString when parsing ASN1. 2227 [Steve Henson] 2228 2229 *) Support NumericString type for name components. 2230 [Steve Henson] 2231 2232 *) Allow CC in the environment to override the automatically chosen 2233 compiler. Note that nothing is done to ensure flags work with the 2234 chosen compiler. 2235 [Ben Laurie] 2236 2237 Changes between 0.9.8i and 0.9.8j [07 Jan 2009] 2238 2239 *) Properly check EVP_VerifyFinal() and similar return values 2240 (CVE-2008-5077). 2241 [Ben Laurie, Bodo Moeller, Google Security Team] 2242 2243 *) Enable TLS extensions by default. 2244 [Ben Laurie] 2245 2246 *) Allow the CHIL engine to be loaded, whether the application is 2247 multithreaded or not. (This does not release the developer from the 2248 obligation to set up the dynamic locking callbacks.) 2249 [Sander Temme <sander@temme.net>] 2250 2251 *) Use correct exit code if there is an error in dgst command. 2252 [Steve Henson; problem pointed out by Roland Dirlewanger] 2253 2254 *) Tweak Configure so that you need to say "experimental-jpake" to enable 2255 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. 2256 [Bodo Moeller] 2257 2258 *) Add experimental JPAKE support, including demo authentication in 2259 s_client and s_server. 2260 [Ben Laurie] 2261 2262 *) Set the comparison function in v3_addr_canonize(). 2263 [Rob Austein <sra@hactrn.net>] 2264 2265 *) Add support for XMPP STARTTLS in s_client. 2266 [Philip Paeps <philip@freebsd.org>] 2267 2268 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior 2269 to ensure that even with this option, only ciphersuites in the 2270 server's preference list will be accepted. (Note that the option 2271 applies only when resuming a session, so the earlier behavior was 2272 just about the algorithm choice for symmetric cryptography.) 2273 [Bodo Moeller] 2274 2275 Changes between 0.9.8h and 0.9.8i [15 Sep 2008] 2276 2277 *) Fix NULL pointer dereference if a DTLS server received 2278 ChangeCipherSpec as first record (CVE-2009-1386). 2279 [PR #1679] 2280 2281 *) Fix a state transitition in s3_srvr.c and d1_srvr.c 2282 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...). 2283 [Nagendra Modadugu] 2284 2285 *) The fix in 0.9.8c that supposedly got rid of unsafe 2286 double-checked locking was incomplete for RSA blinding, 2287 addressing just one layer of what turns out to have been 2288 doubly unsafe triple-checked locking. 2289 2290 So now fix this for real by retiring the MONT_HELPER macro 2291 in crypto/rsa/rsa_eay.c. 2292 2293 [Bodo Moeller; problem pointed out by Marius Schilder] 2294 2295 *) Various precautionary measures: 2296 2297 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). 2298 2299 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). 2300 (NB: This would require knowledge of the secret session ticket key 2301 to exploit, in which case you'd be SOL either way.) 2302 2303 - Change bn_nist.c so that it will properly handle input BIGNUMs 2304 outside the expected range. 2305 2306 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG 2307 builds. 2308 2309 [Neel Mehta, Bodo Moeller] 2310 2311 *) Allow engines to be "soft loaded" - i.e. optionally don't die if 2312 the load fails. Useful for distros. 2313 [Ben Laurie and the FreeBSD team] 2314 2315 *) Add support for Local Machine Keyset attribute in PKCS#12 files. 2316 [Steve Henson] 2317 2318 *) Fix BN_GF2m_mod_arr() top-bit cleanup code. 2319 [Huang Ying] 2320 2321 *) Expand ENGINE to support engine supplied SSL client certificate functions. 2322 2323 This work was sponsored by Logica. 2324 [Steve Henson] 2325 2326 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows 2327 keystores. Support for SSL/TLS client authentication too. 2328 Not compiled unless enable-capieng specified to Configure. 2329 2330 This work was sponsored by Logica. 2331 [Steve Henson] 2332 2333 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using 2334 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain 2335 attribute creation routines such as certifcate requests and PKCS#12 2336 files. 2337 [Steve Henson] 2338 2339 Changes between 0.9.8g and 0.9.8h [28 May 2008] 2340 2341 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS 2342 handshake which could lead to a cilent crash as found using the 2343 Codenomicon TLS test suite (CVE-2008-1672) 2344 [Steve Henson, Mark Cox] 2345 2346 *) Fix double free in TLS server name extensions which could lead to 2347 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891) 2348 [Joe Orton] 2349 2350 *) Clear error queue in SSL_CTX_use_certificate_chain_file() 2351 2352 Clear the error queue to ensure that error entries left from 2353 older function calls do not interfere with the correct operation. 2354 [Lutz Jaenicke, Erik de Castro Lopo] 2355 2356 *) Remove root CA certificates of commercial CAs: 2357 2358 The OpenSSL project does not recommend any specific CA and does not 2359 have any policy with respect to including or excluding any CA. 2360 Therefore it does not make any sense to ship an arbitrary selection 2361 of root CA certificates with the OpenSSL software. 2362 [Lutz Jaenicke] 2363 2364 *) RSA OAEP patches to fix two separate invalid memory reads. 2365 The first one involves inputs when 'lzero' is greater than 2366 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes 2367 before the beginning of from). The second one involves inputs where 2368 the 'db' section contains nothing but zeroes (there is a one-byte 2369 invalid read after the end of 'db'). 2370 [Ivan Nestlerode <inestlerode@us.ibm.com>] 2371 2372 *) Partial backport from 0.9.9-dev: 2373 2374 Introduce bn_mul_mont (dedicated Montgomery multiplication 2375 procedure) as a candidate for BIGNUM assembler implementation. 2376 While 0.9.9-dev uses assembler for various architectures, only 2377 x86_64 is available by default here in the 0.9.8 branch, and 2378 32-bit x86 is available through a compile-time setting. 2379 2380 To try the 32-bit x86 assembler implementation, use Configure 2381 option "enable-montasm" (which exists only for this backport). 2382 2383 As "enable-montasm" for 32-bit x86 disclaims code stability 2384 anyway, in this constellation we activate additional code 2385 backported from 0.9.9-dev for further performance improvements, 2386 namely BN_from_montgomery_word. (To enable this otherwise, 2387 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".) 2388 2389 [Andy Polyakov (backport partially by Bodo Moeller)] 2390 2391 *) Add TLS session ticket callback. This allows an application to set 2392 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed 2393 values. This is useful for key rollover for example where several key 2394 sets may exist with different names. 2395 [Steve Henson] 2396 2397 *) Reverse ENGINE-internal logic for caching default ENGINE handles. 2398 This was broken until now in 0.9.8 releases, such that the only way 2399 a registered ENGINE could be used (assuming it initialises 2400 successfully on the host) was to explicitly set it as the default 2401 for the relevant algorithms. This is in contradiction with 0.9.7 2402 behaviour and the documentation. With this fix, when an ENGINE is 2403 registered into a given algorithm's table of implementations, the 2404 'uptodate' flag is reset so that auto-discovery will be used next 2405 time a new context for that algorithm attempts to select an 2406 implementation. 2407 [Ian Lister (tweaked by Geoff Thorpe)] 2408 2409 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 2410 implemention in the following ways: 2411 2412 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be 2413 hard coded. 2414 2415 Lack of BER streaming support means one pass streaming processing is 2416 only supported if data is detached: setting the streaming flag is 2417 ignored for embedded content. 2418 2419 CMS support is disabled by default and must be explicitly enabled 2420 with the enable-cms configuration option. 2421 [Steve Henson] 2422 2423 *) Update the GMP engine glue to do direct copies between BIGNUM and 2424 mpz_t when openssl and GMP use the same limb size. Otherwise the 2425 existing "conversion via a text string export" trick is still used. 2426 [Paul Sheer <paulsheer@gmail.com>] 2427 2428 *) Zlib compression BIO. This is a filter BIO which compressed and 2429 uncompresses any data passed through it. 2430 [Steve Henson] 2431 2432 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement 2433 RFC3394 compatible AES key wrapping. 2434 [Steve Henson] 2435 2436 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): 2437 sets string data without copying. X509_ALGOR_set0() and 2438 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) 2439 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data 2440 from an X509_ATTRIBUTE structure optionally checking it occurs only 2441 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied 2442 data. 2443 [Steve Henson] 2444 2445 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() 2446 to get the expected BN_FLG_CONSTTIME behavior. 2447 [Bodo Moeller (Google)] 2448 2449 *) Netware support: 2450 2451 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets 2452 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) 2453 - added some more tests to do_tests.pl 2454 - fixed RunningProcess usage so that it works with newer LIBC NDKs too 2455 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency 2456 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, 2457 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc 2458 - various changes to netware.pl to enable gcc-cross builds on Win32 2459 platform 2460 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) 2461 - various changes to fix missing prototype warnings 2462 - fixed x86nasm.pl to create correct asm files for NASM COFF output 2463 - added AES, WHIRLPOOL and CPUID assembler code to build files 2464 - added missing AES assembler make rules to mk1mf.pl 2465 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply 2466 [Guenter Knauf <eflash@gmx.net>] 2467 2468 *) Implement certificate status request TLS extension defined in RFC3546. 2469 A client can set the appropriate parameters and receive the encoded 2470 OCSP response via a callback. A server can query the supplied parameters 2471 and set the encoded OCSP response in the callback. Add simplified examples 2472 to s_client and s_server. 2473 [Steve Henson] 2474 2475 Changes between 0.9.8f and 0.9.8g [19 Oct 2007] 2476 2477 *) Fix various bugs: 2478 + Binary incompatibility of ssl_ctx_st structure 2479 + DTLS interoperation with non-compliant servers 2480 + Don't call get_session_cb() without proposed session 2481 + Fix ia64 assembler code 2482 [Andy Polyakov, Steve Henson] 2483 2484 Changes between 0.9.8e and 0.9.8f [11 Oct 2007] 2485 2486 *) DTLS Handshake overhaul. There were longstanding issues with 2487 OpenSSL DTLS implementation, which were making it impossible for 2488 RFC 4347 compliant client to communicate with OpenSSL server. 2489 Unfortunately just fixing these incompatibilities would "cut off" 2490 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e 2491 server keeps tolerating non RFC compliant syntax. The opposite is 2492 not true, 0.9.8f client can not communicate with earlier server. 2493 This update even addresses CVE-2007-4995. 2494 [Andy Polyakov] 2495 2496 *) Changes to avoid need for function casts in OpenSSL: some compilers 2497 (gcc 4.2 and later) reject their use. 2498 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, 2499 Steve Henson] 2500 2501 *) Add RFC4507 support to OpenSSL. This includes the corrections in 2502 RFC4507bis. The encrypted ticket format is an encrypted encoded 2503 SSL_SESSION structure, that way new session features are automatically 2504 supported. 2505 2506 If a client application caches session in an SSL_SESSION structure 2507 support is transparent because tickets are now stored in the encoded 2508 SSL_SESSION. 2509 2510 The SSL_CTX structure automatically generates keys for ticket 2511 protection in servers so again support should be possible 2512 with no application modification. 2513 2514 If a client or server wishes to disable RFC4507 support then the option 2515 SSL_OP_NO_TICKET can be set. 2516 2517 Add a TLS extension debugging callback to allow the contents of any client 2518 or server extensions to be examined. 2519 2520 This work was sponsored by Google. 2521 [Steve Henson] 2522 2523 *) Add initial support for TLS extensions, specifically for the server_name 2524 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 2525 have new members for a host name. The SSL data structure has an 2526 additional member SSL_CTX *initial_ctx so that new sessions can be 2527 stored in that context to allow for session resumption, even after the 2528 SSL has been switched to a new SSL_CTX in reaction to a client's 2529 server_name extension. 2530 2531 New functions (subject to change): 2532 2533 SSL_get_servername() 2534 SSL_get_servername_type() 2535 SSL_set_SSL_CTX() 2536 2537 New CTRL codes and macros (subject to change): 2538 2539 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 2540 - SSL_CTX_set_tlsext_servername_callback() 2541 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 2542 - SSL_CTX_set_tlsext_servername_arg() 2543 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 2544 2545 openssl s_client has a new '-servername ...' option. 2546 2547 openssl s_server has new options '-servername_host ...', '-cert2 ...', 2548 '-key2 ...', '-servername_fatal' (subject to change). This allows 2549 testing the HostName extension for a specific single host name ('-cert' 2550 and '-key' remain fallbacks for handshakes without HostName 2551 negotiation). If the unrecogninzed_name alert has to be sent, this by 2552 default is a warning; it becomes fatal with the '-servername_fatal' 2553 option. 2554 2555 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson] 2556 2557 *) Add AES and SSE2 assembly language support to VC++ build. 2558 [Steve Henson] 2559 2560 *) Mitigate attack on final subtraction in Montgomery reduction. 2561 [Andy Polyakov] 2562 2563 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 2564 (which previously caused an internal error). 2565 [Bodo Moeller] 2566 2567 *) Squeeze another 10% out of IGE mode when in != out. 2568 [Ben Laurie] 2569 2570 *) AES IGE mode speedup. 2571 [Dean Gaudet (Google)] 2572 2573 *) Add the Korean symmetric 128-bit cipher SEED (see 2574 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and 2575 add SEED ciphersuites from RFC 4162: 2576 2577 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" 2578 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" 2579 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" 2580 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" 2581 2582 To minimize changes between patchlevels in the OpenSSL 0.9.8 2583 series, SEED remains excluded from compilation unless OpenSSL 2584 is configured with 'enable-seed'. 2585 [KISA, Bodo Moeller] 2586 2587 *) Mitigate branch prediction attacks, which can be practical if a 2588 single processor is shared, allowing a spy process to extract 2589 information. For detailed background information, see 2590 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron, 2591 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL 2592 and Necessary Software Countermeasures"). The core of the change 2593 are new versions BN_div_no_branch() and 2594 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), 2595 respectively, which are slower, but avoid the security-relevant 2596 conditional branches. These are automatically called by BN_div() 2597 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one 2598 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to 2599 remove a conditional branch. 2600 2601 BN_FLG_CONSTTIME is the new name for the previous 2602 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just 2603 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag 2604 in the exponent causes BN_mod_exp_mont() to use the alternative 2605 implementation in BN_mod_exp_mont_consttime().) The old name 2606 remains as a deprecated alias. 2607 2608 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general 2609 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses 2610 constant-time implementations for more than just exponentiation. 2611 Here too the old name is kept as a deprecated alias. 2612 2613 BN_BLINDING_new() will now use BN_dup() for the modulus so that 2614 the BN_BLINDING structure gets an independent copy of the 2615 modulus. This means that the previous "BIGNUM *m" argument to 2616 BN_BLINDING_new() and to BN_BLINDING_create_param() now 2617 essentially becomes "const BIGNUM *m", although we can't actually 2618 change this in the header file before 0.9.9. It allows 2619 RSA_setup_blinding() to use BN_with_flags() on the modulus to 2620 enable BN_FLG_CONSTTIME. 2621 2622 [Matthew D Wood (Intel Corp)] 2623 2624 *) In the SSL/TLS server implementation, be strict about session ID 2625 context matching (which matters if an application uses a single 2626 external cache for different purposes). Previously, 2627 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was 2628 set. This did ensure strict client verification, but meant that, 2629 with applications using a single external cache for quite 2630 different requirements, clients could circumvent ciphersuite 2631 restrictions for a given session ID context by starting a session 2632 in a different context. 2633 [Bodo Moeller] 2634 2635 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 2636 a ciphersuite string such as "DEFAULT:RSA" cannot enable 2637 authentication-only ciphersuites. 2638 [Bodo Moeller] 2639 2640 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was 2641 not complete and could lead to a possible single byte overflow 2642 (CVE-2007-5135) [Ben Laurie] 2643 2644 Changes between 0.9.8d and 0.9.8e [23 Feb 2007] 2645 2646 *) Since AES128 and AES256 (and similarly Camellia128 and 2647 Camellia256) share a single mask bit in the logic of 2648 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 2649 kludge to work properly if AES128 is available and AES256 isn't 2650 (or if Camellia128 is available and Camellia256 isn't). 2651 [Victor Duchovni] 2652 2653 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c 2654 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): 2655 When a point or a seed is encoded in a BIT STRING, we need to 2656 prevent the removal of trailing zero bits to get the proper DER 2657 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case 2658 of a NamedBitList, for which trailing 0 bits need to be removed.) 2659 [Bodo Moeller] 2660 2661 *) Have SSL/TLS server implementation tolerate "mismatched" record 2662 protocol version while receiving ClientHello even if the 2663 ClientHello is fragmented. (The server can't insist on the 2664 particular protocol version it has chosen before the ServerHello 2665 message has informed the client about his choice.) 2666 [Bodo Moeller] 2667 2668 *) Add RFC 3779 support. 2669 [Rob Austein for ARIN, Ben Laurie] 2670 2671 *) Load error codes if they are not already present instead of using a 2672 static variable. This allows them to be cleanly unloaded and reloaded. 2673 Improve header file function name parsing. 2674 [Steve Henson] 2675 2676 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO 2677 or CAPABILITY handshake as required by RFCs. 2678 [Goetz Babin-Ebell] 2679 2680 Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 2681 2682 *) Introduce limits to prevent malicious keys being able to 2683 cause a denial of service. (CVE-2006-2940) 2684 [Steve Henson, Bodo Moeller] 2685 2686 *) Fix ASN.1 parsing of certain invalid structures that can result 2687 in a denial of service. (CVE-2006-2937) [Steve Henson] 2688 2689 *) Fix buffer overflow in SSL_get_shared_ciphers() function. 2690 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 2691 2692 *) Fix SSL client code which could crash if connecting to a 2693 malicious SSLv2 server. (CVE-2006-4343) 2694 [Tavis Ormandy and Will Drewry, Google Security Team] 2695 2696 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 2697 match only those. Before that, "AES256-SHA" would be interpreted 2698 as a pattern and match "AES128-SHA" too (since AES128-SHA got 2699 the same strength classification in 0.9.7h) as we currently only 2700 have a single AES bit in the ciphersuite description bitmap. 2701 That change, however, also applied to ciphersuite strings such as 2702 "RC4-MD5" that intentionally matched multiple ciphersuites -- 2703 namely, SSL 2.0 ciphersuites in addition to the more common ones 2704 from SSL 3.0/TLS 1.0. 2705 2706 So we change the selection algorithm again: Naming an explicit 2707 ciphersuite selects this one ciphersuite, and any other similar 2708 ciphersuite (same bitmap) from *other* protocol versions. 2709 Thus, "RC4-MD5" again will properly select both the SSL 2.0 2710 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 2711 2712 Since SSL 2.0 does not have any ciphersuites for which the 2713 128/256 bit distinction would be relevant, this works for now. 2714 The proper fix will be to use different bits for AES128 and 2715 AES256, which would have avoided the problems from the beginning; 2716 however, bits are scarce, so we can only do this in a new release 2717 (not just a patchlevel) when we can change the SSL_CIPHER 2718 definition to split the single 'unsigned long mask' bitmap into 2719 multiple values to extend the available space. 2720 2721 [Bodo Moeller] 2722 2723 Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 2724 2725 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 2726 (CVE-2006-4339) [Ben Laurie and Google Security Team] 2727 2728 *) Add AES IGE and biIGE modes. 2729 [Ben Laurie] 2730 2731 *) Change the Unix randomness entropy gathering to use poll() when 2732 possible instead of select(), since the latter has some 2733 undesirable limitations. 2734 [Darryl Miles via Richard Levitte and Bodo Moeller] 2735 2736 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special 2737 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 2738 cannot be implicitly activated as part of, e.g., the "AES" alias. 2739 However, please upgrade to OpenSSL 0.9.9[-dev] for 2740 non-experimental use of the ECC ciphersuites to get TLS extension 2741 support, which is required for curve and point format negotiation 2742 to avoid potential handshake problems. 2743 [Bodo Moeller] 2744 2745 *) Disable rogue ciphersuites: 2746 2747 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 2748 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 2749 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 2750 2751 The latter two were purportedly from 2752 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 2753 appear there. 2754 2755 Also deactivate the remaining ciphersuites from 2756 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 2757 unofficial, and the ID has long expired. 2758 [Bodo Moeller] 2759 2760 *) Fix RSA blinding Heisenbug (problems sometimes occured on 2761 dual-core machines) and other potential thread-safety issues. 2762 [Bodo Moeller] 2763 2764 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 2765 versions), which is now available for royalty-free use 2766 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). 2767 Also, add Camellia TLS ciphersuites from RFC 4132. 2768 2769 To minimize changes between patchlevels in the OpenSSL 0.9.8 2770 series, Camellia remains excluded from compilation unless OpenSSL 2771 is configured with 'enable-camellia'. 2772 [NTT] 2773 2774 *) Disable the padding bug check when compression is in use. The padding 2775 bug check assumes the first packet is of even length, this is not 2776 necessarily true if compresssion is enabled and can result in false 2777 positives causing handshake failure. The actual bug test is ancient 2778 code so it is hoped that implementations will either have fixed it by 2779 now or any which still have the bug do not support compression. 2780 [Steve Henson] 2781 2782 Changes between 0.9.8a and 0.9.8b [04 May 2006] 2783 2784 *) When applying a cipher rule check to see if string match is an explicit 2785 cipher suite and only match that one cipher suite if it is. 2786 [Steve Henson] 2787 2788 *) Link in manifests for VC++ if needed. 2789 [Austin Ziegler <halostatue@gmail.com>] 2790 2791 *) Update support for ECC-based TLS ciphersuites according to 2792 draft-ietf-tls-ecc-12.txt with proposed changes (but without 2793 TLS extensions, which are supported starting with the 0.9.9 2794 branch, not in the OpenSSL 0.9.8 branch). 2795 [Douglas Stebila] 2796 2797 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 2798 opaque EVP_CIPHER_CTX handling. 2799 [Steve Henson] 2800 2801 *) Fixes and enhancements to zlib compression code. We now only use 2802 "zlib1.dll" and use the default __cdecl calling convention on Win32 2803 to conform with the standards mentioned here: 2804 http://www.zlib.net/DLL_FAQ.txt 2805 Static zlib linking now works on Windows and the new --with-zlib-include 2806 --with-zlib-lib options to Configure can be used to supply the location 2807 of the headers and library. Gracefully handle case where zlib library 2808 can't be loaded. 2809 [Steve Henson] 2810 2811 *) Several fixes and enhancements to the OID generation code. The old code 2812 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 2813 handle numbers larger than ULONG_MAX, truncated printing and had a 2814 non standard OBJ_obj2txt() behaviour. 2815 [Steve Henson] 2816 2817 *) Add support for building of engines under engine/ as shared libraries 2818 under VC++ build system. 2819 [Steve Henson] 2820 2821 *) Corrected the numerous bugs in the Win32 path splitter in DSO. 2822 Hopefully, we will not see any false combination of paths any more. 2823 [Richard Levitte] 2824 2825 Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 2826 2827 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 2828 (part of SSL_OP_ALL). This option used to disable the 2829 countermeasure against man-in-the-middle protocol-version 2830 rollback in the SSL 2.0 server implementation, which is a bad 2831 idea. (CVE-2005-2969) 2832 2833 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 2834 for Information Security, National Institute of Advanced Industrial 2835 Science and Technology [AIST], Japan)] 2836 2837 *) Add two function to clear and return the verify parameter flags. 2838 [Steve Henson] 2839 2840 *) Keep cipherlists sorted in the source instead of sorting them at 2841 runtime, thus removing the need for a lock. 2842 [Nils Larsch] 2843 2844 *) Avoid some small subgroup attacks in Diffie-Hellman. 2845 [Nick Mathewson and Ben Laurie] 2846 2847 *) Add functions for well-known primes. 2848 [Nick Mathewson] 2849 2850 *) Extended Windows CE support. 2851 [Satoshi Nakamura and Andy Polyakov] 2852 2853 *) Initialize SSL_METHOD structures at compile time instead of during 2854 runtime, thus removing the need for a lock. 2855 [Steve Henson] 2856 2857 *) Make PKCS7_decrypt() work even if no certificate is supplied by 2858 attempting to decrypt each encrypted key in turn. Add support to 2859 smime utility. 2860 [Steve Henson] 2861 2862 Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 2863 2864 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 2865 OpenSSL 0.9.8.] 2866 2867 *) Add libcrypto.pc and libssl.pc for those who feel they need them. 2868 [Richard Levitte] 2869 2870 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 2871 key into the same file any more. 2872 [Richard Levitte] 2873 2874 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 2875 [Andy Polyakov] 2876 2877 *) Add -utf8 command line and config file option to 'ca'. 2878 [Stefan <stf@udoma.org] 2879 2880 *) Removed the macro des_crypt(), as it seems to conflict with some 2881 libraries. Use DES_crypt(). 2882 [Richard Levitte] 2883 2884 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 2885 involves renaming the source and generated shared-libs for 2886 both. The engines will accept the corrected or legacy ids 2887 ('ncipher' and '4758_cca' respectively) when binding. NB, 2888 this only applies when building 'shared'. 2889 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 2890 2891 *) Add attribute functions to EVP_PKEY structure. Modify 2892 PKCS12_create() to recognize a CSP name attribute and 2893 use it. Make -CSP option work again in pkcs12 utility. 2894 [Steve Henson] 2895 2896 *) Add new functionality to the bn blinding code: 2897 - automatic re-creation of the BN_BLINDING parameters after 2898 a fixed number of uses (currently 32) 2899 - add new function for parameter creation 2900 - introduce flags to control the update behaviour of the 2901 BN_BLINDING parameters 2902 - hide BN_BLINDING structure 2903 Add a second BN_BLINDING slot to the RSA structure to improve 2904 performance when a single RSA object is shared among several 2905 threads. 2906 [Nils Larsch] 2907 2908 *) Add support for DTLS. 2909 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 2910 2911 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 2912 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 2913 [Walter Goulet] 2914 2915 *) Remove buggy and incompletet DH cert support from 2916 ssl/ssl_rsa.c and ssl/s3_both.c 2917 [Nils Larsch] 2918 2919 *) Use SHA-1 instead of MD5 as the default digest algorithm for 2920 the apps/openssl applications. 2921 [Nils Larsch] 2922 2923 *) Compile clean with "-Wall -Wmissing-prototypes 2924 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 2925 DEBUG_SAFESTACK must also be set. 2926 [Ben Laurie] 2927 2928 *) Change ./Configure so that certain algorithms can be disabled by default. 2929 The new counterpiece to "no-xxx" is "enable-xxx". 2930 2931 The patented RC5 and MDC2 algorithms will now be disabled unless 2932 "enable-rc5" and "enable-mdc2", respectively, are specified. 2933 2934 (IDEA remains enabled despite being patented. This is because IDEA 2935 is frequently required for interoperability, and there is no license 2936 fee for non-commercial use. As before, "no-idea" can be used to 2937 avoid this algorithm.) 2938 2939 [Bodo Moeller] 2940 2941 *) Add processing of proxy certificates (see RFC 3820). This work was 2942 sponsored by KTH (The Royal Institute of Technology in Stockholm) and 2943 EGEE (Enabling Grids for E-science in Europe). 2944 [Richard Levitte] 2945 2946 *) RC4 performance overhaul on modern architectures/implementations, such 2947 as Intel P4, IA-64 and AMD64. 2948 [Andy Polyakov] 2949 2950 *) New utility extract-section.pl. This can be used specify an alternative 2951 section number in a pod file instead of having to treat each file as 2952 a separate case in Makefile. This can be done by adding two lines to the 2953 pod file: 2954 2955 =for comment openssl_section:XXX 2956 2957 The blank line is mandatory. 2958 2959 [Steve Henson] 2960 2961 *) New arguments -certform, -keyform and -pass for s_client and s_server 2962 to allow alternative format key and certificate files and passphrase 2963 sources. 2964 [Steve Henson] 2965 2966 *) New structure X509_VERIFY_PARAM which combines current verify parameters, 2967 update associated structures and add various utility functions. 2968 2969 Add new policy related verify parameters, include policy checking in 2970 standard verify code. Enhance 'smime' application with extra parameters 2971 to support policy checking and print out. 2972 [Steve Henson] 2973 2974 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 2975 Nehemiah processors. These extensions support AES encryption in hardware 2976 as well as RNG (though RNG support is currently disabled). 2977 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 2978 2979 *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 2980 [Geoff Thorpe] 2981 2982 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 2983 [Andy Polyakov and a number of other people] 2984 2985 *) Improved PowerPC platform support. Most notably BIGNUM assembler 2986 implementation contributed by IBM. 2987 [Suresh Chari, Peter Waltenberg, Andy Polyakov] 2988 2989 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 2990 exponent rather than 'unsigned long'. There is a corresponding change to 2991 the new 'rsa_keygen' element of the RSA_METHOD structure. 2992 [Jelte Jansen, Geoff Thorpe] 2993 2994 *) Functionality for creating the initial serial number file is now 2995 moved from CA.pl to the 'ca' utility with a new option -create_serial. 2996 2997 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 2998 number file to 1, which is bound to cause problems. To avoid 2999 the problems while respecting compatibility between different 0.9.7 3000 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 3001 CA.pl for serial number initialization. With the new release 0.9.8, 3002 we can fix the problem directly in the 'ca' utility.) 3003 [Steve Henson] 3004 3005 *) Reduced header interdepencies by declaring more opaque objects in 3006 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 3007 give fewer recursive includes, which could break lazy source code - so 3008 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 3009 developers should define this symbol when building and using openssl to 3010 ensure they track the recommended behaviour, interfaces, [etc], but 3011 backwards-compatible behaviour prevails when this isn't defined. 3012 [Geoff Thorpe] 3013 3014 *) New function X509_POLICY_NODE_print() which prints out policy nodes. 3015 [Steve Henson] 3016 3017 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 3018 This will generate a random key of the appropriate length based on the 3019 cipher context. The EVP_CIPHER can provide its own random key generation 3020 routine to support keys of a specific form. This is used in the des and 3021 3des routines to generate a key of the correct parity. Update S/MIME 3022 code to use new functions and hence generate correct parity DES keys. 3023 Add EVP_CHECK_DES_KEY #define to return an error if the key is not 3024 valid (weak or incorrect parity). 3025 [Steve Henson] 3026 3027 *) Add a local set of CRLs that can be used by X509_verify_cert() as well 3028 as looking them up. This is useful when the verified structure may contain 3029 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 3030 present unless the new PKCS7_NO_CRL flag is asserted. 3031 [Steve Henson] 3032 3033 *) Extend ASN1 oid configuration module. It now additionally accepts the 3034 syntax: 3035 3036 shortName = some long name, 1.2.3.4 3037 [Steve Henson] 3038 3039 *) Reimplemented the BN_CTX implementation. There is now no more static 3040 limitation on the number of variables it can handle nor the depth of the 3041 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 3042 information can now expand as required, and rather than having a single 3043 static array of bignums, BN_CTX now uses a linked-list of such arrays 3044 allowing it to expand on demand whilst maintaining the usefulness of 3045 BN_CTX's "bundling". 3046 [Geoff Thorpe] 3047 3048 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 3049 to allow all RSA operations to function using a single BN_CTX. 3050 [Geoff Thorpe] 3051 3052 *) Preliminary support for certificate policy evaluation and checking. This 3053 is initially intended to pass the tests outlined in "Conformance Testing 3054 of Relying Party Client Certificate Path Processing Logic" v1.07. 3055 [Steve Henson] 3056 3057 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 3058 remained unused and not that useful. A variety of other little bignum 3059 tweaks and fixes have also been made continuing on from the audit (see 3060 below). 3061 [Geoff Thorpe] 3062 3063 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 3064 associated ASN1, EVP and SSL functions and old ASN1 macros. 3065 [Richard Levitte] 3066 3067 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 3068 and this should never fail. So the return value from the use of 3069 BN_set_word() (which can fail due to needless expansion) is now deprecated; 3070 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 3071 [Geoff Thorpe] 3072 3073 *) BN_CTX_get() should return zero-valued bignums, providing the same 3074 initialised value as BN_new(). 3075 [Geoff Thorpe, suggested by Ulf M��ller] 3076 3077 *) Support for inhibitAnyPolicy certificate extension. 3078 [Steve Henson] 3079 3080 *) An audit of the BIGNUM code is underway, for which debugging code is 3081 enabled when BN_DEBUG is defined. This makes stricter enforcements on what 3082 is considered valid when processing BIGNUMs, and causes execution to 3083 assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 3084 further steps are taken to deliberately pollute unused data in BIGNUM 3085 structures to try and expose faulty code further on. For now, openssl will 3086 (in its default mode of operation) continue to tolerate the inconsistent 3087 forms that it has tolerated in the past, but authors and packagers should 3088 consider trying openssl and their own applications when compiled with 3089 these debugging symbols defined. It will help highlight potential bugs in 3090 their own code, and will improve the test coverage for OpenSSL itself. At 3091 some point, these tighter rules will become openssl's default to improve 3092 maintainability, though the assert()s and other overheads will remain only 3093 in debugging configurations. See bn.h for more details. 3094 [Geoff Thorpe, Nils Larsch, Ulf M��ller] 3095 3096 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 3097 that can only be obtained through BN_CTX_new() (which implicitly 3098 initialises it). The presence of this function only made it possible 3099 to overwrite an existing structure (and cause memory leaks). 3100 [Geoff Thorpe] 3101 3102 *) Because of the callback-based approach for implementing LHASH as a 3103 template type, lh_insert() adds opaque objects to hash-tables and 3104 lh_doall() or lh_doall_arg() are typically used with a destructor callback 3105 to clean up those corresponding objects before destroying the hash table 3106 (and losing the object pointers). So some over-zealous constifications in 3107 LHASH have been relaxed so that lh_insert() does not take (nor store) the 3108 objects as "const" and the lh_doall[_arg] callback wrappers are not 3109 prototyped to have "const" restrictions on the object pointers they are 3110 given (and so aren't required to cast them away any more). 3111 [Geoff Thorpe] 3112 3113 *) The tmdiff.h API was so ugly and minimal that our own timing utility 3114 (speed) prefers to use its own implementation. The two implementations 3115 haven't been consolidated as yet (volunteers?) but the tmdiff API has had 3116 its object type properly exposed (MS_TM) instead of casting to/from "char 3117 *". This may still change yet if someone realises MS_TM and "ms_time_***" 3118 aren't necessarily the greatest nomenclatures - but this is what was used 3119 internally to the implementation so I've used that for now. 3120 [Geoff Thorpe] 3121 3122 *) Ensure that deprecated functions do not get compiled when 3123 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 3124 the self-tests were still using deprecated key-generation functions so 3125 these have been updated also. 3126 [Geoff Thorpe] 3127 3128 *) Reorganise PKCS#7 code to separate the digest location functionality 3129 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 3130 New function PKCS7_set_digest() to set the digest type for PKCS#7 3131 digestedData type. Add additional code to correctly generate the 3132 digestedData type and add support for this type in PKCS7 initialization 3133 functions. 3134 [Steve Henson] 3135 3136 *) New function PKCS7_set0_type_other() this initializes a PKCS7 3137 structure of type "other". 3138 [Steve Henson] 3139 3140 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 3141 sure the loop does correctly stop and breaking ("division by zero") 3142 modulus operations are not performed. The (pre-generated) prime 3143 table crypto/bn/bn_prime.h was already correct, but it could not be 3144 re-generated on some platforms because of the "division by zero" 3145 situation in the script. 3146 [Ralf S. Engelschall] 3147 3148 *) Update support for ECC-based TLS ciphersuites according to 3149 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 3150 SHA-1 now is only used for "small" curves (where the 3151 representation of a field element takes up to 24 bytes); for 3152 larger curves, the field element resulting from ECDH is directly 3153 used as premaster secret. 3154 [Douglas Stebila (Sun Microsystems Laboratories)] 3155 3156 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 3157 curve secp160r1 to the tests. 3158 [Douglas Stebila (Sun Microsystems Laboratories)] 3159 3160 *) Add the possibility to load symbols globally with DSO. 3161 [G��tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 3162 3163 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 3164 control of the error stack. 3165 [Richard Levitte] 3166 3167 *) Add support for STORE in ENGINE. 3168 [Richard Levitte] 3169 3170 *) Add the STORE type. The intention is to provide a common interface 3171 to certificate and key stores, be they simple file-based stores, or 3172 HSM-type store, or LDAP stores, or... 3173 NOTE: The code is currently UNTESTED and isn't really used anywhere. 3174 [Richard Levitte] 3175 3176 *) Add a generic structure called OPENSSL_ITEM. This can be used to 3177 pass a list of arguments to any function as well as provide a way 3178 for a function to pass data back to the caller. 3179 [Richard Levitte] 3180 3181 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 3182 works like BUF_strdup() but can be used to duplicate a portion of 3183 a string. The copy gets NUL-terminated. BUF_memdup() duplicates 3184 a memory area. 3185 [Richard Levitte] 3186 3187 *) Add the function sk_find_ex() which works like sk_find(), but will 3188 return an index to an element even if an exact match couldn't be 3189 found. The index is guaranteed to point at the element where the 3190 searched-for key would be inserted to preserve sorting order. 3191 [Richard Levitte] 3192 3193 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 3194 takes an extra flags argument for optional functionality. Currently, 3195 the following flags are defined: 3196 3197 OBJ_BSEARCH_VALUE_ON_NOMATCH 3198 This one gets OBJ_bsearch_ex() to return a pointer to the first 3199 element where the comparing function returns a negative or zero 3200 number. 3201 3202 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 3203 This one gets OBJ_bsearch_ex() to return a pointer to the first 3204 element where the comparing function returns zero. This is useful 3205 if there are more than one element where the comparing function 3206 returns zero. 3207 [Richard Levitte] 3208 3209 *) Make it possible to create self-signed certificates with 'openssl ca' 3210 in such a way that the self-signed certificate becomes part of the 3211 CA database and uses the same mechanisms for serial number generation 3212 as all other certificate signing. The new flag '-selfsign' enables 3213 this functionality. Adapt CA.sh and CA.pl.in. 3214 [Richard Levitte] 3215 3216 *) Add functionality to check the public key of a certificate request 3217 against a given private. This is useful to check that a certificate 3218 request can be signed by that key (self-signing). 3219 [Richard Levitte] 3220 3221 *) Make it possible to have multiple active certificates with the same 3222 subject in the CA index file. This is done only if the keyword 3223 'unique_subject' is set to 'no' in the main CA section (default 3224 if 'CA_default') of the configuration file. The value is saved 3225 with the database itself in a separate index attribute file, 3226 named like the index file with '.attr' appended to the name. 3227 [Richard Levitte] 3228 3229 *) Generate muti valued AVAs using '+' notation in config files for 3230 req and dirName. 3231 [Steve Henson] 3232 3233 *) Support for nameConstraints certificate extension. 3234 [Steve Henson] 3235 3236 *) Support for policyConstraints certificate extension. 3237 [Steve Henson] 3238 3239 *) Support for policyMappings certificate extension. 3240 [Steve Henson] 3241 3242 *) Make sure the default DSA_METHOD implementation only uses its 3243 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 3244 and change its own handlers to be NULL so as to remove unnecessary 3245 indirection. This lets alternative implementations fallback to the 3246 default implementation more easily. 3247 [Geoff Thorpe] 3248 3249 *) Support for directoryName in GeneralName related extensions 3250 in config files. 3251 [Steve Henson] 3252 3253 *) Make it possible to link applications using Makefile.shared. 3254 Make that possible even when linking against static libraries! 3255 [Richard Levitte] 3256 3257 *) Support for single pass processing for S/MIME signing. This now 3258 means that S/MIME signing can be done from a pipe, in addition 3259 cleartext signing (multipart/signed type) is effectively streaming 3260 and the signed data does not need to be all held in memory. 3261 3262 This is done with a new flag PKCS7_STREAM. When this flag is set 3263 PKCS7_sign() only initializes the PKCS7 structure and the actual signing 3264 is done after the data is output (and digests calculated) in 3265 SMIME_write_PKCS7(). 3266 [Steve Henson] 3267 3268 *) Add full support for -rpath/-R, both in shared libraries and 3269 applications, at least on the platforms where it's known how 3270 to do it. 3271 [Richard Levitte] 3272 3273 *) In crypto/ec/ec_mult.c, implement fast point multiplication with 3274 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 3275 will now compute a table of multiples of the generator that 3276 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 3277 faster (notably in the case of a single point multiplication, 3278 scalar * generator). 3279 [Nils Larsch, Bodo Moeller] 3280 3281 *) IPv6 support for certificate extensions. The various extensions 3282 which use the IP:a.b.c.d can now take IPv6 addresses using the 3283 formats of RFC1884 2.2 . IPv6 addresses are now also displayed 3284 correctly. 3285 [Steve Henson] 3286 3287 *) Added an ENGINE that implements RSA by performing private key 3288 exponentiations with the GMP library. The conversions to and from 3289 GMP's mpz_t format aren't optimised nor are any montgomery forms 3290 cached, and on x86 it appears OpenSSL's own performance has caught up. 3291 However there are likely to be other architectures where GMP could 3292 provide a boost. This ENGINE is not built in by default, but it can be 3293 specified at Configure time and should be accompanied by the necessary 3294 linker additions, eg; 3295 ./config -DOPENSSL_USE_GMP -lgmp 3296 [Geoff Thorpe] 3297 3298 *) "openssl engine" will not display ENGINE/DSO load failure errors when 3299 testing availability of engines with "-t" - the old behaviour is 3300 produced by increasing the feature's verbosity with "-tt". 3301 [Geoff Thorpe] 3302 3303 *) ECDSA routines: under certain error conditions uninitialized BN objects 3304 could be freed. Solution: make sure initialization is performed early 3305 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 3306 via PR#459) 3307 [Lutz Jaenicke] 3308 3309 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 3310 and DH_METHOD (eg. by ENGINE implementations) to override the normal 3311 software implementations. For DSA and DH, parameter generation can 3312 also be overriden by providing the appropriate method callbacks. 3313 [Geoff Thorpe] 3314 3315 *) Change the "progress" mechanism used in key-generation and 3316 primality testing to functions that take a new BN_GENCB pointer in 3317 place of callback/argument pairs. The new API functions have "_ex" 3318 postfixes and the older functions are reimplemented as wrappers for 3319 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 3320 declarations of the old functions to help (graceful) attempts to 3321 migrate to the new functions. Also, the new key-generation API 3322 functions operate on a caller-supplied key-structure and return 3323 success/failure rather than returning a key or NULL - this is to 3324 help make "keygen" another member function of RSA_METHOD etc. 3325 3326 Example for using the new callback interface: 3327 3328 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 3329 void *my_arg = ...; 3330 BN_GENCB my_cb; 3331 3332 BN_GENCB_set(&my_cb, my_callback, my_arg); 3333 3334 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 3335 /* For the meaning of a, b in calls to my_callback(), see the 3336 * documentation of the function that calls the callback. 3337 * cb will point to my_cb; my_arg can be retrieved as cb->arg. 3338 * my_callback should return 1 if it wants BN_is_prime_ex() 3339 * to continue, or 0 to stop. 3340 */ 3341 3342 [Geoff Thorpe] 3343 3344 *) Change the ZLIB compression method to be stateful, and make it 3345 available to TLS with the number defined in 3346 draft-ietf-tls-compression-04.txt. 3347 [Richard Levitte] 3348 3349 *) Add the ASN.1 structures and functions for CertificatePair, which 3350 is defined as follows (according to X.509_4thEditionDraftV6.pdf): 3351 3352 CertificatePair ::= SEQUENCE { 3353 forward [0] Certificate OPTIONAL, 3354 reverse [1] Certificate OPTIONAL, 3355 -- at least one of the pair shall be present -- } 3356 3357 Also implement the PEM functions to read and write certificate 3358 pairs, and defined the PEM tag as "CERTIFICATE PAIR". 3359 3360 This needed to be defined, mostly for the sake of the LDAP 3361 attribute crossCertificatePair, but may prove useful elsewhere as 3362 well. 3363 [Richard Levitte] 3364 3365 *) Make it possible to inhibit symlinking of shared libraries in 3366 Makefile.shared, for Cygwin's sake. 3367 [Richard Levitte] 3368 3369 *) Extend the BIGNUM API by creating a function 3370 void BN_set_negative(BIGNUM *a, int neg); 3371 and a macro that behave like 3372 int BN_is_negative(const BIGNUM *a); 3373 3374 to avoid the need to access 'a->neg' directly in applications. 3375 [Nils Larsch] 3376 3377 *) Implement fast modular reduction for pseudo-Mersenne primes 3378 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 3379 EC_GROUP_new_curve_GFp() will now automatically use this 3380 if applicable. 3381 [Nils Larsch <nla@trustcenter.de>] 3382 3383 *) Add new lock type (CRYPTO_LOCK_BN). 3384 [Bodo Moeller] 3385 3386 *) Change the ENGINE framework to automatically load engines 3387 dynamically from specific directories unless they could be 3388 found to already be built in or loaded. Move all the 3389 current engines except for the cryptodev one to a new 3390 directory engines/. 3391 The engines in engines/ are built as shared libraries if 3392 the "shared" options was given to ./Configure or ./config. 3393 Otherwise, they are inserted in libcrypto.a. 3394 /usr/local/ssl/engines is the default directory for dynamic 3395 engines, but that can be overriden at configure time through 3396 the usual use of --prefix and/or --openssldir, and at run 3397 time with the environment variable OPENSSL_ENGINES. 3398 [Geoff Thorpe and Richard Levitte] 3399 3400 *) Add Makefile.shared, a helper makefile to build shared 3401 libraries. Addapt Makefile.org. 3402 [Richard Levitte] 3403 3404 *) Add version info to Win32 DLLs. 3405 [Peter 'Luna' Runestig" <peter@runestig.com>] 3406 3407 *) Add new 'medium level' PKCS#12 API. Certificates and keys 3408 can be added using this API to created arbitrary PKCS#12 3409 files while avoiding the low level API. 3410 3411 New options to PKCS12_create(), key or cert can be NULL and 3412 will then be omitted from the output file. The encryption 3413 algorithm NIDs can be set to -1 for no encryption, the mac 3414 iteration count can be set to 0 to omit the mac. 3415 3416 Enhance pkcs12 utility by making the -nokeys and -nocerts 3417 options work when creating a PKCS#12 file. New option -nomac 3418 to omit the mac, NONE can be set for an encryption algorithm. 3419 New code is modified to use the enhanced PKCS12_create() 3420 instead of the low level API. 3421 [Steve Henson] 3422 3423 *) Extend ASN1 encoder to support indefinite length constructed 3424 encoding. This can output sequences tags and octet strings in 3425 this form. Modify pk7_asn1.c to support indefinite length 3426 encoding. This is experimental and needs additional code to 3427 be useful, such as an ASN1 bio and some enhanced streaming 3428 PKCS#7 code. 3429 3430 Extend template encode functionality so that tagging is passed 3431 down to the template encoder. 3432 [Steve Henson] 3433 3434 *) Let 'openssl req' fail if an argument to '-newkey' is not 3435 recognized instead of using RSA as a default. 3436 [Bodo Moeller] 3437 3438 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 3439 As these are not official, they are not included in "ALL"; 3440 the "ECCdraft" ciphersuite group alias can be used to select them. 3441 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 3442 3443 *) Add ECDH engine support. 3444 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 3445 3446 *) Add ECDH in new directory crypto/ecdh/. 3447 [Douglas Stebila (Sun Microsystems Laboratories)] 3448 3449 *) Let BN_rand_range() abort with an error after 100 iterations 3450 without success (which indicates a broken PRNG). 3451 [Bodo Moeller] 3452 3453 *) Change BN_mod_sqrt() so that it verifies that the input value 3454 is really the square of the return value. (Previously, 3455 BN_mod_sqrt would show GIGO behaviour.) 3456 [Bodo Moeller] 3457 3458 *) Add named elliptic curves over binary fields from X9.62, SECG, 3459 and WAP/WTLS; add OIDs that were still missing. 3460 3461 [Sheueling Chang Shantz and Douglas Stebila 3462 (Sun Microsystems Laboratories)] 3463 3464 *) Extend the EC library for elliptic curves over binary fields 3465 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 3466 New EC_METHOD: 3467 3468 EC_GF2m_simple_method 3469 3470 New API functions: 3471 3472 EC_GROUP_new_curve_GF2m 3473 EC_GROUP_set_curve_GF2m 3474 EC_GROUP_get_curve_GF2m 3475 EC_POINT_set_affine_coordinates_GF2m 3476 EC_POINT_get_affine_coordinates_GF2m 3477 EC_POINT_set_compressed_coordinates_GF2m 3478 3479 Point compression for binary fields is disabled by default for 3480 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 3481 enable it). 3482 3483 As binary polynomials are represented as BIGNUMs, various members 3484 of the EC_GROUP and EC_POINT data structures can be shared 3485 between the implementations for prime fields and binary fields; 3486 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 3487 are essentially identical to their ..._GFp counterparts. 3488 (For simplicity, the '..._GFp' prefix has been dropped from 3489 various internal method names.) 3490 3491 An internal 'field_div' method (similar to 'field_mul' and 3492 'field_sqr') has been added; this is used only for binary fields. 3493 3494 [Sheueling Chang Shantz and Douglas Stebila 3495 (Sun Microsystems Laboratories)] 3496 3497 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 3498 through methods ('mul', 'precompute_mult'). 3499 3500 The generic implementations (now internally called 'ec_wNAF_mul' 3501 and 'ec_wNAF_precomputed_mult') remain the default if these 3502 methods are undefined. 3503 3504 [Sheueling Chang Shantz and Douglas Stebila 3505 (Sun Microsystems Laboratories)] 3506 3507 *) New function EC_GROUP_get_degree, which is defined through 3508 EC_METHOD. For curves over prime fields, this returns the bit 3509 length of the modulus. 3510 3511 [Sheueling Chang Shantz and Douglas Stebila 3512 (Sun Microsystems Laboratories)] 3513 3514 *) New functions EC_GROUP_dup, EC_POINT_dup. 3515 (These simply call ..._new and ..._copy). 3516 3517 [Sheueling Chang Shantz and Douglas Stebila 3518 (Sun Microsystems Laboratories)] 3519 3520 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 3521 Polynomials are represented as BIGNUMs (where the sign bit is not 3522 used) in the following functions [macros]: 3523 3524 BN_GF2m_add 3525 BN_GF2m_sub [= BN_GF2m_add] 3526 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 3527 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 3528 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 3529 BN_GF2m_mod_inv 3530 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 3531 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 3532 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 3533 BN_GF2m_cmp [= BN_ucmp] 3534 3535 (Note that only the 'mod' functions are actually for fields GF(2^m). 3536 BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 3537 3538 For some functions, an the irreducible polynomial defining a 3539 field can be given as an 'unsigned int[]' with strictly 3540 decreasing elements giving the indices of those bits that are set; 3541 i.e., p[] represents the polynomial 3542 f(t) = t^p[0] + t^p[1] + ... + t^p[k] 3543 where 3544 p[0] > p[1] > ... > p[k] = 0. 3545 This applies to the following functions: 3546 3547 BN_GF2m_mod_arr 3548 BN_GF2m_mod_mul_arr 3549 BN_GF2m_mod_sqr_arr 3550 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 3551 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 3552 BN_GF2m_mod_exp_arr 3553 BN_GF2m_mod_sqrt_arr 3554 BN_GF2m_mod_solve_quad_arr 3555 BN_GF2m_poly2arr 3556 BN_GF2m_arr2poly 3557 3558 Conversion can be performed by the following functions: 3559 3560 BN_GF2m_poly2arr 3561 BN_GF2m_arr2poly 3562 3563 bntest.c has additional tests for binary polynomial arithmetic. 3564 3565 Two implementations for BN_GF2m_mod_div() are available. 3566 The default algorithm simply uses BN_GF2m_mod_inv() and 3567 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 3568 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 3569 copyright notice in crypto/bn/bn_gf2m.c before enabling it). 3570 3571 [Sheueling Chang Shantz and Douglas Stebila 3572 (Sun Microsystems Laboratories)] 3573 3574 *) Add new error code 'ERR_R_DISABLED' that can be used when some 3575 functionality is disabled at compile-time. 3576 [Douglas Stebila <douglas.stebila@sun.com>] 3577 3578 *) Change default behaviour of 'openssl asn1parse' so that more 3579 information is visible when viewing, e.g., a certificate: 3580 3581 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 3582 mode the content of non-printable OCTET STRINGs is output in a 3583 style similar to INTEGERs, but with '[HEX DUMP]' prepended to 3584 avoid the appearance of a printable string. 3585 [Nils Larsch <nla@trustcenter.de>] 3586 3587 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 3588 functions 3589 EC_GROUP_set_asn1_flag() 3590 EC_GROUP_get_asn1_flag() 3591 EC_GROUP_set_point_conversion_form() 3592 EC_GROUP_get_point_conversion_form() 3593 These control ASN1 encoding details: 3594 - Curves (i.e., groups) are encoded explicitly unless asn1_flag 3595 has been set to OPENSSL_EC_NAMED_CURVE. 3596 - Points are encoded in uncompressed form by default; options for 3597 asn1_for are as for point2oct, namely 3598 POINT_CONVERSION_COMPRESSED 3599 POINT_CONVERSION_UNCOMPRESSED 3600 POINT_CONVERSION_HYBRID 3601 3602 Also add 'seed' and 'seed_len' members to EC_GROUP with access 3603 functions 3604 EC_GROUP_set_seed() 3605 EC_GROUP_get0_seed() 3606 EC_GROUP_get_seed_len() 3607 This is used only for ASN1 purposes (so far). 3608 [Nils Larsch <nla@trustcenter.de>] 3609 3610 *) Add 'field_type' member to EC_METHOD, which holds the NID 3611 of the appropriate field type OID. The new function 3612 EC_METHOD_get_field_type() returns this value. 3613 [Nils Larsch <nla@trustcenter.de>] 3614 3615 *) Add functions 3616 EC_POINT_point2bn() 3617 EC_POINT_bn2point() 3618 EC_POINT_point2hex() 3619 EC_POINT_hex2point() 3620 providing useful interfaces to EC_POINT_point2oct() and 3621 EC_POINT_oct2point(). 3622 [Nils Larsch <nla@trustcenter.de>] 3623 3624 *) Change internals of the EC library so that the functions 3625 EC_GROUP_set_generator() 3626 EC_GROUP_get_generator() 3627 EC_GROUP_get_order() 3628 EC_GROUP_get_cofactor() 3629 are implemented directly in crypto/ec/ec_lib.c and not dispatched 3630 to methods, which would lead to unnecessary code duplication when 3631 adding different types of curves. 3632 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 3633 3634 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 3635 arithmetic, and such that modified wNAFs are generated 3636 (which avoid length expansion in many cases). 3637 [Bodo Moeller] 3638 3639 *) Add a function EC_GROUP_check_discriminant() (defined via 3640 EC_METHOD) that verifies that the curve discriminant is non-zero. 3641 3642 Add a function EC_GROUP_check() that makes some sanity tests 3643 on a EC_GROUP, its generator and order. This includes 3644 EC_GROUP_check_discriminant(). 3645 [Nils Larsch <nla@trustcenter.de>] 3646 3647 *) Add ECDSA in new directory crypto/ecdsa/. 3648 3649 Add applications 'openssl ecparam' and 'openssl ecdsa' 3650 (these are based on 'openssl dsaparam' and 'openssl dsa'). 3651 3652 ECDSA support is also included in various other files across the 3653 library. Most notably, 3654 - 'openssl req' now has a '-newkey ecdsa:file' option; 3655 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 3656 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 3657 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 3658 them suitable for ECDSA where domain parameters must be 3659 extracted before the specific public key; 3660 - ECDSA engine support has been added. 3661 [Nils Larsch <nla@trustcenter.de>] 3662 3663 *) Include some named elliptic curves, and add OIDs from X9.62, 3664 SECG, and WAP/WTLS. Each curve can be obtained from the new 3665 function 3666 EC_GROUP_new_by_curve_name(), 3667 and the list of available named curves can be obtained with 3668 EC_get_builtin_curves(). 3669 Also add a 'curve_name' member to EC_GROUP objects, which can be 3670 accessed via 3671 EC_GROUP_set_curve_name() 3672 EC_GROUP_get_curve_name() 3673 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 3674 3675 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 3676 was actually never needed) and in BN_mul(). The removal in BN_mul() 3677 required a small change in bn_mul_part_recursive() and the addition 3678 of the functions bn_cmp_part_words(), bn_sub_part_words() and 3679 bn_add_part_words(), which do the same thing as bn_cmp_words(), 3680 bn_sub_words() and bn_add_words() except they take arrays with 3681 differing sizes. 3682 [Richard Levitte] 3683 3684 Changes between 0.9.7l and 0.9.7m [23 Feb 2007] 3685 3686 *) Cleanse PEM buffers before freeing them since they may contain 3687 sensitive data. 3688 [Benjamin Bennett <ben@psc.edu>] 3689 3690 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 3691 a ciphersuite string such as "DEFAULT:RSA" cannot enable 3692 authentication-only ciphersuites. 3693 [Bodo Moeller] 3694 3695 *) Since AES128 and AES256 share a single mask bit in the logic of 3696 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 3697 kludge to work properly if AES128 is available and AES256 isn't. 3698 [Victor Duchovni] 3699 3700 *) Expand security boundary to match 1.1.1 module. 3701 [Steve Henson] 3702 3703 *) Remove redundant features: hash file source, editing of test vectors 3704 modify fipsld to use external fips_premain.c signature. 3705 [Steve Henson] 3706 3707 *) New perl script mkfipsscr.pl to create shell scripts or batch files to 3708 run algorithm test programs. 3709 [Steve Henson] 3710 3711 *) Make algorithm test programs more tolerant of whitespace. 3712 [Steve Henson] 3713 3714 *) Have SSL/TLS server implementation tolerate "mismatched" record 3715 protocol version while receiving ClientHello even if the 3716 ClientHello is fragmented. (The server can't insist on the 3717 particular protocol version it has chosen before the ServerHello 3718 message has informed the client about his choice.) 3719 [Bodo Moeller] 3720 3721 *) Load error codes if they are not already present instead of using a 3722 static variable. This allows them to be cleanly unloaded and reloaded. 3723 [Steve Henson] 3724 3725 Changes between 0.9.7k and 0.9.7l [28 Sep 2006] 3726 3727 *) Introduce limits to prevent malicious keys being able to 3728 cause a denial of service. (CVE-2006-2940) 3729 [Steve Henson, Bodo Moeller] 3730 3731 *) Fix ASN.1 parsing of certain invalid structures that can result 3732 in a denial of service. (CVE-2006-2937) [Steve Henson] 3733 3734 *) Fix buffer overflow in SSL_get_shared_ciphers() function. 3735 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 3736 3737 *) Fix SSL client code which could crash if connecting to a 3738 malicious SSLv2 server. (CVE-2006-4343) 3739 [Tavis Ormandy and Will Drewry, Google Security Team] 3740 3741 *) Change ciphersuite string processing so that an explicit 3742 ciphersuite selects this one ciphersuite (so that "AES256-SHA" 3743 will no longer include "AES128-SHA"), and any other similar 3744 ciphersuite (same bitmap) from *other* protocol versions (so that 3745 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 3746 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 3747 changes from 0.9.8b and 0.9.8d. 3748 [Bodo Moeller] 3749 3750 Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 3751 3752 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 3753 (CVE-2006-4339) [Ben Laurie and Google Security Team] 3754 3755 *) Change the Unix randomness entropy gathering to use poll() when 3756 possible instead of select(), since the latter has some 3757 undesirable limitations. 3758 [Darryl Miles via Richard Levitte and Bodo Moeller] 3759 3760 *) Disable rogue ciphersuites: 3761 3762 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 3763 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 3764 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 3765 3766 The latter two were purportedly from 3767 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 3768 appear there. 3769 3770 Also deactive the remaining ciphersuites from 3771 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 3772 unofficial, and the ID has long expired. 3773 [Bodo Moeller] 3774 3775 *) Fix RSA blinding Heisenbug (problems sometimes occured on 3776 dual-core machines) and other potential thread-safety issues. 3777 [Bodo Moeller] 3778 3779 Changes between 0.9.7i and 0.9.7j [04 May 2006] 3780 3781 *) Adapt fipsld and the build system to link against the validated FIPS 3782 module in FIPS mode. 3783 [Steve Henson] 3784 3785 *) Fixes for VC++ 2005 build under Windows. 3786 [Steve Henson] 3787 3788 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 3789 from a Windows bash shell such as MSYS. It is autodetected from the 3790 "config" script when run from a VC++ environment. Modify standard VC++ 3791 build to use fipscanister.o from the GNU make build. 3792 [Steve Henson] 3793 3794 Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 3795 3796 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 3797 The value now differs depending on if you build for FIPS or not. 3798 BEWARE! A program linked with a shared FIPSed libcrypto can't be 3799 safely run with a non-FIPSed libcrypto, as it may crash because of 3800 the difference induced by this change. 3801 [Andy Polyakov] 3802 3803 Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 3804 3805 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 3806 (part of SSL_OP_ALL). This option used to disable the 3807 countermeasure against man-in-the-middle protocol-version 3808 rollback in the SSL 2.0 server implementation, which is a bad 3809 idea. (CVE-2005-2969) 3810 3811 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 3812 for Information Security, National Institute of Advanced Industrial 3813 Science and Technology [AIST], Japan)] 3814 3815 *) Minimal support for X9.31 signatures and PSS padding modes. This is 3816 mainly for FIPS compliance and not fully integrated at this stage. 3817 [Steve Henson] 3818 3819 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 3820 the exponentiation using a fixed-length exponent. (Otherwise, 3821 the information leaked through timing could expose the secret key 3822 after many signatures; cf. Bleichenbacher's attack on DSA with 3823 biased k.) 3824 [Bodo Moeller] 3825 3826 *) Make a new fixed-window mod_exp implementation the default for 3827 RSA, DSA, and DH private-key operations so that the sequence of 3828 squares and multiplies and the memory access pattern are 3829 independent of the particular secret key. This will mitigate 3830 cache-timing and potential related attacks. 3831 3832 BN_mod_exp_mont_consttime() is the new exponentiation implementation, 3833 and this is automatically used by BN_mod_exp_mont() if the new flag 3834 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 3835 will use this BN flag for private exponents unless the flag 3836 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 3837 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 3838 3839 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 3840 3841 *) Change the client implementation for SSLv23_method() and 3842 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 3843 Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 3844 (Previously, the SSL 2.0 backwards compatible Client Hello 3845 message format would be used even with SSL_OP_NO_SSLv2.) 3846 [Bodo Moeller] 3847 3848 *) Add support for smime-type MIME parameter in S/MIME messages which some 3849 clients need. 3850 [Steve Henson] 3851 3852 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 3853 a threadsafe manner. Modify rsa code to use new function and add calls 3854 to dsa and dh code (which had race conditions before). 3855 [Steve Henson] 3856 3857 *) Include the fixed error library code in the C error file definitions 3858 instead of fixing them up at runtime. This keeps the error code 3859 structures constant. 3860 [Steve Henson] 3861 3862 Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 3863 3864 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 3865 OpenSSL 0.9.8.] 3866 3867 *) Fixes for newer kerberos headers. NB: the casts are needed because 3868 the 'length' field is signed on one version and unsigned on another 3869 with no (?) obvious way to tell the difference, without these VC++ 3870 complains. Also the "definition" of FAR (blank) is no longer included 3871 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 3872 some needed definitions. 3873 [Steve Henson] 3874 3875 *) Undo Cygwin change. 3876 [Ulf M��ller] 3877 3878 *) Added support for proxy certificates according to RFC 3820. 3879 Because they may be a security thread to unaware applications, 3880 they must be explicitely allowed in run-time. See 3881 docs/HOWTO/proxy_certificates.txt for further information. 3882 [Richard Levitte] 3883 3884 Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 3885 3886 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 3887 server and client random values. Previously 3888 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 3889 less random data when sizeof(time_t) > 4 (some 64 bit platforms). 3890 3891 This change has negligible security impact because: 3892 3893 1. Server and client random values still have 24 bytes of pseudo random 3894 data. 3895 3896 2. Server and client random values are sent in the clear in the initial 3897 handshake. 3898 3899 3. The master secret is derived using the premaster secret (48 bytes in 3900 size for static RSA ciphersuites) as well as client server and random 3901 values. 3902 3903 The OpenSSL team would like to thank the UK NISCC for bringing this issue 3904 to our attention. 3905 3906 [Stephen Henson, reported by UK NISCC] 3907 3908 *) Use Windows randomness collection on Cygwin. 3909 [Ulf M��ller] 3910 3911 *) Fix hang in EGD/PRNGD query when communication socket is closed 3912 prematurely by EGD/PRNGD. 3913 [Darren Tucker <dtucker@zip.com.au> via Lutz J��nicke, resolves #1014] 3914 3915 *) Prompt for pass phrases when appropriate for PKCS12 input format. 3916 [Steve Henson] 3917 3918 *) Back-port of selected performance improvements from development 3919 branch, as well as improved support for PowerPC platforms. 3920 [Andy Polyakov] 3921 3922 *) Add lots of checks for memory allocation failure, error codes to indicate 3923 failure and freeing up memory if a failure occurs. 3924 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 3925 3926 *) Add new -passin argument to dgst. 3927 [Steve Henson] 3928 3929 *) Perform some character comparisons of different types in X509_NAME_cmp: 3930 this is needed for some certificates that reencode DNs into UTF8Strings 3931 (in violation of RFC3280) and can't or wont issue name rollover 3932 certificates. 3933 [Steve Henson] 3934 3935 *) Make an explicit check during certificate validation to see that 3936 the CA setting in each certificate on the chain is correct. As a 3937 side effect always do the following basic checks on extensions, 3938 not just when there's an associated purpose to the check: 3939 3940 - if there is an unhandled critical extension (unless the user 3941 has chosen to ignore this fault) 3942 - if the path length has been exceeded (if one is set at all) 3943 - that certain extensions fit the associated purpose (if one has 3944 been given) 3945 [Richard Levitte] 3946 3947 Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 3948 3949 *) Avoid a race condition when CRLs are checked in a multi threaded 3950 environment. This would happen due to the reordering of the revoked 3951 entries during signature checking and serial number lookup. Now the 3952 encoding is cached and the serial number sort performed under a lock. 3953 Add new STACK function sk_is_sorted(). 3954 [Steve Henson] 3955 3956 *) Add Delta CRL to the extension code. 3957 [Steve Henson] 3958 3959 *) Various fixes to s3_pkt.c so alerts are sent properly. 3960 [David Holmes <d.holmes@f5.com>] 3961 3962 *) Reduce the chances of duplicate issuer name and serial numbers (in 3963 violation of RFC3280) using the OpenSSL certificate creation utilities. 3964 This is done by creating a random 64 bit value for the initial serial 3965 number when a serial number file is created or when a self signed 3966 certificate is created using 'openssl req -x509'. The initial serial 3967 number file is created using 'openssl x509 -next_serial' in CA.pl 3968 rather than being initialized to 1. 3969 [Steve Henson] 3970 3971 Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 3972 3973 *) Fix null-pointer assignment in do_change_cipher_spec() revealed 3974 by using the Codenomicon TLS Test Tool (CVE-2004-0079) 3975 [Joe Orton, Steve Henson] 3976 3977 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 3978 (CVE-2004-0112) 3979 [Joe Orton, Steve Henson] 3980 3981 *) Make it possible to have multiple active certificates with the same 3982 subject in the CA index file. This is done only if the keyword 3983 'unique_subject' is set to 'no' in the main CA section (default 3984 if 'CA_default') of the configuration file. The value is saved 3985 with the database itself in a separate index attribute file, 3986 named like the index file with '.attr' appended to the name. 3987 [Richard Levitte] 3988 3989 *) X509 verify fixes. Disable broken certificate workarounds when 3990 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 3991 keyUsage extension present. Don't accept CRLs with unhandled critical 3992 extensions: since verify currently doesn't process CRL extensions this 3993 rejects a CRL with *any* critical extensions. Add new verify error codes 3994 for these cases. 3995 [Steve Henson] 3996 3997 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 3998 A clarification of RFC2560 will require the use of OCTET STRINGs and 3999 some implementations cannot handle the current raw format. Since OpenSSL 4000 copies and compares OCSP nonces as opaque blobs without any attempt at 4001 parsing them this should not create any compatibility issues. 4002 [Steve Henson] 4003 4004 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 4005 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 4006 this HMAC (and other) operations are several times slower than OpenSSL 4007 < 0.9.7. 4008 [Steve Henson] 4009 4010 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 4011 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 4012 4013 *) Use the correct content when signing type "other". 4014 [Steve Henson] 4015 4016 Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 4017 4018 *) Fix various bugs revealed by running the NISCC test suite: 4019 4020 Stop out of bounds reads in the ASN1 code when presented with 4021 invalid tags (CVE-2003-0543 and CVE-2003-0544). 4022 4023 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 4024 4025 If verify callback ignores invalid public key errors don't try to check 4026 certificate signature with the NULL public key. 4027 4028 [Steve Henson] 4029 4030 *) New -ignore_err option in ocsp application to stop the server 4031 exiting on the first error in a request. 4032 [Steve Henson] 4033 4034 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 4035 if the server requested one: as stated in TLS 1.0 and SSL 3.0 4036 specifications. 4037 [Steve Henson] 4038 4039 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 4040 extra data after the compression methods not only for TLS 1.0 4041 but also for SSL 3.0 (as required by the specification). 4042 [Bodo Moeller; problem pointed out by Matthias Loepfe] 4043 4044 *) Change X509_certificate_type() to mark the key as exported/exportable 4045 when it's 512 *bits* long, not 512 bytes. 4046 [Richard Levitte] 4047 4048 *) Change AES_cbc_encrypt() so it outputs exact multiple of 4049 blocks during encryption. 4050 [Richard Levitte] 4051 4052 *) Various fixes to base64 BIO and non blocking I/O. On write 4053 flushes were not handled properly if the BIO retried. On read 4054 data was not being buffered properly and had various logic bugs. 4055 This also affects blocking I/O when the data being decoded is a 4056 certain size. 4057 [Steve Henson] 4058 4059 *) Various S/MIME bugfixes and compatibility changes: 4060 output correct application/pkcs7 MIME type if 4061 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 4062 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 4063 of files as .eml work). Correctly handle very long lines in MIME 4064 parser. 4065 [Steve Henson] 4066 4067 Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 4068 4069 *) Countermeasure against the Klima-Pokorny-Rosa extension of 4070 Bleichbacher's attack on PKCS #1 v1.5 padding: treat 4071 a protocol version number mismatch like a decryption error 4072 in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 4073 [Bodo Moeller] 4074 4075 *) Turn on RSA blinding by default in the default implementation 4076 to avoid a timing attack. Applications that don't want it can call 4077 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 4078 They would be ill-advised to do so in most cases. 4079 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 4080 4081 *) Change RSA blinding code so that it works when the PRNG is not 4082 seeded (in this case, the secret RSA exponent is abused as 4083 an unpredictable seed -- if it is not unpredictable, there 4084 is no point in blinding anyway). Make RSA blinding thread-safe 4085 by remembering the creator's thread ID in rsa->blinding and 4086 having all other threads use local one-time blinding factors 4087 (this requires more computation than sharing rsa->blinding, but 4088 avoids excessive locking; and if an RSA object is not shared 4089 between threads, blinding will still be very fast). 4090 [Bodo Moeller] 4091 4092 *) Fixed a typo bug that would cause ENGINE_set_default() to set an 4093 ENGINE as defaults for all supported algorithms irrespective of 4094 the 'flags' parameter. 'flags' is now honoured, so applications 4095 should make sure they are passing it correctly. 4096 [Geoff Thorpe] 4097 4098 *) Target "mingw" now allows native Windows code to be generated in 4099 the Cygwin environment as well as with the MinGW compiler. 4100 [Ulf Moeller] 4101 4102 Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 4103 4104 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 4105 via timing by performing a MAC computation even if incorrrect 4106 block cipher padding has been found. This is a countermeasure 4107 against active attacks where the attacker has to distinguish 4108 between bad padding and a MAC verification error. (CVE-2003-0078) 4109 4110 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 4111 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 4112 Martin Vuagnoux (EPFL, Ilion)] 4113 4114 *) Make the no-err option work as intended. The intention with no-err 4115 is not to have the whole error stack handling routines removed from 4116 libcrypto, it's only intended to remove all the function name and 4117 reason texts, thereby removing some of the footprint that may not 4118 be interesting if those errors aren't displayed anyway. 4119 4120 NOTE: it's still possible for any application or module to have it's 4121 own set of error texts inserted. The routines are there, just not 4122 used by default when no-err is given. 4123 [Richard Levitte] 4124 4125 *) Add support for FreeBSD on IA64. 4126 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 4127 4128 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 4129 Kerberos function mit_des_cbc_cksum(). Before this change, 4130 the value returned by DES_cbc_cksum() was like the one from 4131 mit_des_cbc_cksum(), except the bytes were swapped. 4132 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 4133 4134 *) Allow an application to disable the automatic SSL chain building. 4135 Before this a rather primitive chain build was always performed in 4136 ssl3_output_cert_chain(): an application had no way to send the 4137 correct chain if the automatic operation produced an incorrect result. 4138 4139 Now the chain builder is disabled if either: 4140 4141 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 4142 4143 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 4144 4145 The reasoning behind this is that an application would not want the 4146 auto chain building to take place if extra chain certificates are 4147 present and it might also want a means of sending no additional 4148 certificates (for example the chain has two certificates and the 4149 root is omitted). 4150 [Steve Henson] 4151 4152 *) Add the possibility to build without the ENGINE framework. 4153 [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 4154 4155 *) Under Win32 gmtime() can return NULL: check return value in 4156 OPENSSL_gmtime(). Add error code for case where gmtime() fails. 4157 [Steve Henson] 4158 4159 *) DSA routines: under certain error conditions uninitialized BN objects 4160 could be freed. Solution: make sure initialization is performed early 4161 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 4162 Nils Larsch <nla@trustcenter.de> via PR#459) 4163 [Lutz Jaenicke] 4164 4165 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 4166 checked on reconnect on the client side, therefore session resumption 4167 could still fail with a "ssl session id is different" error. This 4168 behaviour is masked when SSL_OP_ALL is used due to 4169 SSL_OP_MICROSOFT_SESS_ID_BUG being set. 4170 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4171 followup to PR #377. 4172 [Lutz Jaenicke] 4173 4174 *) IA-32 assembler support enhancements: unified ELF targets, support 4175 for SCO/Caldera platforms, fix for Cygwin shared build. 4176 [Andy Polyakov] 4177 4178 *) Add support for FreeBSD on sparc64. As a consequence, support for 4179 FreeBSD on non-x86 processors is separate from x86 processors on 4180 the config script, much like the NetBSD support. 4181 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 4182 4183 Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 4184 4185 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 4186 OpenSSL 0.9.7.] 4187 4188 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 4189 code (06) was taken as the first octet of the session ID and the last 4190 octet was ignored consequently. As a result SSLv2 client side session 4191 caching could not have worked due to the session ID mismatch between 4192 client and server. 4193 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4194 PR #377. 4195 [Lutz Jaenicke] 4196 4197 *) Change the declaration of needed Kerberos libraries to use EX_LIBS 4198 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 4199 removed entirely. 4200 [Richard Levitte] 4201 4202 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 4203 seems that in spite of existing for more than a year, many application 4204 author have done nothing to provide the necessary callbacks, which 4205 means that this particular engine will not work properly anywhere. 4206 This is a very unfortunate situation which forces us, in the name 4207 of usability, to give the hw_ncipher.c a static lock, which is part 4208 of libcrypto. 4209 NOTE: This is for the 0.9.7 series ONLY. This hack will never 4210 appear in 0.9.8 or later. We EXPECT application authors to have 4211 dealt properly with this when 0.9.8 is released (unless we actually 4212 make such changes in the libcrypto locking code that changes will 4213 have to be made anyway). 4214 [Richard Levitte] 4215 4216 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 4217 octets have been read, EOF or an error occurs. Without this change 4218 some truncated ASN1 structures will not produce an error. 4219 [Steve Henson] 4220 4221 *) Disable Heimdal support, since it hasn't been fully implemented. 4222 Still give the possibility to force the use of Heimdal, but with 4223 warnings and a request that patches get sent to openssl-dev. 4224 [Richard Levitte] 4225 4226 *) Add the VC-CE target, introduce the WINCE sysname, and add 4227 INSTALL.WCE and appropriate conditionals to make it build. 4228 [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 4229 4230 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 4231 cygssl-x.y.z.dll, where x, y and z are the major, minor and 4232 edit numbers of the version. 4233 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 4234 4235 *) Introduce safe string copy and catenation functions 4236 (BUF_strlcpy() and BUF_strlcat()). 4237 [Ben Laurie (CHATS) and Richard Levitte] 4238 4239 *) Avoid using fixed-size buffers for one-line DNs. 4240 [Ben Laurie (CHATS)] 4241 4242 *) Add BUF_MEM_grow_clean() to avoid information leakage when 4243 resizing buffers containing secrets, and use where appropriate. 4244 [Ben Laurie (CHATS)] 4245 4246 *) Avoid using fixed size buffers for configuration file location. 4247 [Ben Laurie (CHATS)] 4248 4249 *) Avoid filename truncation for various CA files. 4250 [Ben Laurie (CHATS)] 4251 4252 *) Use sizeof in preference to magic numbers. 4253 [Ben Laurie (CHATS)] 4254 4255 *) Avoid filename truncation in cert requests. 4256 [Ben Laurie (CHATS)] 4257 4258 *) Add assertions to check for (supposedly impossible) buffer 4259 overflows. 4260 [Ben Laurie (CHATS)] 4261 4262 *) Don't cache truncated DNS entries in the local cache (this could 4263 potentially lead to a spoofing attack). 4264 [Ben Laurie (CHATS)] 4265 4266 *) Fix various buffers to be large enough for hex/decimal 4267 representations in a platform independent manner. 4268 [Ben Laurie (CHATS)] 4269 4270 *) Add CRYPTO_realloc_clean() to avoid information leakage when 4271 resizing buffers containing secrets, and use where appropriate. 4272 [Ben Laurie (CHATS)] 4273 4274 *) Add BIO_indent() to avoid much slightly worrying code to do 4275 indents. 4276 [Ben Laurie (CHATS)] 4277 4278 *) Convert sprintf()/BIO_puts() to BIO_printf(). 4279 [Ben Laurie (CHATS)] 4280 4281 *) buffer_gets() could terminate with the buffer only half 4282 full. Fixed. 4283 [Ben Laurie (CHATS)] 4284 4285 *) Add assertions to prevent user-supplied crypto functions from 4286 overflowing internal buffers by having large block sizes, etc. 4287 [Ben Laurie (CHATS)] 4288 4289 *) New OPENSSL_assert() macro (similar to assert(), but enabled 4290 unconditionally). 4291 [Ben Laurie (CHATS)] 4292 4293 *) Eliminate unused copy of key in RC4. 4294 [Ben Laurie (CHATS)] 4295 4296 *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 4297 [Ben Laurie (CHATS)] 4298 4299 *) Fix off-by-one error in EGD path. 4300 [Ben Laurie (CHATS)] 4301 4302 *) If RANDFILE path is too long, ignore instead of truncating. 4303 [Ben Laurie (CHATS)] 4304 4305 *) Eliminate unused and incorrectly sized X.509 structure 4306 CBCParameter. 4307 [Ben Laurie (CHATS)] 4308 4309 *) Eliminate unused and dangerous function knumber(). 4310 [Ben Laurie (CHATS)] 4311 4312 *) Eliminate unused and dangerous structure, KSSL_ERR. 4313 [Ben Laurie (CHATS)] 4314 4315 *) Protect against overlong session ID context length in an encoded 4316 session object. Since these are local, this does not appear to be 4317 exploitable. 4318 [Ben Laurie (CHATS)] 4319 4320 *) Change from security patch (see 0.9.6e below) that did not affect 4321 the 0.9.6 release series: 4322 4323 Remote buffer overflow in SSL3 protocol - an attacker could 4324 supply an oversized master key in Kerberos-enabled versions. 4325 (CVE-2002-0657) 4326 [Ben Laurie (CHATS)] 4327 4328 *) Change the SSL kerb5 codes to match RFC 2712. 4329 [Richard Levitte] 4330 4331 *) Make -nameopt work fully for req and add -reqopt switch. 4332 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 4333 4334 *) The "block size" for block ciphers in CFB and OFB mode should be 1. 4335 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 4336 4337 *) Make sure tests can be performed even if the corresponding algorithms 4338 have been removed entirely. This was also the last step to make 4339 OpenSSL compilable with DJGPP under all reasonable conditions. 4340 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 4341 4342 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 4343 to allow version independent disabling of normally unselected ciphers, 4344 which may be activated as a side-effect of selecting a single cipher. 4345 4346 (E.g., cipher list string "RSA" enables ciphersuites that are left 4347 out of "ALL" because they do not provide symmetric encryption. 4348 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 4349 [Lutz Jaenicke, Bodo Moeller] 4350 4351 *) Add appropriate support for separate platform-dependent build 4352 directories. The recommended way to make a platform-dependent 4353 build directory is the following (tested on Linux), maybe with 4354 some local tweaks: 4355 4356 # Place yourself outside of the OpenSSL source tree. In 4357 # this example, the environment variable OPENSSL_SOURCE 4358 # is assumed to contain the absolute OpenSSL source directory. 4359 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 4360 cd objtree/"`uname -s`-`uname -r`-`uname -m`" 4361 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 4362 mkdir -p `dirname $F` 4363 ln -s $OPENSSL_SOURCE/$F $F 4364 done 4365 4366 To be absolutely sure not to disturb the source tree, a "make clean" 4367 is a good thing. If it isn't successfull, don't worry about it, 4368 it probably means the source directory is very clean. 4369 [Richard Levitte] 4370 4371 *) Make sure any ENGINE control commands make local copies of string 4372 pointers passed to them whenever necessary. Otherwise it is possible 4373 the caller may have overwritten (or deallocated) the original string 4374 data when a later ENGINE operation tries to use the stored values. 4375 [G��tz Babin-Ebell <babinebell@trustcenter.de>] 4376 4377 *) Improve diagnostics in file reading and command-line digests. 4378 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 4379 4380 *) Add AES modes CFB and OFB to the object database. Correct an 4381 error in AES-CFB decryption. 4382 [Richard Levitte] 4383 4384 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 4385 allows existing EVP_CIPHER_CTX structures to be reused after 4386 calling EVP_*Final(). This behaviour is used by encryption 4387 BIOs and some applications. This has the side effect that 4388 applications must explicitly clean up cipher contexts with 4389 EVP_CIPHER_CTX_cleanup() or they will leak memory. 4390 [Steve Henson] 4391 4392 *) Check the values of dna and dnb in bn_mul_recursive before calling 4393 bn_mul_comba (a non zero value means the a or b arrays do not contain 4394 n2 elements) and fallback to bn_mul_normal if either is not zero. 4395 [Steve Henson] 4396 4397 *) Fix escaping of non-ASCII characters when using the -subj option 4398 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 4399 [Lutz Jaenicke] 4400 4401 *) Make object definitions compliant to LDAP (RFC2256): SN is the short 4402 form for "surname", serialNumber has no short form. 4403 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 4404 therefore remove "mail" short name for "internet 7". 4405 The OID for unique identifiers in X509 certificates is 4406 x500UniqueIdentifier, not uniqueIdentifier. 4407 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 4408 [Lutz Jaenicke] 4409 4410 *) Add an "init" command to the ENGINE config module and auto initialize 4411 ENGINEs. Without any "init" command the ENGINE will be initialized 4412 after all ctrl commands have been executed on it. If init=1 the 4413 ENGINE is initailized at that point (ctrls before that point are run 4414 on the uninitialized ENGINE and after on the initialized one). If 4415 init=0 then the ENGINE will not be iniatialized at all. 4416 [Steve Henson] 4417 4418 *) Fix the 'app_verify_callback' interface so that the user-defined 4419 argument is actually passed to the callback: In the 4420 SSL_CTX_set_cert_verify_callback() prototype, the callback 4421 declaration has been changed from 4422 int (*cb)() 4423 into 4424 int (*cb)(X509_STORE_CTX *,void *); 4425 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 4426 i=s->ctx->app_verify_callback(&ctx) 4427 has been changed into 4428 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 4429 4430 To update applications using SSL_CTX_set_cert_verify_callback(), 4431 a dummy argument can be added to their callback functions. 4432 [D. K. Smetters <smetters@parc.xerox.com>] 4433 4434 *) Added the '4758cca' ENGINE to support IBM 4758 cards. 4435 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 4436 4437 *) Add and OPENSSL_LOAD_CONF define which will cause 4438 OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 4439 This allows older applications to transparently support certain 4440 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 4441 Two new functions OPENSSL_add_all_algorithms_noconf() which will never 4442 load the config file and OPENSSL_add_all_algorithms_conf() which will 4443 always load it have also been added. 4444 [Steve Henson] 4445 4446 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 4447 Adjust NIDs and EVP layer. 4448 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4449 4450 *) Config modules support in openssl utility. 4451 4452 Most commands now load modules from the config file, 4453 though in a few (such as version) this isn't done 4454 because it couldn't be used for anything. 4455 4456 In the case of ca and req the config file used is 4457 the same as the utility itself: that is the -config 4458 command line option can be used to specify an 4459 alternative file. 4460 [Steve Henson] 4461 4462 *) Move default behaviour from OPENSSL_config(). If appname is NULL 4463 use "openssl_conf" if filename is NULL use default openssl config file. 4464 [Steve Henson] 4465 4466 *) Add an argument to OPENSSL_config() to allow the use of an alternative 4467 config section name. Add a new flag to tolerate a missing config file 4468 and move code to CONF_modules_load_file(). 4469 [Steve Henson] 4470 4471 *) Support for crypto accelerator cards from Accelerated Encryption 4472 Processing, www.aep.ie. (Use engine 'aep') 4473 The support was copied from 0.9.6c [engine] and adapted/corrected 4474 to work with the new engine framework. 4475 [AEP Inc. and Richard Levitte] 4476 4477 *) Support for SureWare crypto accelerator cards from Baltimore 4478 Technologies. (Use engine 'sureware') 4479 The support was copied from 0.9.6c [engine] and adapted 4480 to work with the new engine framework. 4481 [Richard Levitte] 4482 4483 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 4484 make the newer ENGINE framework commands for the CHIL engine work. 4485 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 4486 4487 *) Make it possible to produce shared libraries on ReliantUNIX. 4488 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 4489 4490 *) Add the configuration target debug-linux-ppro. 4491 Make 'openssl rsa' use the general key loading routines 4492 implemented in apps.c, and make those routines able to 4493 handle the key format FORMAT_NETSCAPE and the variant 4494 FORMAT_IISSGC. 4495 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4496 4497 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 4498 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4499 4500 *) Add -keyform to rsautl, and document -engine. 4501 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 4502 4503 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 4504 BIO_R_NO_SUCH_FILE error code rather than the generic 4505 ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 4506 [Ben Laurie] 4507 4508 *) Add new functions 4509 ERR_peek_last_error 4510 ERR_peek_last_error_line 4511 ERR_peek_last_error_line_data. 4512 These are similar to 4513 ERR_peek_error 4514 ERR_peek_error_line 4515 ERR_peek_error_line_data, 4516 but report on the latest error recorded rather than the first one 4517 still in the error queue. 4518 [Ben Laurie, Bodo Moeller] 4519 4520 *) default_algorithms option in ENGINE config module. This allows things 4521 like: 4522 default_algorithms = ALL 4523 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 4524 [Steve Henson] 4525 4526 *) Prelminary ENGINE config module. 4527 [Steve Henson] 4528 4529 *) New experimental application configuration code. 4530 [Steve Henson] 4531 4532 *) Change the AES code to follow the same name structure as all other 4533 symmetric ciphers, and behave the same way. Move everything to 4534 the directory crypto/aes, thereby obsoleting crypto/rijndael. 4535 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4536 4537 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 4538 [Ben Laurie and Theo de Raadt] 4539 4540 *) Add option to output public keys in req command. 4541 [Massimiliano Pala madwolf@openca.org] 4542 4543 *) Use wNAFs in EC_POINTs_mul() for improved efficiency 4544 (up to about 10% better than before for P-192 and P-224). 4545 [Bodo Moeller] 4546 4547 *) New functions/macros 4548 4549 SSL_CTX_set_msg_callback(ctx, cb) 4550 SSL_CTX_set_msg_callback_arg(ctx, arg) 4551 SSL_set_msg_callback(ssl, cb) 4552 SSL_set_msg_callback_arg(ssl, arg) 4553 4554 to request calling a callback function 4555 4556 void cb(int write_p, int version, int content_type, 4557 const void *buf, size_t len, SSL *ssl, void *arg) 4558 4559 whenever a protocol message has been completely received 4560 (write_p == 0) or sent (write_p == 1). Here 'version' is the 4561 protocol version according to which the SSL library interprets 4562 the current protocol message (SSL2_VERSION, SSL3_VERSION, or 4563 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 4564 the content type as defined in the SSL 3.0/TLS 1.0 protocol 4565 specification (change_cipher_spec(20), alert(21), handshake(22)). 4566 'buf' and 'len' point to the actual message, 'ssl' to the 4567 SSL object, and 'arg' is the application-defined value set by 4568 SSL[_CTX]_set_msg_callback_arg(). 4569 4570 'openssl s_client' and 'openssl s_server' have new '-msg' options 4571 to enable a callback that displays all protocol messages. 4572 [Bodo Moeller] 4573 4574 *) Change the shared library support so shared libraries are built as 4575 soon as the corresponding static library is finished, and thereby get 4576 openssl and the test programs linked against the shared library. 4577 This still only happens when the keyword "shard" has been given to 4578 the configuration scripts. 4579 4580 NOTE: shared library support is still an experimental thing, and 4581 backward binary compatibility is still not guaranteed. 4582 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 4583 4584 *) Add support for Subject Information Access extension. 4585 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 4586 4587 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 4588 additional bytes when new memory had to be allocated, not just 4589 when reusing an existing buffer. 4590 [Bodo Moeller] 4591 4592 *) New command line and configuration option 'utf8' for the req command. 4593 This allows field values to be specified as UTF8 strings. 4594 [Steve Henson] 4595 4596 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 4597 runs for the former and machine-readable output for the latter. 4598 [Ben Laurie] 4599 4600 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 4601 of the e-mail address in the DN (i.e., it will go into a certificate 4602 extension only). The new configuration file option 'email_in_dn = no' 4603 has the same effect. 4604 [Massimiliano Pala madwolf@openca.org] 4605 4606 *) Change all functions with names starting with des_ to be starting 4607 with DES_ instead. Add wrappers that are compatible with libdes, 4608 but are named _ossl_old_des_*. Finally, add macros that map the 4609 des_* symbols to the corresponding _ossl_old_des_* if libdes 4610 compatibility is desired. If OpenSSL 0.9.6c compatibility is 4611 desired, the des_* symbols will be mapped to DES_*, with one 4612 exception. 4613 4614 Since we provide two compatibility mappings, the user needs to 4615 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 4616 compatibility is desired. The default (i.e., when that macro 4617 isn't defined) is OpenSSL 0.9.6c compatibility. 4618 4619 There are also macros that enable and disable the support of old 4620 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 4621 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 4622 are defined, the default will apply: to support the old des routines. 4623 4624 In either case, one must include openssl/des.h to get the correct 4625 definitions. Do not try to just include openssl/des_old.h, that 4626 won't work. 4627 4628 NOTE: This is a major break of an old API into a new one. Software 4629 authors are encouraged to switch to the DES_ style functions. Some 4630 time in the future, des_old.h and the libdes compatibility functions 4631 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 4632 default), and then completely removed. 4633 [Richard Levitte] 4634 4635 *) Test for certificates which contain unsupported critical extensions. 4636 If such a certificate is found during a verify operation it is 4637 rejected by default: this behaviour can be overridden by either 4638 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 4639 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 4640 X509_supported_extension() has also been added which returns 1 if a 4641 particular extension is supported. 4642 [Steve Henson] 4643 4644 *) Modify the behaviour of EVP cipher functions in similar way to digests 4645 to retain compatibility with existing code. 4646 [Steve Henson] 4647 4648 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 4649 compatibility with existing code. In particular the 'ctx' parameter does 4650 not have to be to be initialized before the call to EVP_DigestInit() and 4651 it is tidied up after a call to EVP_DigestFinal(). New function 4652 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 4653 EVP_MD_CTX_copy() changed to not require the destination to be 4654 initialized valid and new function EVP_MD_CTX_copy_ex() added which 4655 requires the destination to be valid. 4656 4657 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 4658 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 4659 [Steve Henson] 4660 4661 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 4662 so that complete 'Handshake' protocol structures are kept in memory 4663 instead of overwriting 'msg_type' and 'length' with 'body' data. 4664 [Bodo Moeller] 4665 4666 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 4667 [Massimo Santin via Richard Levitte] 4668 4669 *) Major restructuring to the underlying ENGINE code. This includes 4670 reduction of linker bloat, separation of pure "ENGINE" manipulation 4671 (initialisation, etc) from functionality dealing with implementations 4672 of specific crypto iterfaces. This change also introduces integrated 4673 support for symmetric ciphers and digest implementations - so ENGINEs 4674 can now accelerate these by providing EVP_CIPHER and EVP_MD 4675 implementations of their own. This is detailed in crypto/engine/README 4676 as it couldn't be adequately described here. However, there are a few 4677 API changes worth noting - some RSA, DSA, DH, and RAND functions that 4678 were changed in the original introduction of ENGINE code have now 4679 reverted back - the hooking from this code to ENGINE is now a good 4680 deal more passive and at run-time, operations deal directly with 4681 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 4682 dereferencing through an ENGINE pointer any more. Also, the ENGINE 4683 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 4684 they were not being used by the framework as there is no concept of a 4685 BIGNUM_METHOD and they could not be generalised to the new 4686 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 4687 ENGINE_cpy() has been removed as it cannot be consistently defined in 4688 the new code. 4689 [Geoff Thorpe] 4690 4691 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 4692 [Steve Henson] 4693 4694 *) Change mkdef.pl to sort symbols that get the same entry number, 4695 and make sure the automatically generated functions ERR_load_* 4696 become part of libeay.num as well. 4697 [Richard Levitte] 4698 4699 *) New function SSL_renegotiate_pending(). This returns true once 4700 renegotiation has been requested (either SSL_renegotiate() call 4701 or HelloRequest/ClientHello receveived from the peer) and becomes 4702 false once a handshake has been completed. 4703 (For servers, SSL_renegotiate() followed by SSL_do_handshake() 4704 sends a HelloRequest, but does not ensure that a handshake takes 4705 place. SSL_renegotiate_pending() is useful for checking if the 4706 client has followed the request.) 4707 [Bodo Moeller] 4708 4709 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 4710 By default, clients may request session resumption even during 4711 renegotiation (if session ID contexts permit); with this option, 4712 session resumption is possible only in the first handshake. 4713 4714 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 4715 more bits available for options that should not be part of 4716 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 4717 [Bodo Moeller] 4718 4719 *) Add some demos for certificate and certificate request creation. 4720 [Steve Henson] 4721 4722 *) Make maximum certificate chain size accepted from the peer application 4723 settable (SSL*_get/set_max_cert_list()), as proposed by 4724 "Douglas E. Engert" <deengert@anl.gov>. 4725 [Lutz Jaenicke] 4726 4727 *) Add support for shared libraries for Unixware-7 4728 (Boyd Lynn Gerber <gerberb@zenez.com>). 4729 [Lutz Jaenicke] 4730 4731 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 4732 be done prior to destruction. Use this to unload error strings from 4733 ENGINEs that load their own error strings. NB: This adds two new API 4734 functions to "get" and "set" this destroy handler in an ENGINE. 4735 [Geoff Thorpe] 4736 4737 *) Alter all existing ENGINE implementations (except "openssl" and 4738 "openbsd") to dynamically instantiate their own error strings. This 4739 makes them more flexible to be built both as statically-linked ENGINEs 4740 and self-contained shared-libraries loadable via the "dynamic" ENGINE. 4741 Also, add stub code to each that makes building them as self-contained 4742 shared-libraries easier (see README.ENGINE). 4743 [Geoff Thorpe] 4744 4745 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 4746 implementations into applications that are completely implemented in 4747 self-contained shared-libraries. The "dynamic" ENGINE exposes control 4748 commands that can be used to configure what shared-library to load and 4749 to control aspects of the way it is handled. Also, made an update to 4750 the README.ENGINE file that brings its information up-to-date and 4751 provides some information and instructions on the "dynamic" ENGINE 4752 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 4753 [Geoff Thorpe] 4754 4755 *) Make it possible to unload ranges of ERR strings with a new 4756 "ERR_unload_strings" function. 4757 [Geoff Thorpe] 4758 4759 *) Add a copy() function to EVP_MD. 4760 [Ben Laurie] 4761 4762 *) Make EVP_MD routines take a context pointer instead of just the 4763 md_data void pointer. 4764 [Ben Laurie] 4765 4766 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 4767 that the digest can only process a single chunk of data 4768 (typically because it is provided by a piece of 4769 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 4770 is only going to provide a single chunk of data, and hence the 4771 framework needn't accumulate the data for oneshot drivers. 4772 [Ben Laurie] 4773 4774 *) As with "ERR", make it possible to replace the underlying "ex_data" 4775 functions. This change also alters the storage and management of global 4776 ex_data state - it's now all inside ex_data.c and all "class" code (eg. 4777 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 4778 index counters. The API functions that use this state have been changed 4779 to take a "class_index" rather than pointers to the class's local STACK 4780 and counter, and there is now an API function to dynamically create new 4781 classes. This centralisation allows us to (a) plug a lot of the 4782 thread-safety problems that existed, and (b) makes it possible to clean 4783 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 4784 such data would previously have always leaked in application code and 4785 workarounds were in place to make the memory debugging turn a blind eye 4786 to it. Application code that doesn't use this new function will still 4787 leak as before, but their memory debugging output will announce it now 4788 rather than letting it slide. 4789 4790 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 4791 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 4792 has a return value to indicate success or failure. 4793 [Geoff Thorpe] 4794 4795 *) Make it possible to replace the underlying "ERR" functions such that the 4796 global state (2 LHASH tables and 2 locks) is only used by the "default" 4797 implementation. This change also adds two functions to "get" and "set" 4798 the implementation prior to it being automatically set the first time 4799 any other ERR function takes place. Ie. an application can call "get", 4800 pass the return value to a module it has just loaded, and that module 4801 can call its own "set" function using that value. This means the 4802 module's "ERR" operations will use (and modify) the error state in the 4803 application and not in its own statically linked copy of OpenSSL code. 4804 [Geoff Thorpe] 4805 4806 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 4807 reference counts. This performs normal REF_PRINT/REF_CHECK macros on 4808 the operation, and provides a more encapsulated way for external code 4809 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 4810 to use these functions rather than manually incrementing the counts. 4811 4812 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 4813 [Geoff Thorpe] 4814 4815 *) Add EVP test program. 4816 [Ben Laurie] 4817 4818 *) Add symmetric cipher support to ENGINE. Expect the API to change! 4819 [Ben Laurie] 4820 4821 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 4822 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 4823 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 4824 These allow a CRL to be built without having to access X509_CRL fields 4825 directly. Modify 'ca' application to use new functions. 4826 [Steve Henson] 4827 4828 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 4829 bug workarounds. Rollback attack detection is a security feature. 4830 The problem will only arise on OpenSSL servers when TLSv1 is not 4831 available (sslv3_server_method() or SSL_OP_NO_TLSv1). 4832 Software authors not wanting to support TLSv1 will have special reasons 4833 for their choice and can explicitly enable this option. 4834 [Bodo Moeller, Lutz Jaenicke] 4835 4836 *) Rationalise EVP so it can be extended: don't include a union of 4837 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 4838 (similar to those existing for EVP_CIPHER_CTX). 4839 Usage example: 4840 4841 EVP_MD_CTX md; 4842 4843 EVP_MD_CTX_init(&md); /* new function call */ 4844 EVP_DigestInit(&md, EVP_sha1()); 4845 EVP_DigestUpdate(&md, in, len); 4846 EVP_DigestFinal(&md, out, NULL); 4847 EVP_MD_CTX_cleanup(&md); /* new function call */ 4848 4849 [Ben Laurie] 4850 4851 *) Make DES key schedule conform to the usual scheme, as well as 4852 correcting its structure. This means that calls to DES functions 4853 now have to pass a pointer to a des_key_schedule instead of a 4854 plain des_key_schedule (which was actually always a pointer 4855 anyway): E.g., 4856 4857 des_key_schedule ks; 4858 4859 des_set_key_checked(..., &ks); 4860 des_ncbc_encrypt(..., &ks, ...); 4861 4862 (Note that a later change renames 'des_...' into 'DES_...'.) 4863 [Ben Laurie] 4864 4865 *) Initial reduction of linker bloat: the use of some functions, such as 4866 PEM causes large amounts of unused functions to be linked in due to 4867 poor organisation. For example pem_all.c contains every PEM function 4868 which has a knock on effect of linking in large amounts of (unused) 4869 ASN1 code. Grouping together similar functions and splitting unrelated 4870 functions prevents this. 4871 [Steve Henson] 4872 4873 *) Cleanup of EVP macros. 4874 [Ben Laurie] 4875 4876 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 4877 correct _ecb suffix. 4878 [Ben Laurie] 4879 4880 *) Add initial OCSP responder support to ocsp application. The 4881 revocation information is handled using the text based index 4882 use by the ca application. The responder can either handle 4883 requests generated internally, supplied in files (for example 4884 via a CGI script) or using an internal minimal server. 4885 [Steve Henson] 4886 4887 *) Add configuration choices to get zlib compression for TLS. 4888 [Richard Levitte] 4889 4890 *) Changes to Kerberos SSL for RFC 2712 compliance: 4891 1. Implemented real KerberosWrapper, instead of just using 4892 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 4893 2. Implemented optional authenticator field of KerberosWrapper. 4894 4895 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 4896 and authenticator structs; see crypto/krb5/. 4897 4898 Generalized Kerberos calls to support multiple Kerberos libraries. 4899 [Vern Staats <staatsvr@asc.hpc.mil>, 4900 Jeffrey Altman <jaltman@columbia.edu> 4901 via Richard Levitte] 4902 4903 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 4904 already does with RSA. testdsa.h now has 'priv_key/pub_key' 4905 values for each of the key sizes rather than having just 4906 parameters (and 'speed' generating keys each time). 4907 [Geoff Thorpe] 4908 4909 *) Speed up EVP routines. 4910 Before: 4911encrypt 4912type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 4913des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 4914des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 4915des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 4916decrypt 4917des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 4918des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 4919des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 4920 After: 4921encrypt 4922des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 4923decrypt 4924des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 4925 [Ben Laurie] 4926 4927 *) Added the OS2-EMX target. 4928 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 4929 4930 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 4931 to support NCONF routines in extension code. New function CONF_set_nconf() 4932 to allow functions which take an NCONF to also handle the old LHASH 4933 structure: this means that the old CONF compatible routines can be 4934 retained (in particular wrt extensions) without having to duplicate the 4935 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 4936 [Steve Henson] 4937 4938 *) Enhance the general user interface with mechanisms for inner control 4939 and with possibilities to have yes/no kind of prompts. 4940 [Richard Levitte] 4941 4942 *) Change all calls to low level digest routines in the library and 4943 applications to use EVP. Add missing calls to HMAC_cleanup() and 4944 don't assume HMAC_CTX can be copied using memcpy(). 4945 [Verdon Walker <VWalker@novell.com>, Steve Henson] 4946 4947 *) Add the possibility to control engines through control names but with 4948 arbitrary arguments instead of just a string. 4949 Change the key loaders to take a UI_METHOD instead of a callback 4950 function pointer. NOTE: this breaks binary compatibility with earlier 4951 versions of OpenSSL [engine]. 4952 Adapt the nCipher code for these new conditions and add a card insertion 4953 callback. 4954 [Richard Levitte] 4955 4956 *) Enhance the general user interface with mechanisms to better support 4957 dialog box interfaces, application-defined prompts, the possibility 4958 to use defaults (for example default passwords from somewhere else) 4959 and interrupts/cancellations. 4960 [Richard Levitte] 4961 4962 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 4963 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 4964 [Steve Henson] 4965 4966 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 4967 tidy up some unnecessarily weird code in 'sk_new()'). 4968 [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 4969 4970 *) Change the key loading routines for ENGINEs to use the same kind 4971 callback (pem_password_cb) as all other routines that need this 4972 kind of callback. 4973 [Richard Levitte] 4974 4975 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 4976 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 4977 than this minimum value is recommended. 4978 [Lutz Jaenicke] 4979 4980 *) New random seeder for OpenVMS, using the system process statistics 4981 that are easily reachable. 4982 [Richard Levitte] 4983 4984 *) Windows apparently can't transparently handle global 4985 variables defined in DLLs. Initialisations such as: 4986 4987 const ASN1_ITEM *it = &ASN1_INTEGER_it; 4988 4989 wont compile. This is used by the any applications that need to 4990 declare their own ASN1 modules. This was fixed by adding the option 4991 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 4992 needed for static libraries under Win32. 4993 [Steve Henson] 4994 4995 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 4996 setting of purpose and trust fields. New X509_STORE trust and 4997 purpose functions and tidy up setting in other SSL functions. 4998 [Steve Henson] 4999 5000 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 5001 structure. These are inherited by X509_STORE_CTX when it is 5002 initialised. This allows various defaults to be set in the 5003 X509_STORE structure (such as flags for CRL checking and custom 5004 purpose or trust settings) for functions which only use X509_STORE_CTX 5005 internally such as S/MIME. 5006 5007 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 5008 trust settings if they are not set in X509_STORE. This allows X509_STORE 5009 purposes and trust (in S/MIME for example) to override any set by default. 5010 5011 Add command line options for CRL checking to smime, s_client and s_server 5012 applications. 5013 [Steve Henson] 5014 5015 *) Initial CRL based revocation checking. If the CRL checking flag(s) 5016 are set then the CRL is looked up in the X509_STORE structure and 5017 its validity and signature checked, then if the certificate is found 5018 in the CRL the verify fails with a revoked error. 5019 5020 Various new CRL related callbacks added to X509_STORE_CTX structure. 5021 5022 Command line options added to 'verify' application to support this. 5023 5024 This needs some additional work, such as being able to handle multiple 5025 CRLs with different times, extension based lookup (rather than just 5026 by subject name) and ultimately more complete V2 CRL extension 5027 handling. 5028 [Steve Henson] 5029 5030 *) Add a general user interface API (crypto/ui/). This is designed 5031 to replace things like des_read_password and friends (backward 5032 compatibility functions using this new API are provided). 5033 The purpose is to remove prompting functions from the DES code 5034 section as well as provide for prompting through dialog boxes in 5035 a window system and the like. 5036 [Richard Levitte] 5037 5038 *) Add "ex_data" support to ENGINE so implementations can add state at a 5039 per-structure level rather than having to store it globally. 5040 [Geoff] 5041 5042 *) Make it possible for ENGINE structures to be copied when retrieved by 5043 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 5044 This causes the "original" ENGINE structure to act like a template, 5045 analogous to the RSA vs. RSA_METHOD type of separation. Because of this 5046 operational state can be localised to each ENGINE structure, despite the 5047 fact they all share the same "methods". New ENGINE structures returned in 5048 this case have no functional references and the return value is the single 5049 structural reference. This matches the single structural reference returned 5050 by ENGINE_by_id() normally, when it is incremented on the pre-existing 5051 ENGINE structure. 5052 [Geoff] 5053 5054 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 5055 needs to match any other type at all we need to manually clear the 5056 tag cache. 5057 [Steve Henson] 5058 5059 *) Changes to the "openssl engine" utility to include; 5060 - verbosity levels ('-v', '-vv', and '-vvv') that provide information 5061 about an ENGINE's available control commands. 5062 - executing control commands from command line arguments using the 5063 '-pre' and '-post' switches. '-post' is only used if '-t' is 5064 specified and the ENGINE is successfully initialised. The syntax for 5065 the individual commands are colon-separated, for example; 5066 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 5067 [Geoff] 5068 5069 *) New dynamic control command support for ENGINEs. ENGINEs can now 5070 declare their own commands (numbers), names (strings), descriptions, 5071 and input types for run-time discovery by calling applications. A 5072 subset of these commands are implicitly classed as "executable" 5073 depending on their input type, and only these can be invoked through 5074 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 5075 can be based on user input, config files, etc). The distinction is 5076 that "executable" commands cannot return anything other than a boolean 5077 result and can only support numeric or string input, whereas some 5078 discoverable commands may only be for direct use through 5079 ENGINE_ctrl(), eg. supporting the exchange of binary data, function 5080 pointers, or other custom uses. The "executable" commands are to 5081 support parameterisations of ENGINE behaviour that can be 5082 unambiguously defined by ENGINEs and used consistently across any 5083 OpenSSL-based application. Commands have been added to all the 5084 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 5085 control over shared-library paths without source code alterations. 5086 [Geoff] 5087 5088 *) Changed all ENGINE implementations to dynamically allocate their 5089 ENGINEs rather than declaring them statically. Apart from this being 5090 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 5091 this also allows the implementations to compile without using the 5092 internal engine_int.h header. 5093 [Geoff] 5094 5095 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 5096 'const' value. Any code that should be able to modify a RAND_METHOD 5097 should already have non-const pointers to it (ie. they should only 5098 modify their own ones). 5099 [Geoff] 5100 5101 *) Made a variety of little tweaks to the ENGINE code. 5102 - "atalla" and "ubsec" string definitions were moved from header files 5103 to C code. "nuron" string definitions were placed in variables 5104 rather than hard-coded - allowing parameterisation of these values 5105 later on via ctrl() commands. 5106 - Removed unused "#if 0"'d code. 5107 - Fixed engine list iteration code so it uses ENGINE_free() to release 5108 structural references. 5109 - Constified the RAND_METHOD element of ENGINE structures. 5110 - Constified various get/set functions as appropriate and added 5111 missing functions (including a catch-all ENGINE_cpy that duplicates 5112 all ENGINE values onto a new ENGINE except reference counts/state). 5113 - Removed NULL parameter checks in get/set functions. Setting a method 5114 or function to NULL is a way of cancelling out a previously set 5115 value. Passing a NULL ENGINE parameter is just plain stupid anyway 5116 and doesn't justify the extra error symbols and code. 5117 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 5118 flags from engine_int.h to engine.h. 5119 - Changed prototypes for ENGINE handler functions (init(), finish(), 5120 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 5121 [Geoff] 5122 5123 *) Implement binary inversion algorithm for BN_mod_inverse in addition 5124 to the algorithm using long division. The binary algorithm can be 5125 used only if the modulus is odd. On 32-bit systems, it is faster 5126 only for relatively small moduli (roughly 20-30% for 128-bit moduli, 5127 roughly 5-15% for 256-bit moduli), so we use it only for moduli 5128 up to 450 bits. In 64-bit environments, the binary algorithm 5129 appears to be advantageous for much longer moduli; here we use it 5130 for moduli up to 2048 bits. 5131 [Bodo Moeller] 5132 5133 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 5134 could not support the combine flag in choice fields. 5135 [Steve Henson] 5136 5137 *) Add a 'copy_extensions' option to the 'ca' utility. This copies 5138 extensions from a certificate request to the certificate. 5139 [Steve Henson] 5140 5141 *) Allow multiple 'certopt' and 'nameopt' options to be separated 5142 by commas. Add 'namopt' and 'certopt' options to the 'ca' config 5143 file: this allows the display of the certificate about to be 5144 signed to be customised, to allow certain fields to be included 5145 or excluded and extension details. The old system didn't display 5146 multicharacter strings properly, omitted fields not in the policy 5147 and couldn't display additional details such as extensions. 5148 [Steve Henson] 5149 5150 *) Function EC_POINTs_mul for multiple scalar multiplication 5151 of an arbitrary number of elliptic curve points 5152 \sum scalars[i]*points[i], 5153 optionally including the generator defined for the EC_GROUP: 5154 scalar*generator + \sum scalars[i]*points[i]. 5155 5156 EC_POINT_mul is a simple wrapper function for the typical case 5157 that the point list has just one item (besides the optional 5158 generator). 5159 [Bodo Moeller] 5160 5161 *) First EC_METHODs for curves over GF(p): 5162 5163 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 5164 operations and provides various method functions that can also 5165 operate with faster implementations of modular arithmetic. 5166 5167 EC_GFp_mont_method() reuses most functions that are part of 5168 EC_GFp_simple_method, but uses Montgomery arithmetic. 5169 5170 [Bodo Moeller; point addition and point doubling 5171 implementation directly derived from source code provided by 5172 Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 5173 5174 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 5175 crypto/ec/ec_lib.c): 5176 5177 Curves are EC_GROUP objects (with an optional group generator) 5178 based on EC_METHODs that are built into the library. 5179 5180 Points are EC_POINT objects based on EC_GROUP objects. 5181 5182 Most of the framework would be able to handle curves over arbitrary 5183 finite fields, but as there are no obvious types for fields other 5184 than GF(p), some functions are limited to that for now. 5185 [Bodo Moeller] 5186 5187 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 5188 that the file contains a complete HTTP response. 5189 [Richard Levitte] 5190 5191 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 5192 change the def and num file printf format specifier from "%-40sXXX" 5193 to "%-39s XXX". The latter will always guarantee a space after the 5194 field while the former will cause them to run together if the field 5195 is 40 of more characters long. 5196 [Steve Henson] 5197 5198 *) Constify the cipher and digest 'method' functions and structures 5199 and modify related functions to take constant EVP_MD and EVP_CIPHER 5200 pointers. 5201 [Steve Henson] 5202 5203 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 5204 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 5205 [Bodo Moeller] 5206 5207 *) Modify EVP_Digest*() routines so they now return values. Although the 5208 internal software routines can never fail additional hardware versions 5209 might. 5210 [Steve Henson] 5211 5212 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 5213 5214 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 5215 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 5216 5217 ASN1 error codes 5218 ERR_R_NESTED_ASN1_ERROR 5219 ... 5220 ERR_R_MISSING_ASN1_EOS 5221 were 4 .. 9, conflicting with 5222 ERR_LIB_RSA (= ERR_R_RSA_LIB) 5223 ... 5224 ERR_LIB_PEM (= ERR_R_PEM_LIB). 5225 They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 5226 5227 Add new error code 'ERR_R_INTERNAL_ERROR'. 5228 [Bodo Moeller] 5229 5230 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 5231 suffices. 5232 [Bodo Moeller] 5233 5234 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 5235 sets the subject name for a new request or supersedes the 5236 subject name in a given request. Formats that can be parsed are 5237 'CN=Some Name, OU=myOU, C=IT' 5238 and 5239 'CN=Some Name/OU=myOU/C=IT'. 5240 5241 Add options '-batch' and '-verbose' to 'openssl req'. 5242 [Massimiliano Pala <madwolf@hackmasters.net>] 5243 5244 *) Introduce the possibility to access global variables through 5245 functions on platform were that's the best way to handle exporting 5246 global variables in shared libraries. To enable this functionality, 5247 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 5248 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 5249 is normally done by Configure or something similar). 5250 5251 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 5252 in the source file (foo.c) like this: 5253 5254 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 5255 OPENSSL_IMPLEMENT_GLOBAL(double,bar); 5256 5257 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 5258 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 5259 5260 OPENSSL_DECLARE_GLOBAL(int,foo); 5261 #define foo OPENSSL_GLOBAL_REF(foo) 5262 OPENSSL_DECLARE_GLOBAL(double,bar); 5263 #define bar OPENSSL_GLOBAL_REF(bar) 5264 5265 The #defines are very important, and therefore so is including the 5266 header file everywhere where the defined globals are used. 5267 5268 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 5269 of ASN.1 items, but that structure is a bit different. 5270 5271 The largest change is in util/mkdef.pl which has been enhanced with 5272 better and easier to understand logic to choose which symbols should 5273 go into the Windows .def files as well as a number of fixes and code 5274 cleanup (among others, algorithm keywords are now sorted 5275 lexicographically to avoid constant rewrites). 5276 [Richard Levitte] 5277 5278 *) In BN_div() keep a copy of the sign of 'num' before writing the 5279 result to 'rm' because if rm==num the value will be overwritten 5280 and produce the wrong result if 'num' is negative: this caused 5281 problems with BN_mod() and BN_nnmod(). 5282 [Steve Henson] 5283 5284 *) Function OCSP_request_verify(). This checks the signature on an 5285 OCSP request and verifies the signer certificate. The signer 5286 certificate is just checked for a generic purpose and OCSP request 5287 trust settings. 5288 [Steve Henson] 5289 5290 *) Add OCSP_check_validity() function to check the validity of OCSP 5291 responses. OCSP responses are prepared in real time and may only 5292 be a few seconds old. Simply checking that the current time lies 5293 between thisUpdate and nextUpdate max reject otherwise valid responses 5294 caused by either OCSP responder or client clock inaccuracy. Instead 5295 we allow thisUpdate and nextUpdate to fall within a certain period of 5296 the current time. The age of the response can also optionally be 5297 checked. Two new options -validity_period and -status_age added to 5298 ocsp utility. 5299 [Steve Henson] 5300 5301 *) If signature or public key algorithm is unrecognized print out its 5302 OID rather that just UNKNOWN. 5303 [Steve Henson] 5304 5305 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 5306 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 5307 ID to be generated from the issuer certificate alone which can then be 5308 passed to OCSP_id_issuer_cmp(). 5309 [Steve Henson] 5310 5311 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 5312 ASN1 modules to export functions returning ASN1_ITEM pointers 5313 instead of the ASN1_ITEM structures themselves. This adds several 5314 new macros which allow the underlying ASN1 function/structure to 5315 be accessed transparently. As a result code should not use ASN1_ITEM 5316 references directly (such as &X509_it) but instead use the relevant 5317 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 5318 use of the new ASN1 code on platforms where exporting structures 5319 is problematical (for example in shared libraries) but exporting 5320 functions returning pointers to structures is not. 5321 [Steve Henson] 5322 5323 *) Add support for overriding the generation of SSL/TLS session IDs. 5324 These callbacks can be registered either in an SSL_CTX or per SSL. 5325 The purpose of this is to allow applications to control, if they wish, 5326 the arbitrary values chosen for use as session IDs, particularly as it 5327 can be useful for session caching in multiple-server environments. A 5328 command-line switch for testing this (and any client code that wishes 5329 to use such a feature) has been added to "s_server". 5330 [Geoff Thorpe, Lutz Jaenicke] 5331 5332 *) Modify mkdef.pl to recognise and parse preprocessor conditionals 5333 of the form '#if defined(...) || defined(...) || ...' and 5334 '#if !defined(...) && !defined(...) && ...'. This also avoids 5335 the growing number of special cases it was previously handling. 5336 [Richard Levitte] 5337 5338 *) Make all configuration macros available for application by making 5339 sure they are available in opensslconf.h, by giving them names starting 5340 with "OPENSSL_" to avoid conflicts with other packages and by making 5341 sure e_os2.h will cover all platform-specific cases together with 5342 opensslconf.h. 5343 Additionally, it is now possible to define configuration/platform- 5344 specific names (called "system identities"). In the C code, these 5345 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 5346 macro with the name beginning with "OPENSSL_SYS_", which is determined 5347 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 5348 what is available. 5349 [Richard Levitte] 5350 5351 *) New option -set_serial to 'req' and 'x509' this allows the serial 5352 number to use to be specified on the command line. Previously self 5353 signed certificates were hard coded with serial number 0 and the 5354 CA options of 'x509' had to use a serial number in a file which was 5355 auto incremented. 5356 [Steve Henson] 5357 5358 *) New options to 'ca' utility to support V2 CRL entry extensions. 5359 Currently CRL reason, invalidity date and hold instruction are 5360 supported. Add new CRL extensions to V3 code and some new objects. 5361 [Steve Henson] 5362 5363 *) New function EVP_CIPHER_CTX_set_padding() this is used to 5364 disable standard block padding (aka PKCS#5 padding) in the EVP 5365 API, which was previously mandatory. This means that the data is 5366 not padded in any way and so the total length much be a multiple 5367 of the block size, otherwise an error occurs. 5368 [Steve Henson] 5369 5370 *) Initial (incomplete) OCSP SSL support. 5371 [Steve Henson] 5372 5373 *) New function OCSP_parse_url(). This splits up a URL into its host, 5374 port and path components: primarily to parse OCSP URLs. New -url 5375 option to ocsp utility. 5376 [Steve Henson] 5377 5378 *) New nonce behavior. The return value of OCSP_check_nonce() now 5379 reflects the various checks performed. Applications can decide 5380 whether to tolerate certain situations such as an absent nonce 5381 in a response when one was present in a request: the ocsp application 5382 just prints out a warning. New function OCSP_add1_basic_nonce() 5383 this is to allow responders to include a nonce in a response even if 5384 the request is nonce-less. 5385 [Steve Henson] 5386 5387 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 5388 skipped when using openssl x509 multiple times on a single input file, 5389 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 5390 [Bodo Moeller] 5391 5392 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 5393 set string type: to handle setting ASN1_TIME structures. Fix ca 5394 utility to correctly initialize revocation date of CRLs. 5395 [Steve Henson] 5396 5397 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 5398 the clients preferred ciphersuites and rather use its own preferences. 5399 Should help to work around M$ SGC (Server Gated Cryptography) bug in 5400 Internet Explorer by ensuring unchanged hash method during stepup. 5401 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 5402 [Lutz Jaenicke] 5403 5404 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 5405 to aes and add a new 'exist' option to print out symbols that don't 5406 appear to exist. 5407 [Steve Henson] 5408 5409 *) Additional options to ocsp utility to allow flags to be set and 5410 additional certificates supplied. 5411 [Steve Henson] 5412 5413 *) Add the option -VAfile to 'openssl ocsp', so the user can give the 5414 OCSP client a number of certificate to only verify the response 5415 signature against. 5416 [Richard Levitte] 5417 5418 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 5419 handle the new API. Currently only ECB, CBC modes supported. Add new 5420 AES OIDs. 5421 5422 Add TLS AES ciphersuites as described in RFC3268, "Advanced 5423 Encryption Standard (AES) Ciphersuites for Transport Layer 5424 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 5425 not enabled by default and were not part of the "ALL" ciphersuite 5426 alias because they were not yet official; they could be 5427 explicitly requested by specifying the "AESdraft" ciphersuite 5428 group alias. In the final release of OpenSSL 0.9.7, the group 5429 alias is called "AES" and is part of "ALL".) 5430 [Ben Laurie, Steve Henson, Bodo Moeller] 5431 5432 *) New function OCSP_copy_nonce() to copy nonce value (if present) from 5433 request to response. 5434 [Steve Henson] 5435 5436 *) Functions for OCSP responders. OCSP_request_onereq_count(), 5437 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 5438 extract information from a certificate request. OCSP_response_create() 5439 creates a response and optionally adds a basic response structure. 5440 OCSP_basic_add1_status() adds a complete single response to a basic 5441 response and returns the OCSP_SINGLERESP structure just added (to allow 5442 extensions to be included for example). OCSP_basic_add1_cert() adds a 5443 certificate to a basic response and OCSP_basic_sign() signs a basic 5444 response with various flags. New helper functions ASN1_TIME_check() 5445 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 5446 (converts ASN1_TIME to GeneralizedTime). 5447 [Steve Henson] 5448 5449 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 5450 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 5451 structure from a certificate. X509_pubkey_digest() digests the public_key 5452 contents: this is used in various key identifiers. 5453 [Steve Henson] 5454 5455 *) Make sk_sort() tolerate a NULL argument. 5456 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 5457 5458 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 5459 passed by the function are trusted implicitly. If any of them signed the 5460 response then it is assumed to be valid and is not verified. 5461 [Steve Henson] 5462 5463 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 5464 to data. This was previously part of the PKCS7 ASN1 code. This 5465 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 5466 [Steve Henson, reported by Kenneth R. Robinette 5467 <support@securenetterm.com>] 5468 5469 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 5470 routines: without these tracing memory leaks is very painful. 5471 Fix leaks in PKCS12 and PKCS7 routines. 5472 [Steve Henson] 5473 5474 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 5475 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 5476 effectively meant GeneralizedTime would never be used. Now it 5477 is initialised to -1 but X509_time_adj() now has to check the value 5478 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 5479 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 5480 [Steve Henson, reported by Kenneth R. Robinette 5481 <support@securenetterm.com>] 5482 5483 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 5484 result in a zero length in the ASN1_INTEGER structure which was 5485 not consistent with the structure when d2i_ASN1_INTEGER() was used 5486 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 5487 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 5488 where it did not print out a minus for negative ASN1_INTEGER. 5489 [Steve Henson] 5490 5491 *) Add summary printout to ocsp utility. The various functions which 5492 convert status values to strings have been renamed to: 5493 OCSP_response_status_str(), OCSP_cert_status_str() and 5494 OCSP_crl_reason_str() and are no longer static. New options 5495 to verify nonce values and to disable verification. OCSP response 5496 printout format cleaned up. 5497 [Steve Henson] 5498 5499 *) Add additional OCSP certificate checks. These are those specified 5500 in RFC2560. This consists of two separate checks: the CA of the 5501 certificate being checked must either be the OCSP signer certificate 5502 or the issuer of the OCSP signer certificate. In the latter case the 5503 OCSP signer certificate must contain the OCSP signing extended key 5504 usage. This check is performed by attempting to match the OCSP 5505 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 5506 in the OCSP_CERTID structures of the response. 5507 [Steve Henson] 5508 5509 *) Initial OCSP certificate verification added to OCSP_basic_verify() 5510 and related routines. This uses the standard OpenSSL certificate 5511 verify routines to perform initial checks (just CA validity) and 5512 to obtain the certificate chain. Then additional checks will be 5513 performed on the chain. Currently the root CA is checked to see 5514 if it is explicitly trusted for OCSP signing. This is used to set 5515 a root CA as a global signing root: that is any certificate that 5516 chains to that CA is an acceptable OCSP signing certificate. 5517 [Steve Henson] 5518 5519 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 5520 extensions from a separate configuration file. 5521 As when reading extensions from the main configuration file, 5522 the '-extensions ...' option may be used for specifying the 5523 section to use. 5524 [Massimiliano Pala <madwolf@comune.modena.it>] 5525 5526 *) New OCSP utility. Allows OCSP requests to be generated or 5527 read. The request can be sent to a responder and the output 5528 parsed, outputed or printed in text form. Not complete yet: 5529 still needs to check the OCSP response validity. 5530 [Steve Henson] 5531 5532 *) New subcommands for 'openssl ca': 5533 'openssl ca -status <serial>' prints the status of the cert with 5534 the given serial number (according to the index file). 5535 'openssl ca -updatedb' updates the expiry status of certificates 5536 in the index file. 5537 [Massimiliano Pala <madwolf@comune.modena.it>] 5538 5539 *) New '-newreq-nodes' command option to CA.pl. This is like 5540 '-newreq', but calls 'openssl req' with the '-nodes' option 5541 so that the resulting key is not encrypted. 5542 [Damien Miller <djm@mindrot.org>] 5543 5544 *) New configuration for the GNU Hurd. 5545 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 5546 5547 *) Initial code to implement OCSP basic response verify. This 5548 is currently incomplete. Currently just finds the signer's 5549 certificate and verifies the signature on the response. 5550 [Steve Henson] 5551 5552 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 5553 value of OPENSSLDIR. This is available via the new '-d' option 5554 to 'openssl version', and is also included in 'openssl version -a'. 5555 [Bodo Moeller] 5556 5557 *) Allowing defining memory allocation callbacks that will be given 5558 file name and line number information in additional arguments 5559 (a const char* and an int). The basic functionality remains, as 5560 well as the original possibility to just replace malloc(), 5561 realloc() and free() by functions that do not know about these 5562 additional arguments. To register and find out the current 5563 settings for extended allocation functions, the following 5564 functions are provided: 5565 5566 CRYPTO_set_mem_ex_functions 5567 CRYPTO_set_locked_mem_ex_functions 5568 CRYPTO_get_mem_ex_functions 5569 CRYPTO_get_locked_mem_ex_functions 5570 5571 These work the same way as CRYPTO_set_mem_functions and friends. 5572 CRYPTO_get_[locked_]mem_functions now writes 0 where such an 5573 extended allocation function is enabled. 5574 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 5575 a conventional allocation function is enabled. 5576 [Richard Levitte, Bodo Moeller] 5577 5578 *) Finish off removing the remaining LHASH function pointer casts. 5579 There should no longer be any prototype-casting required when using 5580 the LHASH abstraction, and any casts that remain are "bugs". See 5581 the callback types and macros at the head of lhash.h for details 5582 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 5583 [Geoff Thorpe] 5584 5585 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 5586 If /dev/[u]random devices are not available or do not return enough 5587 entropy, EGD style sockets (served by EGD or PRNGD) will automatically 5588 be queried. 5589 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 5590 /etc/entropy will be queried once each in this sequence, quering stops 5591 when enough entropy was collected without querying more sockets. 5592 [Lutz Jaenicke] 5593 5594 *) Change the Unix RAND_poll() variant to be able to poll several 5595 random devices, as specified by DEVRANDOM, until a sufficient amount 5596 of data has been collected. We spend at most 10 ms on each file 5597 (select timeout) and read in non-blocking mode. DEVRANDOM now 5598 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 5599 (previously it was just the string "/dev/urandom"), so on typical 5600 platforms the 10 ms delay will never occur. 5601 Also separate out the Unix variant to its own file, rand_unix.c. 5602 For VMS, there's a currently-empty rand_vms.c. 5603 [Richard Levitte] 5604 5605 *) Move OCSP client related routines to ocsp_cl.c. These 5606 provide utility functions which an application needing 5607 to issue a request to an OCSP responder and analyse the 5608 response will typically need: as opposed to those which an 5609 OCSP responder itself would need which will be added later. 5610 5611 OCSP_request_sign() signs an OCSP request with an API similar 5612 to PKCS7_sign(). OCSP_response_status() returns status of OCSP 5613 response. OCSP_response_get1_basic() extracts basic response 5614 from response. OCSP_resp_find_status(): finds and extracts status 5615 information from an OCSP_CERTID structure (which will be created 5616 when the request structure is built). These are built from lower 5617 level functions which work on OCSP_SINGLERESP structures but 5618 wont normally be used unless the application wishes to examine 5619 extensions in the OCSP response for example. 5620 5621 Replace nonce routines with a pair of functions. 5622 OCSP_request_add1_nonce() adds a nonce value and optionally 5623 generates a random value. OCSP_check_nonce() checks the 5624 validity of the nonce in an OCSP response. 5625 [Steve Henson] 5626 5627 *) Change function OCSP_request_add() to OCSP_request_add0_id(). 5628 This doesn't copy the supplied OCSP_CERTID and avoids the 5629 need to free up the newly created id. Change return type 5630 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 5631 This can then be used to add extensions to the request. 5632 Deleted OCSP_request_new(), since most of its functionality 5633 is now in OCSP_REQUEST_new() (and the case insensitive name 5634 clash) apart from the ability to set the request name which 5635 will be added elsewhere. 5636 [Steve Henson] 5637 5638 *) Update OCSP API. Remove obsolete extensions argument from 5639 various functions. Extensions are now handled using the new 5640 OCSP extension code. New simple OCSP HTTP function which 5641 can be used to send requests and parse the response. 5642 [Steve Henson] 5643 5644 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 5645 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 5646 uses the special reorder version of SET OF to sort the attributes 5647 and reorder them to match the encoded order. This resolves a long 5648 standing problem: a verify on a PKCS7 structure just after signing 5649 it used to fail because the attribute order did not match the 5650 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 5651 it uses the received order. This is necessary to tolerate some broken 5652 software that does not order SET OF. This is handled by encoding 5653 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 5654 to produce the required SET OF. 5655 [Steve Henson] 5656 5657 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 5658 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 5659 files to get correct declarations of the ASN.1 item variables. 5660 [Richard Levitte] 5661 5662 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 5663 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 5664 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 5665 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 5666 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 5667 ASN1_ITEM and no wrapper functions. 5668 [Steve Henson] 5669 5670 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 5671 replace the old function pointer based I/O routines. Change most of 5672 the *_d2i_bio() and *_d2i_fp() functions to use these. 5673 [Steve Henson] 5674 5675 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 5676 lines, recognice more "algorithms" that can be deselected, and make 5677 it complain about algorithm deselection that isn't recognised. 5678 [Richard Levitte] 5679 5680 *) New ASN1 functions to handle dup, sign, verify, digest, pack and 5681 unpack operations in terms of ASN1_ITEM. Modify existing wrappers 5682 to use new functions. Add NO_ASN1_OLD which can be set to remove 5683 some old style ASN1 functions: this can be used to determine if old 5684 code will still work when these eventually go away. 5685 [Steve Henson] 5686 5687 *) New extension functions for OCSP structures, these follow the 5688 same conventions as certificates and CRLs. 5689 [Steve Henson] 5690 5691 *) New function X509V3_add1_i2d(). This automatically encodes and 5692 adds an extension. Its behaviour can be customised with various 5693 flags to append, replace or delete. Various wrappers added for 5694 certifcates and CRLs. 5695 [Steve Henson] 5696 5697 *) Fix to avoid calling the underlying ASN1 print routine when 5698 an extension cannot be parsed. Correct a typo in the 5699 OCSP_SERVICELOC extension. Tidy up print OCSP format. 5700 [Steve Henson] 5701 5702 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 5703 entries for variables. 5704 [Steve Henson] 5705 5706 *) Add functionality to apps/openssl.c for detecting locking 5707 problems: As the program is single-threaded, all we have 5708 to do is register a locking callback using an array for 5709 storing which locks are currently held by the program. 5710 [Bodo Moeller] 5711 5712 *) Use a lock around the call to CRYPTO_get_ex_new_index() in 5713 SSL_get_ex_data_X509_STORE_idx(), which is used in 5714 ssl_verify_cert_chain() and thus can be called at any time 5715 during TLS/SSL handshakes so that thread-safety is essential. 5716 Unfortunately, the ex_data design is not at all suited 5717 for multi-threaded use, so it probably should be abolished. 5718 [Bodo Moeller] 5719 5720 *) Added Broadcom "ubsec" ENGINE to OpenSSL. 5721 [Broadcom, tweaked and integrated by Geoff Thorpe] 5722 5723 *) Move common extension printing code to new function 5724 X509V3_print_extensions(). Reorganise OCSP print routines and 5725 implement some needed OCSP ASN1 functions. Add OCSP extensions. 5726 [Steve Henson] 5727 5728 *) New function X509_signature_print() to remove duplication in some 5729 print routines. 5730 [Steve Henson] 5731 5732 *) Add a special meaning when SET OF and SEQUENCE OF flags are both 5733 set (this was treated exactly the same as SET OF previously). This 5734 is used to reorder the STACK representing the structure to match the 5735 encoding. This will be used to get round a problem where a PKCS7 5736 structure which was signed could not be verified because the STACK 5737 order did not reflect the encoded order. 5738 [Steve Henson] 5739 5740 *) Reimplement the OCSP ASN1 module using the new code. 5741 [Steve Henson] 5742 5743 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 5744 for its ASN1 operations. The old style function pointers still exist 5745 for now but they will eventually go away. 5746 [Steve Henson] 5747 5748 *) Merge in replacement ASN1 code from the ASN1 branch. This almost 5749 completely replaces the old ASN1 functionality with a table driven 5750 encoder and decoder which interprets an ASN1_ITEM structure describing 5751 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 5752 largely maintained. Almost all of the old asn1_mac.h macro based ASN1 5753 has also been converted to the new form. 5754 [Steve Henson] 5755 5756 *) Change BN_mod_exp_recp so that negative moduli are tolerated 5757 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 5758 so that BN_mod_exp_mont and BN_mod_exp_mont_word work 5759 for negative moduli. 5760 [Bodo Moeller] 5761 5762 *) Fix BN_uadd and BN_usub: Always return non-negative results instead 5763 of not touching the result's sign bit. 5764 [Bodo Moeller] 5765 5766 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 5767 set. 5768 [Bodo Moeller] 5769 5770 *) Changed the LHASH code to use prototypes for callbacks, and created 5771 macros to declare and implement thin (optionally static) functions 5772 that provide type-safety and avoid function pointer casting for the 5773 type-specific callbacks. 5774 [Geoff Thorpe] 5775 5776 *) Added Kerberos Cipher Suites to be used with TLS, as written in 5777 RFC 2712. 5778 [Veers Staats <staatsvr@asc.hpc.mil>, 5779 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 5780 5781 *) Reformat the FAQ so the different questions and answers can be divided 5782 in sections depending on the subject. 5783 [Richard Levitte] 5784 5785 *) Have the zlib compression code load ZLIB.DLL dynamically under 5786 Windows. 5787 [Richard Levitte] 5788 5789 *) New function BN_mod_sqrt for computing square roots modulo a prime 5790 (using the probabilistic Tonelli-Shanks algorithm unless 5791 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 5792 be handled deterministically). 5793 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5794 5795 *) Make BN_mod_inverse faster by explicitly handling small quotients 5796 in the Euclid loop. (Speed gain about 20% for small moduli [256 or 5797 512 bits], about 30% for larger ones [1024 or 2048 bits].) 5798 [Bodo Moeller] 5799 5800 *) New function BN_kronecker. 5801 [Bodo Moeller] 5802 5803 *) Fix BN_gcd so that it works on negative inputs; the result is 5804 positive unless both parameters are zero. 5805 Previously something reasonably close to an infinite loop was 5806 possible because numbers could be growing instead of shrinking 5807 in the implementation of Euclid's algorithm. 5808 [Bodo Moeller] 5809 5810 *) Fix BN_is_word() and BN_is_one() macros to take into account the 5811 sign of the number in question. 5812 5813 Fix BN_is_word(a,w) to work correctly for w == 0. 5814 5815 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 5816 because its test if the absolute value of 'a' equals 'w'. 5817 Note that BN_abs_is_word does *not* handle w == 0 reliably; 5818 it exists mostly for use in the implementations of BN_is_zero(), 5819 BN_is_one(), and BN_is_word(). 5820 [Bodo Moeller] 5821 5822 *) New function BN_swap. 5823 [Bodo Moeller] 5824 5825 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 5826 the exponentiation functions are more likely to produce reasonable 5827 results on negative inputs. 5828 [Bodo Moeller] 5829 5830 *) Change BN_mod_mul so that the result is always non-negative. 5831 Previously, it could be negative if one of the factors was negative; 5832 I don't think anyone really wanted that behaviour. 5833 [Bodo Moeller] 5834 5835 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 5836 (except for exponentiation, which stays in crypto/bn/bn_exp.c, 5837 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 5838 and add new functions: 5839 5840 BN_nnmod 5841 BN_mod_sqr 5842 BN_mod_add 5843 BN_mod_add_quick 5844 BN_mod_sub 5845 BN_mod_sub_quick 5846 BN_mod_lshift1 5847 BN_mod_lshift1_quick 5848 BN_mod_lshift 5849 BN_mod_lshift_quick 5850 5851 These functions always generate non-negative results. 5852 5853 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 5854 such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 5855 5856 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 5857 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 5858 be reduced modulo m. 5859 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5860 5861#if 0 5862 The following entry accidentily appeared in the CHANGES file 5863 distributed with OpenSSL 0.9.7. The modifications described in 5864 it do *not* apply to OpenSSL 0.9.7. 5865 5866 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 5867 was actually never needed) and in BN_mul(). The removal in BN_mul() 5868 required a small change in bn_mul_part_recursive() and the addition 5869 of the functions bn_cmp_part_words(), bn_sub_part_words() and 5870 bn_add_part_words(), which do the same thing as bn_cmp_words(), 5871 bn_sub_words() and bn_add_words() except they take arrays with 5872 differing sizes. 5873 [Richard Levitte] 5874#endif 5875 5876 *) In 'openssl passwd', verify passwords read from the terminal 5877 unless the '-salt' option is used (which usually means that 5878 verification would just waste user's time since the resulting 5879 hash is going to be compared with some given password hash) 5880 or the new '-noverify' option is used. 5881 5882 This is an incompatible change, but it does not affect 5883 non-interactive use of 'openssl passwd' (passwords on the command 5884 line, '-stdin' option, '-in ...' option) and thus should not 5885 cause any problems. 5886 [Bodo Moeller] 5887 5888 *) Remove all references to RSAref, since there's no more need for it. 5889 [Richard Levitte] 5890 5891 *) Make DSO load along a path given through an environment variable 5892 (SHLIB_PATH) with shl_load(). 5893 [Richard Levitte] 5894 5895 *) Constify the ENGINE code as a result of BIGNUM constification. 5896 Also constify the RSA code and most things related to it. In a 5897 few places, most notable in the depth of the ASN.1 code, ugly 5898 casts back to non-const were required (to be solved at a later 5899 time) 5900 [Richard Levitte] 5901 5902 *) Make it so the openssl application has all engines loaded by default. 5903 [Richard Levitte] 5904 5905 *) Constify the BIGNUM routines a little more. 5906 [Richard Levitte] 5907 5908 *) Add the following functions: 5909 5910 ENGINE_load_cswift() 5911 ENGINE_load_chil() 5912 ENGINE_load_atalla() 5913 ENGINE_load_nuron() 5914 ENGINE_load_builtin_engines() 5915 5916 That way, an application can itself choose if external engines that 5917 are built-in in OpenSSL shall ever be used or not. The benefit is 5918 that applications won't have to be linked with libdl or other dso 5919 libraries unless it's really needed. 5920 5921 Changed 'openssl engine' to load all engines on demand. 5922 Changed the engine header files to avoid the duplication of some 5923 declarations (they differed!). 5924 [Richard Levitte] 5925 5926 *) 'openssl engine' can now list capabilities. 5927 [Richard Levitte] 5928 5929 *) Better error reporting in 'openssl engine'. 5930 [Richard Levitte] 5931 5932 *) Never call load_dh_param(NULL) in s_server. 5933 [Bodo Moeller] 5934 5935 *) Add engine application. It can currently list engines by name and 5936 identity, and test if they are actually available. 5937 [Richard Levitte] 5938 5939 *) Improve RPM specification file by forcing symbolic linking and making 5940 sure the installed documentation is also owned by root.root. 5941 [Damien Miller <djm@mindrot.org>] 5942 5943 *) Give the OpenSSL applications more possibilities to make use of 5944 keys (public as well as private) handled by engines. 5945 [Richard Levitte] 5946 5947 *) Add OCSP code that comes from CertCo. 5948 [Richard Levitte] 5949 5950 *) Add VMS support for the Rijndael code. 5951 [Richard Levitte] 5952 5953 *) Added untested support for Nuron crypto accelerator. 5954 [Ben Laurie] 5955 5956 *) Add support for external cryptographic devices. This code was 5957 previously distributed separately as the "engine" branch. 5958 [Geoff Thorpe, Richard Levitte] 5959 5960 *) Rework the filename-translation in the DSO code. It is now possible to 5961 have far greater control over how a "name" is turned into a filename 5962 depending on the operating environment and any oddities about the 5963 different shared library filenames on each system. 5964 [Geoff Thorpe] 5965 5966 *) Support threads on FreeBSD-elf in Configure. 5967 [Richard Levitte] 5968 5969 *) Fix for SHA1 assembly problem with MASM: it produces 5970 warnings about corrupt line number information when assembling 5971 with debugging information. This is caused by the overlapping 5972 of two sections. 5973 [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 5974 5975 *) NCONF changes. 5976 NCONF_get_number() has no error checking at all. As a replacement, 5977 NCONF_get_number_e() is defined (_e for "error checking") and is 5978 promoted strongly. The old NCONF_get_number is kept around for 5979 binary backward compatibility. 5980 Make it possible for methods to load from something other than a BIO, 5981 by providing a function pointer that is given a name instead of a BIO. 5982 For example, this could be used to load configuration data from an 5983 LDAP server. 5984 [Richard Levitte] 5985 5986 *) Fix for non blocking accept BIOs. Added new I/O special reason 5987 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 5988 with non blocking I/O was not possible because no retry code was 5989 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 5990 this case. 5991 [Steve Henson] 5992 5993 *) Added the beginnings of Rijndael support. 5994 [Ben Laurie] 5995 5996 *) Fix for bug in DirectoryString mask setting. Add support for 5997 X509_NAME_print_ex() in 'req' and X509_print_ex() function 5998 to allow certificate printing to more controllable, additional 5999 'certopt' option to 'x509' to allow new printing options to be 6000 set. 6001 [Steve Henson] 6002 6003 *) Clean old EAY MD5 hack from e_os.h. 6004 [Richard Levitte] 6005 6006 Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 6007 6008 *) Fix null-pointer assignment in do_change_cipher_spec() revealed 6009 by using the Codenomicon TLS Test Tool (CVE-2004-0079) 6010 [Joe Orton, Steve Henson] 6011 6012 Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 6013 6014 *) Fix additional bug revealed by the NISCC test suite: 6015 6016 Stop bug triggering large recursion when presented with 6017 certain ASN.1 tags (CVE-2003-0851) 6018 [Steve Henson] 6019 6020 Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 6021 6022 *) Fix various bugs revealed by running the NISCC test suite: 6023 6024 Stop out of bounds reads in the ASN1 code when presented with 6025 invalid tags (CVE-2003-0543 and CVE-2003-0544). 6026 6027 If verify callback ignores invalid public key errors don't try to check 6028 certificate signature with the NULL public key. 6029 6030 [Steve Henson] 6031 6032 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 6033 if the server requested one: as stated in TLS 1.0 and SSL 3.0 6034 specifications. 6035 [Steve Henson] 6036 6037 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 6038 extra data after the compression methods not only for TLS 1.0 6039 but also for SSL 3.0 (as required by the specification). 6040 [Bodo Moeller; problem pointed out by Matthias Loepfe] 6041 6042 *) Change X509_certificate_type() to mark the key as exported/exportable 6043 when it's 512 *bits* long, not 512 bytes. 6044 [Richard Levitte] 6045 6046 Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 6047 6048 *) Countermeasure against the Klima-Pokorny-Rosa extension of 6049 Bleichbacher's attack on PKCS #1 v1.5 padding: treat 6050 a protocol version number mismatch like a decryption error 6051 in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 6052 [Bodo Moeller] 6053 6054 *) Turn on RSA blinding by default in the default implementation 6055 to avoid a timing attack. Applications that don't want it can call 6056 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 6057 They would be ill-advised to do so in most cases. 6058 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 6059 6060 *) Change RSA blinding code so that it works when the PRNG is not 6061 seeded (in this case, the secret RSA exponent is abused as 6062 an unpredictable seed -- if it is not unpredictable, there 6063 is no point in blinding anyway). Make RSA blinding thread-safe 6064 by remembering the creator's thread ID in rsa->blinding and 6065 having all other threads use local one-time blinding factors 6066 (this requires more computation than sharing rsa->blinding, but 6067 avoids excessive locking; and if an RSA object is not shared 6068 between threads, blinding will still be very fast). 6069 [Bodo Moeller] 6070 6071 Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 6072 6073 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 6074 via timing by performing a MAC computation even if incorrrect 6075 block cipher padding has been found. This is a countermeasure 6076 against active attacks where the attacker has to distinguish 6077 between bad padding and a MAC verification error. (CVE-2003-0078) 6078 6079 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 6080 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 6081 Martin Vuagnoux (EPFL, Ilion)] 6082 6083 Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 6084 6085 *) New function OPENSSL_cleanse(), which is used to cleanse a section of 6086 memory from it's contents. This is done with a counter that will 6087 place alternating values in each byte. This can be used to solve 6088 two issues: 1) the removal of calls to memset() by highly optimizing 6089 compilers, and 2) cleansing with other values than 0, since those can 6090 be read through on certain media, for example a swap space on disk. 6091 [Geoff Thorpe] 6092 6093 *) Bugfix: client side session caching did not work with external caching, 6094 because the session->cipher setting was not restored when reloading 6095 from the external cache. This problem was masked, when 6096 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 6097 (Found by Steve Haslam <steve@araqnid.ddts.net>.) 6098 [Lutz Jaenicke] 6099 6100 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 6101 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 6102 [Zeev Lieber <zeev-l@yahoo.com>] 6103 6104 *) Undo an undocumented change introduced in 0.9.6e which caused 6105 repeated calls to OpenSSL_add_all_ciphers() and 6106 OpenSSL_add_all_digests() to be ignored, even after calling 6107 EVP_cleanup(). 6108 [Richard Levitte] 6109 6110 *) Change the default configuration reader to deal with last line not 6111 being properly terminated. 6112 [Richard Levitte] 6113 6114 *) Change X509_NAME_cmp() so it applies the special rules on handling 6115 DN values that are of type PrintableString, as well as RDNs of type 6116 emailAddress where the value has the type ia5String. 6117 [stefank@valicert.com via Richard Levitte] 6118 6119 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 6120 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 6121 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 6122 the bitwise-OR of the two for use by the majority of applications 6123 wanting this behaviour, and update the docs. The documented 6124 behaviour and actual behaviour were inconsistent and had been 6125 changing anyway, so this is more a bug-fix than a behavioural 6126 change. 6127 [Geoff Thorpe, diagnosed by Nadav Har'El] 6128 6129 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 6130 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 6131 [Bodo Moeller] 6132 6133 *) Fix initialization code race conditions in 6134 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 6135 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 6136 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 6137 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 6138 ssl2_get_cipher_by_char(), 6139 ssl3_get_cipher_by_char(). 6140 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 6141 6142 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 6143 the cached sessions are flushed, as the remove_cb() might use ex_data 6144 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 6145 (see [openssl.org #212]). 6146 [Geoff Thorpe, Lutz Jaenicke] 6147 6148 *) Fix typo in OBJ_txt2obj which incorrectly passed the content 6149 length, instead of the encoding length to d2i_ASN1_OBJECT. 6150 [Steve Henson] 6151 6152 Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 6153 6154 *) [In 0.9.6g-engine release:] 6155 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 6156 [Lynn Gazis <lgazis@rainbow.com>] 6157 6158 Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 6159 6160 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 6161 and get fix the header length calculation. 6162 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 6163 Alon Kantor <alonk@checkpoint.com> (and others), 6164 Steve Henson] 6165 6166 *) Use proper error handling instead of 'assertions' in buffer 6167 overflow checks added in 0.9.6e. This prevents DoS (the 6168 assertions could call abort()). 6169 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 6170 6171 Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 6172 6173 *) Add various sanity checks to asn1_get_length() to reject 6174 the ASN1 length bytes if they exceed sizeof(long), will appear 6175 negative or the content length exceeds the length of the 6176 supplied buffer. 6177 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6178 6179 *) Fix cipher selection routines: ciphers without encryption had no flags 6180 for the cipher strength set and where therefore not handled correctly 6181 by the selection routines (PR #130). 6182 [Lutz Jaenicke] 6183 6184 *) Fix EVP_dsa_sha macro. 6185 [Nils Larsch] 6186 6187 *) New option 6188 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 6189 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 6190 that was added in OpenSSL 0.9.6d. 6191 6192 As the countermeasure turned out to be incompatible with some 6193 broken SSL implementations, the new option is part of SSL_OP_ALL. 6194 SSL_OP_ALL is usually employed when compatibility with weird SSL 6195 implementations is desired (e.g. '-bugs' option to 's_client' and 6196 's_server'), so the new option is automatically set in many 6197 applications. 6198 [Bodo Moeller] 6199 6200 *) Changes in security patch: 6201 6202 Changes marked "(CHATS)" were sponsored by the Defense Advanced 6203 Research Projects Agency (DARPA) and Air Force Research Laboratory, 6204 Air Force Materiel Command, USAF, under agreement number 6205 F30602-01-2-0537. 6206 6207 *) Add various sanity checks to asn1_get_length() to reject 6208 the ASN1 length bytes if they exceed sizeof(long), will appear 6209 negative or the content length exceeds the length of the 6210 supplied buffer. (CVE-2002-0659) 6211 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6212 6213 *) Assertions for various potential buffer overflows, not known to 6214 happen in practice. 6215 [Ben Laurie (CHATS)] 6216 6217 *) Various temporary buffers to hold ASCII versions of integers were 6218 too small for 64 bit platforms. (CVE-2002-0655) 6219 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 6220 6221 *) Remote buffer overflow in SSL3 protocol - an attacker could 6222 supply an oversized session ID to a client. (CVE-2002-0656) 6223 [Ben Laurie (CHATS)] 6224 6225 *) Remote buffer overflow in SSL2 protocol - an attacker could 6226 supply an oversized client master key. (CVE-2002-0656) 6227 [Ben Laurie (CHATS)] 6228 6229 Changes between 0.9.6c and 0.9.6d [9 May 2002] 6230 6231 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 6232 encoded as NULL) with id-dsa-with-sha1. 6233 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 6234 6235 *) Check various X509_...() return values in apps/req.c. 6236 [Nils Larsch <nla@trustcenter.de>] 6237 6238 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 6239 an end-of-file condition would erronously be flagged, when the CRLF 6240 was just at the end of a processed block. The bug was discovered when 6241 processing data through a buffering memory BIO handing the data to a 6242 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 6243 <ptsekov@syntrex.com> and Nedelcho Stanev. 6244 [Lutz Jaenicke] 6245 6246 *) Implement a countermeasure against a vulnerability recently found 6247 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 6248 before application data chunks to avoid the use of known IVs 6249 with data potentially chosen by the attacker. 6250 [Bodo Moeller] 6251 6252 *) Fix length checks in ssl3_get_client_hello(). 6253 [Bodo Moeller] 6254 6255 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 6256 to prevent ssl3_read_internal() from incorrectly assuming that 6257 ssl3_read_bytes() found application data while handshake 6258 processing was enabled when in fact s->s3->in_read_app_data was 6259 merely automatically cleared during the initial handshake. 6260 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 6261 6262 *) Fix object definitions for Private and Enterprise: they were not 6263 recognized in their shortname (=lowercase) representation. Extend 6264 obj_dat.pl to issue an error when using undefined keywords instead 6265 of silently ignoring the problem (Svenning Sorensen 6266 <sss@sss.dnsalias.net>). 6267 [Lutz Jaenicke] 6268 6269 *) Fix DH_generate_parameters() so that it works for 'non-standard' 6270 generators, i.e. generators other than 2 and 5. (Previously, the 6271 code did not properly initialise the 'add' and 'rem' values to 6272 BN_generate_prime().) 6273 6274 In the new general case, we do not insist that 'generator' is 6275 actually a primitive root: This requirement is rather pointless; 6276 a generator of the order-q subgroup is just as good, if not 6277 better. 6278 [Bodo Moeller] 6279 6280 *) Map new X509 verification errors to alerts. Discovered and submitted by 6281 Tom Wu <tom@arcot.com>. 6282 [Lutz Jaenicke] 6283 6284 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 6285 returning non-zero before the data has been completely received 6286 when using non-blocking I/O. 6287 [Bodo Moeller; problem pointed out by John Hughes] 6288 6289 *) Some of the ciphers missed the strength entry (SSL_LOW etc). 6290 [Ben Laurie, Lutz Jaenicke] 6291 6292 *) Fix bug in SSL_clear(): bad sessions were not removed (found by 6293 Yoram Zahavi <YoramZ@gilian.com>). 6294 [Lutz Jaenicke] 6295 6296 *) Add information about CygWin 1.3 and on, and preserve proper 6297 configuration for the versions before that. 6298 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 6299 6300 *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 6301 check whether we deal with a copy of a session and do not delete from 6302 the cache in this case. Problem reported by "Izhar Shoshani Levi" 6303 <izhar@checkpoint.com>. 6304 [Lutz Jaenicke] 6305 6306 *) Do not store session data into the internal session cache, if it 6307 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 6308 flag is set). Proposed by Aslam <aslam@funk.com>. 6309 [Lutz Jaenicke] 6310 6311 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 6312 value is 0. 6313 [Richard Levitte] 6314 6315 *) [In 0.9.6d-engine release:] 6316 Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 6317 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 6318 6319 *) Add the configuration target linux-s390x. 6320 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 6321 6322 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 6323 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 6324 variable as an indication that a ClientHello message has been 6325 received. As the flag value will be lost between multiple 6326 invocations of ssl3_accept when using non-blocking I/O, the 6327 function may not be aware that a handshake has actually taken 6328 place, thus preventing a new session from being added to the 6329 session cache. 6330 6331 To avoid this problem, we now set s->new_session to 2 instead of 6332 using a local variable. 6333 [Lutz Jaenicke, Bodo Moeller] 6334 6335 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 6336 if the SSL_R_LENGTH_MISMATCH error is detected. 6337 [Geoff Thorpe, Bodo Moeller] 6338 6339 *) New 'shared_ldflag' column in Configure platform table. 6340 [Richard Levitte] 6341 6342 *) Fix EVP_CIPHER_mode macro. 6343 ["Dan S. Camper" <dan@bti.net>] 6344 6345 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 6346 type, we must throw them away by setting rr->length to 0. 6347 [D P Chang <dpc@qualys.com>] 6348 6349 Changes between 0.9.6b and 0.9.6c [21 dec 2001] 6350 6351 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 6352 <Dominikus.Scherkl@biodata.com>. (The previous implementation 6353 worked incorrectly for those cases where range = 10..._2 and 6354 3*range is two bits longer than range.) 6355 [Bodo Moeller] 6356 6357 *) Only add signing time to PKCS7 structures if it is not already 6358 present. 6359 [Steve Henson] 6360 6361 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 6362 OBJ_ld_ce should be OBJ_id_ce. 6363 Also some ip-pda OIDs in crypto/objects/objects.txt were 6364 incorrect (cf. RFC 3039). 6365 [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 6366 6367 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 6368 returns early because it has nothing to do. 6369 [Andy Schneider <andy.schneider@bjss.co.uk>] 6370 6371 *) [In 0.9.6c-engine release:] 6372 Fix mutex callback return values in crypto/engine/hw_ncipher.c. 6373 [Andy Schneider <andy.schneider@bjss.co.uk>] 6374 6375 *) [In 0.9.6c-engine release:] 6376 Add support for Cryptographic Appliance's keyserver technology. 6377 (Use engine 'keyclient') 6378 [Cryptographic Appliances and Geoff Thorpe] 6379 6380 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 6381 is called via tools/c89.sh because arguments have to be 6382 rearranged (all '-L' options must appear before the first object 6383 modules). 6384 [Richard Shapiro <rshapiro@abinitio.com>] 6385 6386 *) [In 0.9.6c-engine release:] 6387 Add support for Broadcom crypto accelerator cards, backported 6388 from 0.9.7. 6389 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 6390 6391 *) [In 0.9.6c-engine release:] 6392 Add support for SureWare crypto accelerator cards from 6393 Baltimore Technologies. (Use engine 'sureware') 6394 [Baltimore Technologies and Mark Cox] 6395 6396 *) [In 0.9.6c-engine release:] 6397 Add support for crypto accelerator cards from Accelerated 6398 Encryption Processing, www.aep.ie. (Use engine 'aep') 6399 [AEP Inc. and Mark Cox] 6400 6401 *) Add a configuration entry for gcc on UnixWare. 6402 [Gary Benson <gbenson@redhat.com>] 6403 6404 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 6405 messages are stored in a single piece (fixed-length part and 6406 variable-length part combined) and fix various bugs found on the way. 6407 [Bodo Moeller] 6408 6409 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 6410 instead. BIO_gethostbyname() does not know what timeouts are 6411 appropriate, so entries would stay in cache even when they have 6412 become invalid. 6413 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 6414 6415 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 6416 faced with a pathologically small ClientHello fragment that does 6417 not contain client_version: Instead of aborting with an error, 6418 simply choose the highest available protocol version (i.e., 6419 TLS 1.0 unless it is disabled). In practice, ClientHello 6420 messages are never sent like this, but this change gives us 6421 strictly correct behaviour at least for TLS. 6422 [Bodo Moeller] 6423 6424 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 6425 never resets s->method to s->ctx->method when called from within 6426 one of the SSL handshake functions. 6427 [Bodo Moeller; problem pointed out by Niko Baric] 6428 6429 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 6430 (sent using the client's version number) if client_version is 6431 smaller than the protocol version in use. Also change 6432 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 6433 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 6434 the client will at least see that alert. 6435 [Bodo Moeller] 6436 6437 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 6438 correctly. 6439 [Bodo Moeller] 6440 6441 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 6442 client receives HelloRequest while in a handshake. 6443 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 6444 6445 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 6446 should end in 'break', not 'goto end' which circuments various 6447 cleanups done in state SSL_ST_OK. But session related stuff 6448 must be disabled for SSL_ST_OK in the case that we just sent a 6449 HelloRequest. 6450 6451 Also avoid some overhead by not calling ssl_init_wbio_buffer() 6452 before just sending a HelloRequest. 6453 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 6454 6455 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 6456 reveal whether illegal block cipher padding was found or a MAC 6457 verification error occured. (Neither SSLerr() codes nor alerts 6458 are directly visible to potential attackers, but the information 6459 may leak via logfiles.) 6460 6461 Similar changes are not required for the SSL 2.0 implementation 6462 because the number of padding bytes is sent in clear for SSL 2.0, 6463 and the extra bytes are just ignored. However ssl/s2_pkt.c 6464 failed to verify that the purported number of padding bytes is in 6465 the legal range. 6466 [Bodo Moeller] 6467 6468 *) Add OpenUNIX-8 support including shared libraries 6469 (Boyd Lynn Gerber <gerberb@zenez.com>). 6470 [Lutz Jaenicke] 6471 6472 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 6473 'wristwatch attack' using huge encoding parameters (cf. 6474 James H. Manger's CRYPTO 2001 paper). Note that the 6475 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 6476 encoding parameters and hence was not vulnerable. 6477 [Bodo Moeller] 6478 6479 *) BN_sqr() bug fix. 6480 [Ulf M��ller, reported by Jim Ellis <jim.ellis@cavium.com>] 6481 6482 *) Rabin-Miller test analyses assume uniformly distributed witnesses, 6483 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 6484 followed by modular reduction. 6485 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 6486 6487 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 6488 equivalent based on BN_pseudo_rand() instead of BN_rand(). 6489 [Bodo Moeller] 6490 6491 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 6492 This function was broken, as the check for a new client hello message 6493 to handle SGC did not allow these large messages. 6494 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 6495 [Lutz Jaenicke] 6496 6497 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 6498 [Lutz Jaenicke] 6499 6500 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 6501 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 6502 [Lutz Jaenicke] 6503 6504 *) Rework the configuration and shared library support for Tru64 Unix. 6505 The configuration part makes use of modern compiler features and 6506 still retains old compiler behavior for those that run older versions 6507 of the OS. The shared library support part includes a variant that 6508 uses the RPATH feature, and is available through the special 6509 configuration target "alpha-cc-rpath", which will never be selected 6510 automatically. 6511 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 6512 6513 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 6514 with the same message size as in ssl3_get_certificate_request(). 6515 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 6516 messages might inadvertently be reject as too long. 6517 [Petr Lampa <lampa@fee.vutbr.cz>] 6518 6519 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 6520 [Andy Polyakov] 6521 6522 *) Modified SSL library such that the verify_callback that has been set 6523 specificly for an SSL object with SSL_set_verify() is actually being 6524 used. Before the change, a verify_callback set with this function was 6525 ignored and the verify_callback() set in the SSL_CTX at the time of 6526 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 6527 to allow the necessary settings. 6528 [Lutz Jaenicke] 6529 6530 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 6531 explicitly to NULL, as at least on Solaris 8 this seems not always to be 6532 done automatically (in contradiction to the requirements of the C 6533 standard). This made problems when used from OpenSSH. 6534 [Lutz Jaenicke] 6535 6536 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 6537 dh->length and always used 6538 6539 BN_rand_range(priv_key, dh->p). 6540 6541 BN_rand_range() is not necessary for Diffie-Hellman, and this 6542 specific range makes Diffie-Hellman unnecessarily inefficient if 6543 dh->length (recommended exponent length) is much smaller than the 6544 length of dh->p. We could use BN_rand_range() if the order of 6545 the subgroup was stored in the DH structure, but we only have 6546 dh->length. 6547 6548 So switch back to 6549 6550 BN_rand(priv_key, l, ...) 6551 6552 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 6553 otherwise. 6554 [Bodo Moeller] 6555 6556 *) In 6557 6558 RSA_eay_public_encrypt 6559 RSA_eay_private_decrypt 6560 RSA_eay_private_encrypt (signing) 6561 RSA_eay_public_decrypt (signature verification) 6562 6563 (default implementations for RSA_public_encrypt, 6564 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 6565 always reject numbers >= n. 6566 [Bodo Moeller] 6567 6568 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 6569 to synchronize access to 'locking_thread'. This is necessary on 6570 systems where access to 'locking_thread' (an 'unsigned long' 6571 variable) is not atomic. 6572 [Bodo Moeller] 6573 6574 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 6575 *before* setting the 'crypto_lock_rand' flag. The previous code had 6576 a race condition if 0 is a valid thread ID. 6577 [Travis Vitek <vitek@roguewave.com>] 6578 6579 *) Add support for shared libraries under Irix. 6580 [Albert Chin-A-Young <china@thewrittenword.com>] 6581 6582 *) Add configuration option to build on Linux on both big-endian and 6583 little-endian MIPS. 6584 [Ralf Baechle <ralf@uni-koblenz.de>] 6585 6586 *) Add the possibility to create shared libraries on HP-UX. 6587 [Richard Levitte] 6588 6589 Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 6590 6591 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 6592 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 6593 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 6594 PRNG state recovery was possible based on the output of 6595 one PRNG request appropriately sized to gain knowledge on 6596 'md' followed by enough consecutive 1-byte PRNG requests 6597 to traverse all of 'state'. 6598 6599 1. When updating 'md_local' (the current thread's copy of 'md') 6600 during PRNG output generation, hash all of the previous 6601 'md_local' value, not just the half used for PRNG output. 6602 6603 2. Make the number of bytes from 'state' included into the hash 6604 independent from the number of PRNG bytes requested. 6605 6606 The first measure alone would be sufficient to avoid 6607 Markku-Juhani's attack. (Actually it had never occurred 6608 to me that the half of 'md_local' used for chaining was the 6609 half from which PRNG output bytes were taken -- I had always 6610 assumed that the secret half would be used.) The second 6611 measure makes sure that additional data from 'state' is never 6612 mixed into 'md_local' in small portions; this heuristically 6613 further strengthens the PRNG. 6614 [Bodo Moeller] 6615 6616 *) Fix crypto/bn/asm/mips3.s. 6617 [Andy Polyakov] 6618 6619 *) When only the key is given to "enc", the IV is undefined. Print out 6620 an error message in this case. 6621 [Lutz Jaenicke] 6622 6623 *) Handle special case when X509_NAME is empty in X509 printing routines. 6624 [Steve Henson] 6625 6626 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 6627 positive and less than q. 6628 [Bodo Moeller] 6629 6630 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 6631 used: it isn't thread safe and the add_lock_callback should handle 6632 that itself. 6633 [Paul Rose <Paul.Rose@bridge.com>] 6634 6635 *) Verify that incoming data obeys the block size in 6636 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 6637 [Bodo Moeller] 6638 6639 *) Fix OAEP check. 6640 [Ulf M��ller, Bodo M��ller] 6641 6642 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 6643 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 6644 when fixing the server behaviour for backwards-compatible 'client 6645 hello' messages. (Note that the attack is impractical against 6646 SSL 3.0 and TLS 1.0 anyway because length and version checking 6647 means that the probability of guessing a valid ciphertext is 6648 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 6649 paper.) 6650 6651 Before 0.9.5, the countermeasure (hide the error by generating a 6652 random 'decryption result') did not work properly because 6653 ERR_clear_error() was missing, meaning that SSL_get_error() would 6654 detect the supposedly ignored error. 6655 6656 Both problems are now fixed. 6657 [Bodo Moeller] 6658 6659 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 6660 (previously it was 1024). 6661 [Bodo Moeller] 6662 6663 *) Fix for compatibility mode trust settings: ignore trust settings 6664 unless some valid trust or reject settings are present. 6665 [Steve Henson] 6666 6667 *) Fix for blowfish EVP: its a variable length cipher. 6668 [Steve Henson] 6669 6670 *) Fix various bugs related to DSA S/MIME verification. Handle missing 6671 parameters in DSA public key structures and return an error in the 6672 DSA routines if parameters are absent. 6673 [Steve Henson] 6674 6675 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 6676 in the current directory if neither $RANDFILE nor $HOME was set. 6677 RAND_file_name() in 0.9.6a returned NULL in this case. This has 6678 caused some confusion to Windows users who haven't defined $HOME. 6679 Thus RAND_file_name() is changed again: e_os.h can define a 6680 DEFAULT_HOME, which will be used if $HOME is not set. 6681 For Windows, we use "C:"; on other platforms, we still require 6682 environment variables. 6683 6684 *) Move 'if (!initialized) RAND_poll()' into regions protected by 6685 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 6686 having multiple threads call RAND_poll() concurrently. 6687 [Bodo Moeller] 6688 6689 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 6690 combination of a flag and a thread ID variable. 6691 Otherwise while one thread is in ssleay_rand_bytes (which sets the 6692 flag), *other* threads can enter ssleay_add_bytes without obeying 6693 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 6694 that they do not hold after the first thread unsets add_do_not_lock). 6695 [Bodo Moeller] 6696 6697 *) Change bctest again: '-x' expressions are not available in all 6698 versions of 'test'. 6699 [Bodo Moeller] 6700 6701 Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 6702 6703 *) Fix a couple of memory leaks in PKCS7_dataDecode() 6704 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 6705 6706 *) Change Configure and Makefiles to provide EXE_EXT, which will contain 6707 the default extension for executables, if any. Also, make the perl 6708 scripts that use symlink() to test if it really exists and use "cp" 6709 if it doesn't. All this made OpenSSL compilable and installable in 6710 CygWin. 6711 [Richard Levitte] 6712 6713 *) Fix for asn1_GetSequence() for indefinite length constructed data. 6714 If SEQUENCE is length is indefinite just set c->slen to the total 6715 amount of data available. 6716 [Steve Henson, reported by shige@FreeBSD.org] 6717 [This change does not apply to 0.9.7.] 6718 6719 *) Change bctest to avoid here-documents inside command substitution 6720 (workaround for FreeBSD /bin/sh bug). 6721 For compatibility with Ultrix, avoid shell functions (introduced 6722 in the bctest version that searches along $PATH). 6723 [Bodo Moeller] 6724 6725 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 6726 with des_encrypt() defined on some operating systems, like Solaris 6727 and UnixWare. 6728 [Richard Levitte] 6729 6730 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 6731 On the Importance of Eliminating Errors in Cryptographic 6732 Computations, J. Cryptology 14 (2001) 2, 101-119, 6733 http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 6734 [Ulf Moeller] 6735 6736 *) MIPS assembler BIGNUM division bug fix. 6737 [Andy Polyakov] 6738 6739 *) Disabled incorrect Alpha assembler code. 6740 [Richard Levitte] 6741 6742 *) Fix PKCS#7 decode routines so they correctly update the length 6743 after reading an EOC for the EXPLICIT tag. 6744 [Steve Henson] 6745 [This change does not apply to 0.9.7.] 6746 6747 *) Fix bug in PKCS#12 key generation routines. This was triggered 6748 if a 3DES key was generated with a 0 initial byte. Include 6749 PKCS12_BROKEN_KEYGEN compilation option to retain the old 6750 (but broken) behaviour. 6751 [Steve Henson] 6752 6753 *) Enhance bctest to search for a working bc along $PATH and print 6754 it when found. 6755 [Tim Rice <tim@multitalents.net> via Richard Levitte] 6756 6757 *) Fix memory leaks in err.c: free err_data string if necessary; 6758 don't write to the wrong index in ERR_set_error_data. 6759 [Bodo Moeller] 6760 6761 *) Implement ssl23_peek (analogous to ssl23_read), which previously 6762 did not exist. 6763 [Bodo Moeller] 6764 6765 *) Replace rdtsc with _emit statements for VC++ version 5. 6766 [Jeremy Cooper <jeremy@baymoo.org>] 6767 6768 *) Make it possible to reuse SSLv2 sessions. 6769 [Richard Levitte] 6770 6771 *) In copy_email() check for >= 0 as a return value for 6772 X509_NAME_get_index_by_NID() since 0 is a valid index. 6773 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 6774 6775 *) Avoid coredump with unsupported or invalid public keys by checking if 6776 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 6777 PKCS7_verify() fails with non detached data. 6778 [Steve Henson] 6779 6780 *) Don't use getenv in library functions when run as setuid/setgid. 6781 New function OPENSSL_issetugid(). 6782 [Ulf Moeller] 6783 6784 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 6785 due to incorrect handling of multi-threading: 6786 6787 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 6788 6789 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 6790 6791 3. Count how many times MemCheck_off() has been called so that 6792 nested use can be treated correctly. This also avoids 6793 inband-signalling in the previous code (which relied on the 6794 assumption that thread ID 0 is impossible). 6795 [Bodo Moeller] 6796 6797 *) Add "-rand" option also to s_client and s_server. 6798 [Lutz Jaenicke] 6799 6800 *) Fix CPU detection on Irix 6.x. 6801 [Kurt Hockenbury <khockenb@stevens-tech.edu> and 6802 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 6803 6804 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 6805 was empty. 6806 [Steve Henson] 6807 [This change does not apply to 0.9.7.] 6808 6809 *) Use the cached encoding of an X509_NAME structure rather than 6810 copying it. This is apparently the reason for the libsafe "errors" 6811 but the code is actually correct. 6812 [Steve Henson] 6813 6814 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 6815 Bleichenbacher's DSA attack. 6816 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 6817 to be set and top=0 forces the highest bit to be set; top=-1 is new 6818 and leaves the highest bit random. 6819 [Ulf Moeller, Bodo Moeller] 6820 6821 *) In the NCONF_...-based implementations for CONF_... queries 6822 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 6823 a temporary CONF structure with the data component set to NULL 6824 (which gives segmentation faults in lh_retrieve). 6825 Instead, use NULL for the CONF pointer in CONF_get_string and 6826 CONF_get_number (which may use environment variables) and directly 6827 return NULL from CONF_get_section. 6828 [Bodo Moeller] 6829 6830 *) Fix potential buffer overrun for EBCDIC. 6831 [Ulf Moeller] 6832 6833 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 6834 keyUsage if basicConstraints absent for a CA. 6835 [Steve Henson] 6836 6837 *) Make SMIME_write_PKCS7() write mail header values with a format that 6838 is more generally accepted (no spaces before the semicolon), since 6839 some programs can't parse those values properly otherwise. Also make 6840 sure BIO's that break lines after each write do not create invalid 6841 headers. 6842 [Richard Levitte] 6843 6844 *) Make the CRL encoding routines work with empty SEQUENCE OF. The 6845 macros previously used would not encode an empty SEQUENCE OF 6846 and break the signature. 6847 [Steve Henson] 6848 [This change does not apply to 0.9.7.] 6849 6850 *) Zero the premaster secret after deriving the master secret in 6851 DH ciphersuites. 6852 [Steve Henson] 6853 6854 *) Add some EVP_add_digest_alias registrations (as found in 6855 OpenSSL_add_all_digests()) to SSL_library_init() 6856 aka OpenSSL_add_ssl_algorithms(). This provides improved 6857 compatibility with peers using X.509 certificates 6858 with unconventional AlgorithmIdentifier OIDs. 6859 [Bodo Moeller] 6860 6861 *) Fix for Irix with NO_ASM. 6862 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 6863 6864 *) ./config script fixes. 6865 [Ulf Moeller, Richard Levitte] 6866 6867 *) Fix 'openssl passwd -1'. 6868 [Bodo Moeller] 6869 6870 *) Change PKCS12_key_gen_asc() so it can cope with non null 6871 terminated strings whose length is passed in the passlen 6872 parameter, for example from PEM callbacks. This was done 6873 by adding an extra length parameter to asc2uni(). 6874 [Steve Henson, reported by <oddissey@samsung.co.kr>] 6875 6876 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 6877 call failed, free the DSA structure. 6878 [Bodo Moeller] 6879 6880 *) Fix to uni2asc() to cope with zero length Unicode strings. 6881 These are present in some PKCS#12 files. 6882 [Steve Henson] 6883 6884 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 6885 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 6886 when writing a 32767 byte record. 6887 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 6888 6889 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 6890 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 6891 6892 (RSA objects have a reference count access to which is protected 6893 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 6894 so they are meant to be shared between threads.) 6895 [Bodo Moeller, Geoff Thorpe; original patch submitted by 6896 "Reddie, Steven" <Steven.Reddie@ca.com>] 6897 6898 *) Fix a deadlock in CRYPTO_mem_leaks(). 6899 [Bodo Moeller] 6900 6901 *) Use better test patterns in bntest. 6902 [Ulf M��ller] 6903 6904 *) rand_win.c fix for Borland C. 6905 [Ulf M��ller] 6906 6907 *) BN_rshift bugfix for n == 0. 6908 [Bodo Moeller] 6909 6910 *) Add a 'bctest' script that checks for some known 'bc' bugs 6911 so that 'make test' does not abort just because 'bc' is broken. 6912 [Bodo Moeller] 6913 6914 *) Store verify_result within SSL_SESSION also for client side to 6915 avoid potential security hole. (Re-used sessions on the client side 6916 always resulted in verify_result==X509_V_OK, not using the original 6917 result of the server certificate verification.) 6918 [Lutz Jaenicke] 6919 6920 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 6921 SSL3_RT_APPLICATION_DATA, return 0. 6922 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 6923 [Bodo Moeller] 6924 6925 *) Fix SSL_peek: 6926 Both ssl2_peek and ssl3_peek, which were totally broken in earlier 6927 releases, have been re-implemented by renaming the previous 6928 implementations of ssl2_read and ssl3_read to ssl2_read_internal 6929 and ssl3_read_internal, respectively, and adding 'peek' parameters 6930 to them. The new ssl[23]_{read,peek} functions are calls to 6931 ssl[23]_read_internal with the 'peek' flag set appropriately. 6932 A 'peek' parameter has also been added to ssl3_read_bytes, which 6933 does the actual work for ssl3_read_internal. 6934 [Bodo Moeller] 6935 6936 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 6937 the method-specific "init()" handler. Also clean up ex_data after 6938 calling the method-specific "finish()" handler. Previously, this was 6939 happening the other way round. 6940 [Geoff Thorpe] 6941 6942 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 6943 The previous value, 12, was not always sufficient for BN_mod_exp(). 6944 [Bodo Moeller] 6945 6946 *) Make sure that shared libraries get the internal name engine with 6947 the full version number and not just 0. This should mark the 6948 shared libraries as not backward compatible. Of course, this should 6949 be changed again when we can guarantee backward binary compatibility. 6950 [Richard Levitte] 6951 6952 *) Fix typo in get_cert_by_subject() in by_dir.c 6953 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 6954 6955 *) Rework the system to generate shared libraries: 6956 6957 - Make note of the expected extension for the shared libraries and 6958 if there is a need for symbolic links from for example libcrypto.so.0 6959 to libcrypto.so.0.9.7. There is extended info in Configure for 6960 that. 6961 6962 - Make as few rebuilds of the shared libraries as possible. 6963 6964 - Still avoid linking the OpenSSL programs with the shared libraries. 6965 6966 - When installing, install the shared libraries separately from the 6967 static ones. 6968 [Richard Levitte] 6969 6970 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 6971 6972 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 6973 and not in SSL_clear because the latter is also used by the 6974 accept/connect functions; previously, the settings made by 6975 SSL_set_read_ahead would be lost during the handshake. 6976 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 6977 6978 *) Correct util/mkdef.pl to be selective about disabled algorithms. 6979 Previously, it would create entries for disableed algorithms no 6980 matter what. 6981 [Richard Levitte] 6982 6983 *) Added several new manual pages for SSL_* function. 6984 [Lutz Jaenicke] 6985 6986 Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 6987 6988 *) In ssl23_get_client_hello, generate an error message when faced 6989 with an initial SSL 3.0/TLS record that is too small to contain the 6990 first two bytes of the ClientHello message, i.e. client_version. 6991 (Note that this is a pathologic case that probably has never happened 6992 in real life.) The previous approach was to use the version number 6993 from the record header as a substitute; but our protocol choice 6994 should not depend on that one because it is not authenticated 6995 by the Finished messages. 6996 [Bodo Moeller] 6997 6998 *) More robust randomness gathering functions for Windows. 6999 [Jeffrey Altman <jaltman@columbia.edu>] 7000 7001 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 7002 not set then we don't setup the error code for issuer check errors 7003 to avoid possibly overwriting other errors which the callback does 7004 handle. If an application does set the flag then we assume it knows 7005 what it is doing and can handle the new informational codes 7006 appropriately. 7007 [Steve Henson] 7008 7009 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 7010 a general "ANY" type, as such it should be able to decode anything 7011 including tagged types. However it didn't check the class so it would 7012 wrongly interpret tagged types in the same way as their universal 7013 counterpart and unknown types were just rejected. Changed so that the 7014 tagged and unknown types are handled in the same way as a SEQUENCE: 7015 that is the encoding is stored intact. There is also a new type 7016 "V_ASN1_OTHER" which is used when the class is not universal, in this 7017 case we have no idea what the actual type is so we just lump them all 7018 together. 7019 [Steve Henson] 7020 7021 *) On VMS, stdout may very well lead to a file that is written to 7022 in a record-oriented fashion. That means that every write() will 7023 write a separate record, which will be read separately by the 7024 programs trying to read from it. This can be very confusing. 7025 7026 The solution is to put a BIO filter in the way that will buffer 7027 text until a linefeed is reached, and then write everything a 7028 line at a time, so every record written will be an actual line, 7029 not chunks of lines and not (usually doesn't happen, but I've 7030 seen it once) several lines in one record. BIO_f_linebuffer() is 7031 the answer. 7032 7033 Currently, it's a VMS-only method, because that's where it has 7034 been tested well enough. 7035 [Richard Levitte] 7036 7037 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 7038 it can return incorrect results. 7039 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 7040 but it was in 0.9.6-beta[12].) 7041 [Bodo Moeller] 7042 7043 *) Disable the check for content being present when verifying detached 7044 signatures in pk7_smime.c. Some versions of Netscape (wrongly) 7045 include zero length content when signing messages. 7046 [Steve Henson] 7047 7048 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 7049 BIO_ctrl (for BIO pairs). 7050 [Bodo M��ller] 7051 7052 *) Add DSO method for VMS. 7053 [Richard Levitte] 7054 7055 *) Bug fix: Montgomery multiplication could produce results with the 7056 wrong sign. 7057 [Ulf M��ller] 7058 7059 *) Add RPM specification openssl.spec and modify it to build three 7060 packages. The default package contains applications, application 7061 documentation and run-time libraries. The devel package contains 7062 include files, static libraries and function documentation. The 7063 doc package contains the contents of the doc directory. The original 7064 openssl.spec was provided by Damien Miller <djm@mindrot.org>. 7065 [Richard Levitte] 7066 7067 *) Add a large number of documentation files for many SSL routines. 7068 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 7069 7070 *) Add a configuration entry for Sony News 4. 7071 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 7072 7073 *) Don't set the two most significant bits to one when generating a 7074 random number < q in the DSA library. 7075 [Ulf M��ller] 7076 7077 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 7078 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 7079 the underlying transport is blocking) if a handshake took place. 7080 (The default behaviour is needed by applications such as s_client 7081 and s_server that use select() to determine when to use SSL_read; 7082 but for applications that know in advance when to expect data, it 7083 just makes things more complicated.) 7084 [Bodo Moeller] 7085 7086 *) Add RAND_egd_bytes(), which gives control over the number of bytes read 7087 from EGD. 7088 [Ben Laurie] 7089 7090 *) Add a few more EBCDIC conditionals that make `req' and `x509' 7091 work better on such systems. 7092 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 7093 7094 *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 7095 Update PKCS12_parse() so it copies the friendlyName and the 7096 keyid to the certificates aux info. 7097 [Steve Henson] 7098 7099 *) Fix bug in PKCS7_verify() which caused an infinite loop 7100 if there was more than one signature. 7101 [Sven Uszpelkat <su@celocom.de>] 7102 7103 *) Major change in util/mkdef.pl to include extra information 7104 about each symbol, as well as presentig variables as well 7105 as functions. This change means that there's n more need 7106 to rebuild the .num files when some algorithms are excluded. 7107 [Richard Levitte] 7108 7109 *) Allow the verify time to be set by an application, 7110 rather than always using the current time. 7111 [Steve Henson] 7112 7113 *) Phase 2 verify code reorganisation. The certificate 7114 verify code now looks up an issuer certificate by a 7115 number of criteria: subject name, authority key id 7116 and key usage. It also verifies self signed certificates 7117 by the same criteria. The main comparison function is 7118 X509_check_issued() which performs these checks. 7119 7120 Lot of changes were necessary in order to support this 7121 without completely rewriting the lookup code. 7122 7123 Authority and subject key identifier are now cached. 7124 7125 The LHASH 'certs' is X509_STORE has now been replaced 7126 by a STACK_OF(X509_OBJECT). This is mainly because an 7127 LHASH can't store or retrieve multiple objects with 7128 the same hash value. 7129 7130 As a result various functions (which were all internal 7131 use only) have changed to handle the new X509_STORE 7132 structure. This will break anything that messed round 7133 with X509_STORE internally. 7134 7135 The functions X509_STORE_add_cert() now checks for an 7136 exact match, rather than just subject name. 7137 7138 The X509_STORE API doesn't directly support the retrieval 7139 of multiple certificates matching a given criteria, however 7140 this can be worked round by performing a lookup first 7141 (which will fill the cache with candidate certificates) 7142 and then examining the cache for matches. This is probably 7143 the best we can do without throwing out X509_LOOKUP 7144 entirely (maybe later...). 7145 7146 The X509_VERIFY_CTX structure has been enhanced considerably. 7147 7148 All certificate lookup operations now go via a get_issuer() 7149 callback. Although this currently uses an X509_STORE it 7150 can be replaced by custom lookups. This is a simple way 7151 to bypass the X509_STORE hackery necessary to make this 7152 work and makes it possible to use more efficient techniques 7153 in future. A very simple version which uses a simple 7154 STACK for its trusted certificate store is also provided 7155 using X509_STORE_CTX_trusted_stack(). 7156 7157 The verify_cb() and verify() callbacks now have equivalents 7158 in the X509_STORE_CTX structure. 7159 7160 X509_STORE_CTX also has a 'flags' field which can be used 7161 to customise the verify behaviour. 7162 [Steve Henson] 7163 7164 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 7165 excludes S/MIME capabilities. 7166 [Steve Henson] 7167 7168 *) When a certificate request is read in keep a copy of the 7169 original encoding of the signed data and use it when outputing 7170 again. Signatures then use the original encoding rather than 7171 a decoded, encoded version which may cause problems if the 7172 request is improperly encoded. 7173 [Steve Henson] 7174 7175 *) For consistency with other BIO_puts implementations, call 7176 buffer_write(b, ...) directly in buffer_puts instead of calling 7177 BIO_write(b, ...). 7178 7179 In BIO_puts, increment b->num_write as in BIO_write. 7180 [Peter.Sylvester@EdelWeb.fr] 7181 7182 *) Fix BN_mul_word for the case where the word is 0. (We have to use 7183 BN_zero, we may not return a BIGNUM with an array consisting of 7184 words set to zero.) 7185 [Bodo Moeller] 7186 7187 *) Avoid calling abort() from within the library when problems are 7188 detected, except if preprocessor symbols have been defined 7189 (such as REF_CHECK, BN_DEBUG etc.). 7190 [Bodo Moeller] 7191 7192 *) New openssl application 'rsautl'. This utility can be 7193 used for low level RSA operations. DER public key 7194 BIO/fp routines also added. 7195 [Steve Henson] 7196 7197 *) New Configure entry and patches for compiling on QNX 4. 7198 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 7199 7200 *) A demo state-machine implementation was sponsored by 7201 Nuron (http://www.nuron.com/) and is now available in 7202 demos/state_machine. 7203 [Ben Laurie] 7204 7205 *) New options added to the 'dgst' utility for signature 7206 generation and verification. 7207 [Steve Henson] 7208 7209 *) Unrecognized PKCS#7 content types are now handled via a 7210 catch all ASN1_TYPE structure. This allows unsupported 7211 types to be stored as a "blob" and an application can 7212 encode and decode it manually. 7213 [Steve Henson] 7214 7215 *) Fix various signed/unsigned issues to make a_strex.c 7216 compile under VC++. 7217 [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 7218 7219 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 7220 length if passed a buffer. ASN1_INTEGER_to_BN failed 7221 if passed a NULL BN and its argument was negative. 7222 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 7223 7224 *) Modification to PKCS#7 encoding routines to output definite 7225 length encoding. Since currently the whole structures are in 7226 memory there's not real point in using indefinite length 7227 constructed encoding. However if OpenSSL is compiled with 7228 the flag PKCS7_INDEFINITE_ENCODING the old form is used. 7229 [Steve Henson] 7230 7231 *) Added BIO_vprintf() and BIO_vsnprintf(). 7232 [Richard Levitte] 7233 7234 *) Added more prefixes to parse for in the the strings written 7235 through a logging bio, to cover all the levels that are available 7236 through syslog. The prefixes are now: 7237 7238 PANIC, EMERG, EMR => LOG_EMERG 7239 ALERT, ALR => LOG_ALERT 7240 CRIT, CRI => LOG_CRIT 7241 ERROR, ERR => LOG_ERR 7242 WARNING, WARN, WAR => LOG_WARNING 7243 NOTICE, NOTE, NOT => LOG_NOTICE 7244 INFO, INF => LOG_INFO 7245 DEBUG, DBG => LOG_DEBUG 7246 7247 and as before, if none of those prefixes are present at the 7248 beginning of the string, LOG_ERR is chosen. 7249 7250 On Win32, the LOG_* levels are mapped according to this: 7251 7252 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 7253 LOG_WARNING => EVENTLOG_WARNING_TYPE 7254 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 7255 7256 [Richard Levitte] 7257 7258 *) Made it possible to reconfigure with just the configuration 7259 argument "reconf" or "reconfigure". The command line arguments 7260 are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 7261 and are retrieved from there when reconfiguring. 7262 [Richard Levitte] 7263 7264 *) MD4 implemented. 7265 [Assar Westerlund <assar@sics.se>, Richard Levitte] 7266 7267 *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 7268 [Richard Levitte] 7269 7270 *) The obj_dat.pl script was messing up the sorting of object 7271 names. The reason was that it compared the quoted version 7272 of strings as a result "OCSP" > "OCSP Signing" because 7273 " > SPACE. Changed script to store unquoted versions of 7274 names and add quotes on output. It was also omitting some 7275 names from the lookup table if they were given a default 7276 value (that is if SN is missing it is given the same 7277 value as LN and vice versa), these are now added on the 7278 grounds that if an object has a name we should be able to 7279 look it up. Finally added warning output when duplicate 7280 short or long names are found. 7281 [Steve Henson] 7282 7283 *) Changes needed for Tandem NSK. 7284 [Scott Uroff <scott@xypro.com>] 7285 7286 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 7287 RSA_padding_check_SSLv23(), special padding was never detected 7288 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 7289 version rollback attacks was not effective. 7290 7291 In s23_clnt.c, don't use special rollback-attack detection padding 7292 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 7293 client; similarly, in s23_srvr.c, don't do the rollback check if 7294 SSL 2.0 is the only protocol enabled in the server. 7295 [Bodo Moeller] 7296 7297 *) Make it possible to get hexdumps of unprintable data with 'openssl 7298 asn1parse'. By implication, the functions ASN1_parse_dump() and 7299 BIO_dump_indent() are added. 7300 [Richard Levitte] 7301 7302 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 7303 these print out strings and name structures based on various 7304 flags including RFC2253 support and proper handling of 7305 multibyte characters. Added options to the 'x509' utility 7306 to allow the various flags to be set. 7307 [Steve Henson] 7308 7309 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 7310 Also change the functions X509_cmp_current_time() and 7311 X509_gmtime_adj() work with an ASN1_TIME structure, 7312 this will enable certificates using GeneralizedTime in validity 7313 dates to be checked. 7314 [Steve Henson] 7315 7316 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 7317 negative public key encodings) on by default, 7318 NO_NEG_PUBKEY_BUG can be set to disable it. 7319 [Steve Henson] 7320 7321 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 7322 content octets. An i2c_ASN1_OBJECT is unnecessary because 7323 the encoding can be trivially obtained from the structure. 7324 [Steve Henson] 7325 7326 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 7327 not read locks (CRYPTO_r_[un]lock). 7328 [Bodo Moeller] 7329 7330 *) A first attempt at creating official support for shared 7331 libraries through configuration. I've kept it so the 7332 default is static libraries only, and the OpenSSL programs 7333 are always statically linked for now, but there are 7334 preparations for dynamic linking in place. 7335 This has been tested on Linux and Tru64. 7336 [Richard Levitte] 7337 7338 *) Randomness polling function for Win9x, as described in: 7339 Peter Gutmann, Software Generation of Practically Strong 7340 Random Numbers. 7341 [Ulf M��ller] 7342 7343 *) Fix so PRNG is seeded in req if using an already existing 7344 DSA key. 7345 [Steve Henson] 7346 7347 *) New options to smime application. -inform and -outform 7348 allow alternative formats for the S/MIME message including 7349 PEM and DER. The -content option allows the content to be 7350 specified separately. This should allow things like Netscape 7351 form signing output easier to verify. 7352 [Steve Henson] 7353 7354 *) Fix the ASN1 encoding of tags using the 'long form'. 7355 [Steve Henson] 7356 7357 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 7358 STRING types. These convert content octets to and from the 7359 underlying type. The actual tag and length octets are 7360 already assumed to have been read in and checked. These 7361 are needed because all other string types have virtually 7362 identical handling apart from the tag. By having versions 7363 of the ASN1 functions that just operate on content octets 7364 IMPLICIT tagging can be handled properly. It also allows 7365 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 7366 and ASN1_INTEGER are identical apart from the tag. 7367 [Steve Henson] 7368 7369 *) Change the handling of OID objects as follows: 7370 7371 - New object identifiers are inserted in objects.txt, following 7372 the syntax given in objects.README. 7373 - objects.pl is used to process obj_mac.num and create a new 7374 obj_mac.h. 7375 - obj_dat.pl is used to create a new obj_dat.h, using the data in 7376 obj_mac.h. 7377 7378 This is currently kind of a hack, and the perl code in objects.pl 7379 isn't very elegant, but it works as I intended. The simplest way 7380 to check that it worked correctly is to look in obj_dat.h and 7381 check the array nid_objs and make sure the objects haven't moved 7382 around (this is important!). Additions are OK, as well as 7383 consistent name changes. 7384 [Richard Levitte] 7385 7386 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 7387 [Bodo Moeller] 7388 7389 *) Addition of the command line parameter '-rand file' to 'openssl req'. 7390 The given file adds to whatever has already been seeded into the 7391 random pool through the RANDFILE configuration file option or 7392 environment variable, or the default random state file. 7393 [Richard Levitte] 7394 7395 *) mkstack.pl now sorts each macro group into lexical order. 7396 Previously the output order depended on the order the files 7397 appeared in the directory, resulting in needless rewriting 7398 of safestack.h . 7399 [Steve Henson] 7400 7401 *) Patches to make OpenSSL compile under Win32 again. Mostly 7402 work arounds for the VC++ problem that it treats func() as 7403 func(void). Also stripped out the parts of mkdef.pl that 7404 added extra typesafe functions: these no longer exist. 7405 [Steve Henson] 7406 7407 *) Reorganisation of the stack code. The macros are now all 7408 collected in safestack.h . Each macro is defined in terms of 7409 a "stack macro" of the form SKM_<name>(type, a, b). The 7410 DEBUG_SAFESTACK is now handled in terms of function casts, 7411 this has the advantage of retaining type safety without the 7412 use of additional functions. If DEBUG_SAFESTACK is not defined 7413 then the non typesafe macros are used instead. Also modified the 7414 mkstack.pl script to handle the new form. Needs testing to see 7415 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 7416 the default if no major problems. Similar behaviour for ASN1_SET_OF 7417 and PKCS12_STACK_OF. 7418 [Steve Henson] 7419 7420 *) When some versions of IIS use the 'NET' form of private key the 7421 key derivation algorithm is different. Normally MD5(password) is 7422 used as a 128 bit RC4 key. In the modified case 7423 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 7424 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 7425 as the old Netscape_RSA functions except they have an additional 7426 'sgckey' parameter which uses the modified algorithm. Also added 7427 an -sgckey command line option to the rsa utility. Thanks to 7428 Adrian Peck <bertie@ncipher.com> for posting details of the modified 7429 algorithm to openssl-dev. 7430 [Steve Henson] 7431 7432 *) The evp_local.h macros were using 'c.##kname' which resulted in 7433 invalid expansion on some systems (SCO 5.0.5 for example). 7434 Corrected to 'c.kname'. 7435 [Phillip Porch <root@theporch.com>] 7436 7437 *) New X509_get1_email() and X509_REQ_get1_email() functions that return 7438 a STACK of email addresses from a certificate or request, these look 7439 in the subject name and the subject alternative name extensions and 7440 omit any duplicate addresses. 7441 [Steve Henson] 7442 7443 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 7444 This makes DSA verification about 2 % faster. 7445 [Bodo Moeller] 7446 7447 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 7448 (meaning that now 2^5 values will be precomputed, which is only 4 KB 7449 plus overhead for 1024 bit moduli). 7450 This makes exponentiations about 0.5 % faster for 1024 bit 7451 exponents (as measured by "openssl speed rsa2048"). 7452 [Bodo Moeller] 7453 7454 *) Rename memory handling macros to avoid conflicts with other 7455 software: 7456 Malloc => OPENSSL_malloc 7457 Malloc_locked => OPENSSL_malloc_locked 7458 Realloc => OPENSSL_realloc 7459 Free => OPENSSL_free 7460 [Richard Levitte] 7461 7462 *) New function BN_mod_exp_mont_word for small bases (roughly 15% 7463 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 7464 [Bodo Moeller] 7465 7466 *) CygWin32 support. 7467 [John Jarvie <jjarvie@newsguy.com>] 7468 7469 *) The type-safe stack code has been rejigged. It is now only compiled 7470 in when OpenSSL is configured with the DEBUG_SAFESTACK option and 7471 by default all type-specific stack functions are "#define"d back to 7472 standard stack functions. This results in more streamlined output 7473 but retains the type-safety checking possibilities of the original 7474 approach. 7475 [Geoff Thorpe] 7476 7477 *) The STACK code has been cleaned up, and certain type declarations 7478 that didn't make a lot of sense have been brought in line. This has 7479 also involved a cleanup of sorts in safestack.h to more correctly 7480 map type-safe stack functions onto their plain stack counterparts. 7481 This work has also resulted in a variety of "const"ifications of 7482 lots of the code, especially "_cmp" operations which should normally 7483 be prototyped with "const" parameters anyway. 7484 [Geoff Thorpe] 7485 7486 *) When generating bytes for the first time in md_rand.c, 'stir the pool' 7487 by seeding with STATE_SIZE dummy bytes (with zero entropy count). 7488 (The PRNG state consists of two parts, the large pool 'state' and 'md', 7489 where all of 'md' is used each time the PRNG is used, but 'state' 7490 is used only indexed by a cyclic counter. As entropy may not be 7491 well distributed from the beginning, 'md' is important as a 7492 chaining variable. However, the output function chains only half 7493 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 7494 all of 'md', and seeding with STATE_SIZE dummy bytes will result 7495 in all of 'state' being rewritten, with the new values depending 7496 on virtually all of 'md'. This overcomes the 80 bit limitation.) 7497 [Bodo Moeller] 7498 7499 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 7500 the handshake is continued after ssl_verify_cert_chain(); 7501 otherwise, if SSL_VERIFY_NONE is set, remaining error codes 7502 can lead to 'unexplainable' connection aborts later. 7503 [Bodo Moeller; problem tracked down by Lutz Jaenicke] 7504 7505 *) Major EVP API cipher revision. 7506 Add hooks for extra EVP features. This allows various cipher 7507 parameters to be set in the EVP interface. Support added for variable 7508 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 7509 setting of RC2 and RC5 parameters. 7510 7511 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 7512 ciphers. 7513 7514 Remove lots of duplicated code from the EVP library. For example *every* 7515 cipher init() function handles the 'iv' in the same way according to the 7516 cipher mode. They also all do nothing if the 'key' parameter is NULL and 7517 for CFB and OFB modes they zero ctx->num. 7518 7519 New functionality allows removal of S/MIME code RC2 hack. 7520 7521 Most of the routines have the same form and so can be declared in terms 7522 of macros. 7523 7524 By shifting this to the top level EVP_CipherInit() it can be removed from 7525 all individual ciphers. If the cipher wants to handle IVs or keys 7526 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 7527 flags. 7528 7529 Change lots of functions like EVP_EncryptUpdate() to now return a 7530 value: although software versions of the algorithms cannot fail 7531 any installed hardware versions can. 7532 [Steve Henson] 7533 7534 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 7535 this option is set, tolerate broken clients that send the negotiated 7536 protocol version number instead of the requested protocol version 7537 number. 7538 [Bodo Moeller] 7539 7540 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 7541 i.e. non-zero for export ciphersuites, zero otherwise. 7542 Previous versions had this flag inverted, inconsistent with 7543 rsa_tmp_cb (..._TMP_RSA_CB). 7544 [Bodo Moeller; problem reported by Amit Chopra] 7545 7546 *) Add missing DSA library text string. Work around for some IIS 7547 key files with invalid SEQUENCE encoding. 7548 [Steve Henson] 7549 7550 *) Add a document (doc/standards.txt) that list all kinds of standards 7551 and so on that are implemented in OpenSSL. 7552 [Richard Levitte] 7553 7554 *) Enhance c_rehash script. Old version would mishandle certificates 7555 with the same subject name hash and wouldn't handle CRLs at all. 7556 Added -fingerprint option to crl utility, to support new c_rehash 7557 features. 7558 [Steve Henson] 7559 7560 *) Eliminate non-ANSI declarations in crypto.h and stack.h. 7561 [Ulf M��ller] 7562 7563 *) Fix for SSL server purpose checking. Server checking was 7564 rejecting certificates which had extended key usage present 7565 but no ssl client purpose. 7566 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 7567 7568 *) Make PKCS#12 code work with no password. The PKCS#12 spec 7569 is a little unclear about how a blank password is handled. 7570 Since the password in encoded as a BMPString with terminating 7571 double NULL a zero length password would end up as just the 7572 double NULL. However no password at all is different and is 7573 handled differently in the PKCS#12 key generation code. NS 7574 treats a blank password as zero length. MSIE treats it as no 7575 password on export: but it will try both on import. We now do 7576 the same: PKCS12_parse() tries zero length and no password if 7577 the password is set to "" or NULL (NULL is now a valid password: 7578 it wasn't before) as does the pkcs12 application. 7579 [Steve Henson] 7580 7581 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 7582 perror when PEM_read_bio_X509_REQ fails, the error message must 7583 be obtained from the error queue. 7584 [Bodo Moeller] 7585 7586 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 7587 it in ERR_remove_state if appropriate, and change ERR_get_state 7588 accordingly to avoid race conditions (this is necessary because 7589 thread_hash is no longer constant once set). 7590 [Bodo Moeller] 7591 7592 *) Bugfix for linux-elf makefile.one. 7593 [Ulf M��ller] 7594 7595 *) RSA_get_default_method() will now cause a default 7596 RSA_METHOD to be chosen if one doesn't exist already. 7597 Previously this was only set during a call to RSA_new() 7598 or RSA_new_method(NULL) meaning it was possible for 7599 RSA_get_default_method() to return NULL. 7600 [Geoff Thorpe] 7601 7602 *) Added native name translation to the existing DSO code 7603 that will convert (if the flag to do so is set) filenames 7604 that are sufficiently small and have no path information 7605 into a canonical native form. Eg. "blah" converted to 7606 "libblah.so" or "blah.dll" etc. 7607 [Geoff Thorpe] 7608 7609 *) New function ERR_error_string_n(e, buf, len) which is like 7610 ERR_error_string(e, buf), but writes at most 'len' bytes 7611 including the 0 terminator. For ERR_error_string_n, 'buf' 7612 may not be NULL. 7613 [Damien Miller <djm@mindrot.org>, Bodo Moeller] 7614 7615 *) CONF library reworked to become more general. A new CONF 7616 configuration file reader "class" is implemented as well as a 7617 new functions (NCONF_*, for "New CONF") to handle it. The now 7618 old CONF_* functions are still there, but are reimplemented to 7619 work in terms of the new functions. Also, a set of functions 7620 to handle the internal storage of the configuration data is 7621 provided to make it easier to write new configuration file 7622 reader "classes" (I can definitely see something reading a 7623 configuration file in XML format, for example), called _CONF_*, 7624 or "the configuration storage API"... 7625 7626 The new configuration file reading functions are: 7627 7628 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 7629 NCONF_get_section, NCONF_get_string, NCONF_get_numbre 7630 7631 NCONF_default, NCONF_WIN32 7632 7633 NCONF_dump_fp, NCONF_dump_bio 7634 7635 NCONF_default and NCONF_WIN32 are method (or "class") choosers, 7636 NCONF_new creates a new CONF object. This works in the same way 7637 as other interfaces in OpenSSL, like the BIO interface. 7638 NCONF_dump_* dump the internal storage of the configuration file, 7639 which is useful for debugging. All other functions take the same 7640 arguments as the old CONF_* functions wth the exception of the 7641 first that must be a `CONF *' instead of a `LHASH *'. 7642 7643 To make it easer to use the new classes with the old CONF_* functions, 7644 the function CONF_set_default_method is provided. 7645 [Richard Levitte] 7646 7647 *) Add '-tls1' option to 'openssl ciphers', which was already 7648 mentioned in the documentation but had not been implemented. 7649 (This option is not yet really useful because even the additional 7650 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 7651 [Bodo Moeller] 7652 7653 *) Initial DSO code added into libcrypto for letting OpenSSL (and 7654 OpenSSL-based applications) load shared libraries and bind to 7655 them in a portable way. 7656 [Geoff Thorpe, with contributions from Richard Levitte] 7657 7658 Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 7659 7660 *) Make sure _lrotl and _lrotr are only used with MSVC. 7661 7662 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 7663 (the default implementation of RAND_status). 7664 7665 *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 7666 to '-clrext' (= clear extensions), as intended and documented. 7667 [Bodo Moeller; inconsistency pointed out by Michael Attili 7668 <attili@amaxo.com>] 7669 7670 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 7671 was larger than the MD block size. 7672 [Steve Henson, pointed out by Yost William <YostW@tce.com>] 7673 7674 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 7675 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 7676 using the passed key: if the passed key was a private key the result 7677 of X509_print(), for example, would be to print out all the private key 7678 components. 7679 [Steve Henson] 7680 7681 *) des_quad_cksum() byte order bug fix. 7682 [Ulf M��ller, using the problem description in krb4-0.9.7, where 7683 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 7684 7685 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 7686 discouraged. 7687 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 7688 7689 *) For easily testing in shell scripts whether some command 7690 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 7691 returns with exit code 0 iff no command of the given name is available. 7692 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 7693 the output goes to stdout and nothing is printed to stderr. 7694 Additional arguments are always ignored. 7695 7696 Since for each cipher there is a command of the same name, 7697 the 'no-cipher' compilation switches can be tested this way. 7698 7699 ('openssl no-XXX' is not able to detect pseudo-commands such 7700 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 7701 [Bodo Moeller] 7702 7703 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 7704 [Bodo Moeller] 7705 7706 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 7707 is set; it will be thrown away anyway because each handshake creates 7708 its own key. 7709 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 7710 to parameters -- in previous versions (since OpenSSL 0.9.3) the 7711 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 7712 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 7713 [Bodo Moeller] 7714 7715 *) New s_client option -ign_eof: EOF at stdin is ignored, and 7716 'Q' and 'R' lose their special meanings (quit/renegotiate). 7717 This is part of what -quiet does; unlike -quiet, -ign_eof 7718 does not suppress any output. 7719 [Richard Levitte] 7720 7721 *) Add compatibility options to the purpose and trust code. The 7722 purpose X509_PURPOSE_ANY is "any purpose" which automatically 7723 accepts a certificate or CA, this was the previous behaviour, 7724 with all the associated security issues. 7725 7726 X509_TRUST_COMPAT is the old trust behaviour: only and 7727 automatically trust self signed roots in certificate store. A 7728 new trust setting X509_TRUST_DEFAULT is used to specify that 7729 a purpose has no associated trust setting and it should instead 7730 use the value in the default purpose. 7731 [Steve Henson] 7732 7733 *) Fix the PKCS#8 DSA private key code so it decodes keys again 7734 and fix a memory leak. 7735 [Steve Henson] 7736 7737 *) In util/mkerr.pl (which implements 'make errors'), preserve 7738 reason strings from the previous version of the .c file, as 7739 the default to have only downcase letters (and digits) in 7740 automatically generated reasons codes is not always appropriate. 7741 [Bodo Moeller] 7742 7743 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 7744 using strerror. Previously, ERR_reason_error_string() returned 7745 library names as reason strings for SYSerr; but SYSerr is a special 7746 case where small numbers are errno values, not library numbers. 7747 [Bodo Moeller] 7748 7749 *) Add '-dsaparam' option to 'openssl dhparam' application. This 7750 converts DSA parameters into DH parameters. (When creating parameters, 7751 DSA_generate_parameters is used.) 7752 [Bodo Moeller] 7753 7754 *) Include 'length' (recommended exponent length) in C code generated 7755 by 'openssl dhparam -C'. 7756 [Bodo Moeller] 7757 7758 *) The second argument to set_label in perlasm was already being used 7759 so couldn't be used as a "file scope" flag. Moved to third argument 7760 which was free. 7761 [Steve Henson] 7762 7763 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 7764 instead of RAND_bytes for encryption IVs and salts. 7765 [Bodo Moeller] 7766 7767 *) Include RAND_status() into RAND_METHOD instead of implementing 7768 it only for md_rand.c Otherwise replacing the PRNG by calling 7769 RAND_set_rand_method would be impossible. 7770 [Bodo Moeller] 7771 7772 *) Don't let DSA_generate_key() enter an infinite loop if the random 7773 number generation fails. 7774 [Bodo Moeller] 7775 7776 *) New 'rand' application for creating pseudo-random output. 7777 [Bodo Moeller] 7778 7779 *) Added configuration support for Linux/IA64 7780 [Rolf Haberrecker <rolf@suse.de>] 7781 7782 *) Assembler module support for Mingw32. 7783 [Ulf M��ller] 7784 7785 *) Shared library support for HPUX (in shlib/). 7786 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 7787 7788 *) Shared library support for Solaris gcc. 7789 [Lutz Behnke <behnke@trustcenter.de>] 7790 7791 Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 7792 7793 *) PKCS7_encrypt() was adding text MIME headers twice because they 7794 were added manually and by SMIME_crlf_copy(). 7795 [Steve Henson] 7796 7797 *) In bntest.c don't call BN_rand with zero bits argument. 7798 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 7799 7800 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 7801 case was implemented. This caused BN_div_recp() to fail occasionally. 7802 [Ulf M��ller] 7803 7804 *) Add an optional second argument to the set_label() in the perl 7805 assembly language builder. If this argument exists and is set 7806 to 1 it signals that the assembler should use a symbol whose 7807 scope is the entire file, not just the current function. This 7808 is needed with MASM which uses the format label:: for this scope. 7809 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 7810 7811 *) Change the ASN1 types so they are typedefs by default. Before 7812 almost all types were #define'd to ASN1_STRING which was causing 7813 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 7814 for example. 7815 [Steve Henson] 7816 7817 *) Change names of new functions to the new get1/get0 naming 7818 convention: After 'get1', the caller owns a reference count 7819 and has to call ..._free; 'get0' returns a pointer to some 7820 data structure without incrementing reference counters. 7821 (Some of the existing 'get' functions increment a reference 7822 counter, some don't.) 7823 Similarly, 'set1' and 'add1' functions increase reference 7824 counters or duplicate objects. 7825 [Steve Henson] 7826 7827 *) Allow for the possibility of temp RSA key generation failure: 7828 the code used to assume it always worked and crashed on failure. 7829 [Steve Henson] 7830 7831 *) Fix potential buffer overrun problem in BIO_printf(). 7832 [Ulf M��ller, using public domain code by Patrick Powell; problem 7833 pointed out by David Sacerdote <das33@cornell.edu>] 7834 7835 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 7836 RAND_egd() and RAND_status(). In the command line application, 7837 the EGD socket can be specified like a seed file using RANDFILE 7838 or -rand. 7839 [Ulf M��ller] 7840 7841 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 7842 Some CAs (e.g. Verisign) distribute certificates in this form. 7843 [Steve Henson] 7844 7845 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 7846 list to exclude them. This means that no special compilation option 7847 is needed to use anonymous DH: it just needs to be included in the 7848 cipher list. 7849 [Steve Henson] 7850 7851 *) Change the EVP_MD_CTX_type macro so its meaning consistent with 7852 EVP_MD_type. The old functionality is available in a new macro called 7853 EVP_MD_md(). Change code that uses it and update docs. 7854 [Steve Henson] 7855 7856 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 7857 where the 'void *' argument is replaced by a function pointer argument. 7858 Previously 'void *' was abused to point to functions, which works on 7859 many platforms, but is not correct. As these functions are usually 7860 called by macros defined in OpenSSL header files, most source code 7861 should work without changes. 7862 [Richard Levitte] 7863 7864 *) <openssl/opensslconf.h> (which is created by Configure) now contains 7865 sections with information on -D... compiler switches used for 7866 compiling the library so that applications can see them. To enable 7867 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 7868 must be defined. E.g., 7869 #define OPENSSL_ALGORITHM_DEFINES 7870 #include <openssl/opensslconf.h> 7871 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 7872 [Richard Levitte, Ulf and Bodo M��ller] 7873 7874 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 7875 record layer. 7876 [Bodo Moeller] 7877 7878 *) Change the 'other' type in certificate aux info to a STACK_OF 7879 X509_ALGOR. Although not an AlgorithmIdentifier as such it has 7880 the required ASN1 format: arbitrary types determined by an OID. 7881 [Steve Henson] 7882 7883 *) Add some PEM_write_X509_REQ_NEW() functions and a command line 7884 argument to 'req'. This is not because the function is newer or 7885 better than others it just uses the work 'NEW' in the certificate 7886 request header lines. Some software needs this. 7887 [Steve Henson] 7888 7889 *) Reorganise password command line arguments: now passwords can be 7890 obtained from various sources. Delete the PEM_cb function and make 7891 it the default behaviour: i.e. if the callback is NULL and the 7892 usrdata argument is not NULL interpret it as a null terminated pass 7893 phrase. If usrdata and the callback are NULL then the pass phrase 7894 is prompted for as usual. 7895 [Steve Henson] 7896 7897 *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 7898 the support is automatically enabled. The resulting binaries will 7899 autodetect the card and use it if present. 7900 [Ben Laurie and Compaq Inc.] 7901 7902 *) Work around for Netscape hang bug. This sends certificate request 7903 and server done in one record. Since this is perfectly legal in the 7904 SSL/TLS protocol it isn't a "bug" option and is on by default. See 7905 the bugs/SSLv3 entry for more info. 7906 [Steve Henson] 7907 7908 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 7909 [Andy Polyakov] 7910 7911 *) Add -rand argument to smime and pkcs12 applications and read/write 7912 of seed file. 7913 [Steve Henson] 7914 7915 *) New 'passwd' tool for crypt(3) and apr1 password hashes. 7916 [Bodo Moeller] 7917 7918 *) Add command line password options to the remaining applications. 7919 [Steve Henson] 7920 7921 *) Bug fix for BN_div_recp() for numerators with an even number of 7922 bits. 7923 [Ulf M��ller] 7924 7925 *) More tests in bntest.c, and changed test_bn output. 7926 [Ulf M��ller] 7927 7928 *) ./config recognizes MacOS X now. 7929 [Andy Polyakov] 7930 7931 *) Bug fix for BN_div() when the first words of num and divsor are 7932 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 7933 [Ulf M��ller] 7934 7935 *) Add support for various broken PKCS#8 formats, and command line 7936 options to produce them. 7937 [Steve Henson] 7938 7939 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 7940 get temporary BIGNUMs from a BN_CTX. 7941 [Ulf M��ller] 7942 7943 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 7944 for p == 0. 7945 [Ulf M��ller] 7946 7947 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 7948 include a #define from the old name to the new. The original intent 7949 was that statically linked binaries could for example just call 7950 SSLeay_add_all_ciphers() to just add ciphers to the table and not 7951 link with digests. This never worked becayse SSLeay_add_all_digests() 7952 and SSLeay_add_all_ciphers() were in the same source file so calling 7953 one would link with the other. They are now in separate source files. 7954 [Steve Henson] 7955 7956 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 7957 [Steve Henson] 7958 7959 *) Use a less unusual form of the Miller-Rabin primality test (it used 7960 a binary algorithm for exponentiation integrated into the Miller-Rabin 7961 loop, our standard modexp algorithms are faster). 7962 [Bodo Moeller] 7963 7964 *) Support for the EBCDIC character set completed. 7965 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 7966 7967 *) Source code cleanups: use const where appropriate, eliminate casts, 7968 use void * instead of char * in lhash. 7969 [Ulf M��ller] 7970 7971 *) Bugfix: ssl3_send_server_key_exchange was not restartable 7972 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 7973 this the server could overwrite ephemeral keys that the client 7974 has already seen). 7975 [Bodo Moeller] 7976 7977 *) Turn DSA_is_prime into a macro that calls BN_is_prime, 7978 using 50 iterations of the Rabin-Miller test. 7979 7980 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 7981 iterations of the Rabin-Miller test as required by the appendix 7982 to FIPS PUB 186[-1]) instead of DSA_is_prime. 7983 As BN_is_prime_fasttest includes trial division, DSA parameter 7984 generation becomes much faster. 7985 7986 This implies a change for the callback functions in DSA_is_prime 7987 and DSA_generate_parameters: The callback function is called once 7988 for each positive witness in the Rabin-Miller test, not just 7989 occasionally in the inner loop; and the parameters to the 7990 callback function now provide an iteration count for the outer 7991 loop rather than for the current invocation of the inner loop. 7992 DSA_generate_parameters additionally can call the callback 7993 function with an 'iteration count' of -1, meaning that a 7994 candidate has passed the trial division test (when q is generated 7995 from an application-provided seed, trial division is skipped). 7996 [Bodo Moeller] 7997 7998 *) New function BN_is_prime_fasttest that optionally does trial 7999 division before starting the Rabin-Miller test and has 8000 an additional BN_CTX * argument (whereas BN_is_prime always 8001 has to allocate at least one BN_CTX). 8002 'callback(1, -1, cb_arg)' is called when a number has passed the 8003 trial division stage. 8004 [Bodo Moeller] 8005 8006 *) Fix for bug in CRL encoding. The validity dates weren't being handled 8007 as ASN1_TIME. 8008 [Steve Henson] 8009 8010 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 8011 [Steve Henson] 8012 8013 *) New function BN_pseudo_rand(). 8014 [Ulf M��ller] 8015 8016 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 8017 bignum version of BN_from_montgomery() with the working code from 8018 SSLeay 0.9.0 (the word based version is faster anyway), and clean up 8019 the comments. 8020 [Ulf M��ller] 8021 8022 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 8023 made it impossible to use the same SSL_SESSION data structure in 8024 SSL2 clients in multiple threads. 8025 [Bodo Moeller] 8026 8027 *) The return value of RAND_load_file() no longer counts bytes obtained 8028 by stat(). RAND_load_file(..., -1) is new and uses the complete file 8029 to seed the PRNG (previously an explicit byte count was required). 8030 [Ulf M��ller, Bodo M��ller] 8031 8032 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 8033 used (char *) instead of (void *) and had casts all over the place. 8034 [Steve Henson] 8035 8036 *) Make BN_generate_prime() return NULL on error if ret!=NULL. 8037 [Ulf M��ller] 8038 8039 *) Retain source code compatibility for BN_prime_checks macro: 8040 BN_is_prime(..., BN_prime_checks, ...) now uses 8041 BN_prime_checks_for_size to determine the appropriate number of 8042 Rabin-Miller iterations. 8043 [Ulf M��ller] 8044 8045 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 8046 DH_CHECK_P_NOT_SAFE_PRIME. 8047 (Check if this is true? OpenPGP calls them "strong".) 8048 [Ulf M��ller] 8049 8050 *) Merge the functionality of "dh" and "gendh" programs into a new program 8051 "dhparam". The old programs are retained for now but will handle DH keys 8052 (instead of parameters) in future. 8053 [Steve Henson] 8054 8055 *) Make the ciphers, s_server and s_client programs check the return values 8056 when a new cipher list is set. 8057 [Steve Henson] 8058 8059 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 8060 ciphers. Before when the 56bit ciphers were enabled the sorting was 8061 wrong. 8062 8063 The syntax for the cipher sorting has been extended to support sorting by 8064 cipher-strength (using the strength_bits hard coded in the tables). 8065 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 8066 8067 Fix a bug in the cipher-command parser: when supplying a cipher command 8068 string with an "undefined" symbol (neither command nor alphanumeric 8069 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 8070 an error is flagged. 8071 8072 Due to the strength-sorting extension, the code of the 8073 ssl_create_cipher_list() function was completely rearranged. I hope that 8074 the readability was also increased :-) 8075 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 8076 8077 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 8078 for the first serial number and places 2 in the serial number file. This 8079 avoids problems when the root CA is created with serial number zero and 8080 the first user certificate has the same issuer name and serial number 8081 as the root CA. 8082 [Steve Henson] 8083 8084 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 8085 the new code. Add documentation for this stuff. 8086 [Steve Henson] 8087 8088 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 8089 X509_*() to X509at_*() on the grounds that they don't handle X509 8090 structures and behave in an analagous way to the X509v3 functions: 8091 they shouldn't be called directly but wrapper functions should be used 8092 instead. 8093 8094 So we also now have some wrapper functions that call the X509at functions 8095 when passed certificate requests. (TO DO: similar things can be done with 8096 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 8097 things. Some of these need some d2i or i2d and print functionality 8098 because they handle more complex structures.) 8099 [Steve Henson] 8100 8101 *) Add missing #ifndefs that caused missing symbols when building libssl 8102 as a shared library without RSA. Use #ifndef NO_SSL2 instead of 8103 NO_RSA in ssl/s2*.c. 8104 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M��ller] 8105 8106 *) Precautions against using the PRNG uninitialized: RAND_bytes() now 8107 has a return value which indicates the quality of the random data 8108 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 8109 error queue. New function RAND_pseudo_bytes() generates output that is 8110 guaranteed to be unique but not unpredictable. RAND_add is like 8111 RAND_seed, but takes an extra argument for an entropy estimate 8112 (RAND_seed always assumes full entropy). 8113 [Ulf M��ller] 8114 8115 *) Do more iterations of Rabin-Miller probable prime test (specifically, 8116 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 8117 instead of only 2 for all lengths; see BN_prime_checks_for_size definition 8118 in crypto/bn/bn_prime.c for the complete table). This guarantees a 8119 false-positive rate of at most 2^-80 for random input. 8120 [Bodo Moeller] 8121 8122 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 8123 [Bodo Moeller] 8124 8125 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 8126 in the 0.9.5 release), this returns the chain 8127 from an X509_CTX structure with a dup of the stack and all 8128 the X509 reference counts upped: so the stack will exist 8129 after X509_CTX_cleanup() has been called. Modify pkcs12.c 8130 to use this. 8131 8132 Also make SSL_SESSION_print() print out the verify return 8133 code. 8134 [Steve Henson] 8135 8136 *) Add manpage for the pkcs12 command. Also change the default 8137 behaviour so MAC iteration counts are used unless the new 8138 -nomaciter option is used. This improves file security and 8139 only older versions of MSIE (4.0 for example) need it. 8140 [Steve Henson] 8141 8142 *) Honor the no-xxx Configure options when creating .DEF files. 8143 [Ulf M��ller] 8144 8145 *) Add PKCS#10 attributes to field table: challengePassword, 8146 unstructuredName and unstructuredAddress. These are taken from 8147 draft PKCS#9 v2.0 but are compatible with v1.2 provided no 8148 international characters are used. 8149 8150 More changes to X509_ATTRIBUTE code: allow the setting of types 8151 based on strings. Remove the 'loc' parameter when adding 8152 attributes because these will be a SET OF encoding which is sorted 8153 in ASN1 order. 8154 [Steve Henson] 8155 8156 *) Initial changes to the 'req' utility to allow request generation 8157 automation. This will allow an application to just generate a template 8158 file containing all the field values and have req construct the 8159 request. 8160 8161 Initial support for X509_ATTRIBUTE handling. Stacks of these are 8162 used all over the place including certificate requests and PKCS#7 8163 structures. They are currently handled manually where necessary with 8164 some primitive wrappers for PKCS#7. The new functions behave in a 8165 manner analogous to the X509 extension functions: they allow 8166 attributes to be looked up by NID and added. 8167 8168 Later something similar to the X509V3 code would be desirable to 8169 automatically handle the encoding, decoding and printing of the 8170 more complex types. The string types like challengePassword can 8171 be handled by the string table functions. 8172 8173 Also modified the multi byte string table handling. Now there is 8174 a 'global mask' which masks out certain types. The table itself 8175 can use the flag STABLE_NO_MASK to ignore the mask setting: this 8176 is useful when for example there is only one permissible type 8177 (as in countryName) and using the mask might result in no valid 8178 types at all. 8179 [Steve Henson] 8180 8181 *) Clean up 'Finished' handling, and add functions SSL_get_finished and 8182 SSL_get_peer_finished to allow applications to obtain the latest 8183 Finished messages sent to the peer or expected from the peer, 8184 respectively. (SSL_get_peer_finished is usually the Finished message 8185 actually received from the peer, otherwise the protocol will be aborted.) 8186 8187 As the Finished message are message digests of the complete handshake 8188 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 8189 be used for external authentication procedures when the authentication 8190 provided by SSL/TLS is not desired or is not enough. 8191 [Bodo Moeller] 8192 8193 *) Enhanced support for Alpha Linux is added. Now ./config checks if 8194 the host supports BWX extension and if Compaq C is present on the 8195 $PATH. Just exploiting of the BWX extension results in 20-30% 8196 performance kick for some algorithms, e.g. DES and RC4 to mention 8197 a couple. Compaq C in turn generates ~20% faster code for MD5 and 8198 SHA1. 8199 [Andy Polyakov] 8200 8201 *) Add support for MS "fast SGC". This is arguably a violation of the 8202 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 8203 weak crypto and after checking the certificate is SGC a second one 8204 with strong crypto. MS SGC stops the first handshake after receiving 8205 the server certificate message and sends a second client hello. Since 8206 a server will typically do all the time consuming operations before 8207 expecting any further messages from the client (server key exchange 8208 is the most expensive) there is little difference between the two. 8209 8210 To get OpenSSL to support MS SGC we have to permit a second client 8211 hello message after we have sent server done. In addition we have to 8212 reset the MAC if we do get this second client hello. 8213 [Steve Henson] 8214 8215 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 8216 if a DER encoded private key is RSA or DSA traditional format. Changed 8217 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 8218 format DER encoded private key. Newer code should use PKCS#8 format which 8219 has the key type encoded in the ASN1 structure. Added DER private key 8220 support to pkcs8 application. 8221 [Steve Henson] 8222 8223 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 8224 ciphersuites has been selected (as required by the SSL 3/TLS 1 8225 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 8226 is set, we interpret this as a request to violate the specification 8227 (the worst that can happen is a handshake failure, and 'correct' 8228 behaviour would result in a handshake failure anyway). 8229 [Bodo Moeller] 8230 8231 *) In SSL_CTX_add_session, take into account that there might be multiple 8232 SSL_SESSION structures with the same session ID (e.g. when two threads 8233 concurrently obtain them from an external cache). 8234 The internal cache can handle only one SSL_SESSION with a given ID, 8235 so if there's a conflict, we now throw out the old one to achieve 8236 consistency. 8237 [Bodo Moeller] 8238 8239 *) Add OIDs for idea and blowfish in CBC mode. This will allow both 8240 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 8241 some routines that use cipher OIDs: some ciphers do not have OIDs 8242 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 8243 example. 8244 [Steve Henson] 8245 8246 *) Simplify the trust setting structure and code. Now we just have 8247 two sequences of OIDs for trusted and rejected settings. These will 8248 typically have values the same as the extended key usage extension 8249 and any application specific purposes. 8250 8251 The trust checking code now has a default behaviour: it will just 8252 check for an object with the same NID as the passed id. Functions can 8253 be provided to override either the default behaviour or the behaviour 8254 for a given id. SSL client, server and email already have functions 8255 in place for compatibility: they check the NID and also return "trusted" 8256 if the certificate is self signed. 8257 [Steve Henson] 8258 8259 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 8260 traditional format into an EVP_PKEY structure. 8261 [Steve Henson] 8262 8263 *) Add a password callback function PEM_cb() which either prompts for 8264 a password if usr_data is NULL or otherwise assumes it is a null 8265 terminated password. Allow passwords to be passed on command line 8266 environment or config files in a few more utilities. 8267 [Steve Henson] 8268 8269 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 8270 keys. Add some short names for PKCS#8 PBE algorithms and allow them 8271 to be specified on the command line for the pkcs8 and pkcs12 utilities. 8272 Update documentation. 8273 [Steve Henson] 8274 8275 *) Support for ASN1 "NULL" type. This could be handled before by using 8276 ASN1_TYPE but there wasn't any function that would try to read a NULL 8277 and produce an error if it couldn't. For compatibility we also have 8278 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 8279 don't allocate anything because they don't need to. 8280 [Steve Henson] 8281 8282 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 8283 for details. 8284 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 8285 8286 *) Rebuild of the memory allocation routines used by OpenSSL code and 8287 possibly others as well. The purpose is to make an interface that 8288 provide hooks so anyone can build a separate set of allocation and 8289 deallocation routines to be used by OpenSSL, for example memory 8290 pool implementations, or something else, which was previously hard 8291 since Malloc(), Realloc() and Free() were defined as macros having 8292 the values malloc, realloc and free, respectively (except for Win32 8293 compilations). The same is provided for memory debugging code. 8294 OpenSSL already comes with functionality to find memory leaks, but 8295 this gives people a chance to debug other memory problems. 8296 8297 With these changes, a new set of functions and macros have appeared: 8298 8299 CRYPTO_set_mem_debug_functions() [F] 8300 CRYPTO_get_mem_debug_functions() [F] 8301 CRYPTO_dbg_set_options() [F] 8302 CRYPTO_dbg_get_options() [F] 8303 CRYPTO_malloc_debug_init() [M] 8304 8305 The memory debug functions are NULL by default, unless the library 8306 is compiled with CRYPTO_MDEBUG or friends is defined. If someone 8307 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 8308 gives the standard debugging functions that come with OpenSSL) or 8309 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 8310 provided by the library user) must be used. When the standard 8311 debugging functions are used, CRYPTO_dbg_set_options can be used to 8312 request additional information: 8313 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 8314 the CRYPTO_MDEBUG_xxx macro when compiling the library. 8315 8316 Also, things like CRYPTO_set_mem_functions will always give the 8317 expected result (the new set of functions is used for allocation 8318 and deallocation) at all times, regardless of platform and compiler 8319 options. 8320 8321 To finish it up, some functions that were never use in any other 8322 way than through macros have a new API and new semantic: 8323 8324 CRYPTO_dbg_malloc() 8325 CRYPTO_dbg_realloc() 8326 CRYPTO_dbg_free() 8327 8328 All macros of value have retained their old syntax. 8329 [Richard Levitte and Bodo Moeller] 8330 8331 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 8332 ordering of SMIMECapabilities wasn't in "strength order" and there 8333 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 8334 algorithm. 8335 [Steve Henson] 8336 8337 *) Some ASN1 types with illegal zero length encoding (INTEGER, 8338 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 8339 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 8340 8341 *) Merge in my S/MIME library for OpenSSL. This provides a simple 8342 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 8343 functionality to handle multipart/signed properly) and a utility 8344 called 'smime' to call all this stuff. This is based on code I 8345 originally wrote for Celo who have kindly allowed it to be 8346 included in OpenSSL. 8347 [Steve Henson] 8348 8349 *) Add variants des_set_key_checked and des_set_key_unchecked of 8350 des_set_key (aka des_key_sched). Global variable des_check_key 8351 decides which of these is called by des_set_key; this way 8352 des_check_key behaves as it always did, but applications and 8353 the library itself, which was buggy for des_check_key == 1, 8354 have a cleaner way to pick the version they need. 8355 [Bodo Moeller] 8356 8357 *) New function PKCS12_newpass() which changes the password of a 8358 PKCS12 structure. 8359 [Steve Henson] 8360 8361 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 8362 dynamic mix. In both cases the ids can be used as an index into the 8363 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 8364 functions so they accept a list of the field values and the 8365 application doesn't need to directly manipulate the X509_TRUST 8366 structure. 8367 [Steve Henson] 8368 8369 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 8370 need initialising. 8371 [Steve Henson] 8372 8373 *) Modify the way the V3 extension code looks up extensions. This now 8374 works in a similar way to the object code: we have some "standard" 8375 extensions in a static table which is searched with OBJ_bsearch() 8376 and the application can add dynamic ones if needed. The file 8377 crypto/x509v3/ext_dat.h now has the info: this file needs to be 8378 updated whenever a new extension is added to the core code and kept 8379 in ext_nid order. There is a simple program 'tabtest.c' which checks 8380 this. New extensions are not added too often so this file can readily 8381 be maintained manually. 8382 8383 There are two big advantages in doing things this way. The extensions 8384 can be looked up immediately and no longer need to be "added" using 8385 X509V3_add_standard_extensions(): this function now does nothing. 8386 [Side note: I get *lots* of email saying the extension code doesn't 8387 work because people forget to call this function] 8388 Also no dynamic allocation is done unless new extensions are added: 8389 so if we don't add custom extensions there is no need to call 8390 X509V3_EXT_cleanup(). 8391 [Steve Henson] 8392 8393 *) Modify enc utility's salting as follows: make salting the default. Add a 8394 magic header, so unsalted files fail gracefully instead of just decrypting 8395 to garbage. This is because not salting is a big security hole, so people 8396 should be discouraged from doing it. 8397 [Ben Laurie] 8398 8399 *) Fixes and enhancements to the 'x509' utility. It allowed a message 8400 digest to be passed on the command line but it only used this 8401 parameter when signing a certificate. Modified so all relevant 8402 operations are affected by the digest parameter including the 8403 -fingerprint and -x509toreq options. Also -x509toreq choked if a 8404 DSA key was used because it didn't fix the digest. 8405 [Steve Henson] 8406 8407 *) Initial certificate chain verify code. Currently tests the untrusted 8408 certificates for consistency with the verify purpose (which is set 8409 when the X509_STORE_CTX structure is set up) and checks the pathlength. 8410 8411 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 8412 this is because it will reject chains with invalid extensions whereas 8413 every previous version of OpenSSL and SSLeay made no checks at all. 8414 8415 Trust code: checks the root CA for the relevant trust settings. Trust 8416 settings have an initial value consistent with the verify purpose: e.g. 8417 if the verify purpose is for SSL client use it expects the CA to be 8418 trusted for SSL client use. However the default value can be changed to 8419 permit custom trust settings: one example of this would be to only trust 8420 certificates from a specific "secure" set of CAs. 8421 8422 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 8423 which should be used for version portability: especially since the 8424 verify structure is likely to change more often now. 8425 8426 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 8427 to set them. If not set then assume SSL clients will verify SSL servers 8428 and vice versa. 8429 8430 Two new options to the verify program: -untrusted allows a set of 8431 untrusted certificates to be passed in and -purpose which sets the 8432 intended purpose of the certificate. If a purpose is set then the 8433 new chain verify code is used to check extension consistency. 8434 [Steve Henson] 8435 8436 *) Support for the authority information access extension. 8437 [Steve Henson] 8438 8439 *) Modify RSA and DSA PEM read routines to transparently handle 8440 PKCS#8 format private keys. New *_PUBKEY_* functions that handle 8441 public keys in a format compatible with certificate 8442 SubjectPublicKeyInfo structures. Unfortunately there were already 8443 functions called *_PublicKey_* which used various odd formats so 8444 these are retained for compatibility: however the DSA variants were 8445 never in a public release so they have been deleted. Changed dsa/rsa 8446 utilities to handle the new format: note no releases ever handled public 8447 keys so we should be OK. 8448 8449 The primary motivation for this change is to avoid the same fiasco 8450 that dogs private keys: there are several incompatible private key 8451 formats some of which are standard and some OpenSSL specific and 8452 require various evil hacks to allow partial transparent handling and 8453 even then it doesn't work with DER formats. Given the option anything 8454 other than PKCS#8 should be dumped: but the other formats have to 8455 stay in the name of compatibility. 8456 8457 With public keys and the benefit of hindsight one standard format 8458 is used which works with EVP_PKEY, RSA or DSA structures: though 8459 it clearly returns an error if you try to read the wrong kind of key. 8460 8461 Added a -pubkey option to the 'x509' utility to output the public key. 8462 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 8463 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 8464 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 8465 that do the same as the EVP_PKEY_assign_*() except they up the 8466 reference count of the added key (they don't "swallow" the 8467 supplied key). 8468 [Steve Henson] 8469 8470 *) Fixes to crypto/x509/by_file.c the code to read in certificates and 8471 CRLs would fail if the file contained no certificates or no CRLs: 8472 added a new function to read in both types and return the number 8473 read: this means that if none are read it will be an error. The 8474 DER versions of the certificate and CRL reader would always fail 8475 because it isn't possible to mix certificates and CRLs in DER format 8476 without choking one or the other routine. Changed this to just read 8477 a certificate: this is the best we can do. Also modified the code 8478 in apps/verify.c to take notice of return codes: it was previously 8479 attempting to read in certificates from NULL pointers and ignoring 8480 any errors: this is one reason why the cert and CRL reader seemed 8481 to work. It doesn't check return codes from the default certificate 8482 routines: these may well fail if the certificates aren't installed. 8483 [Steve Henson] 8484 8485 *) Code to support otherName option in GeneralName. 8486 [Steve Henson] 8487 8488 *) First update to verify code. Change the verify utility 8489 so it warns if it is passed a self signed certificate: 8490 for consistency with the normal behaviour. X509_verify 8491 has been modified to it will now verify a self signed 8492 certificate if *exactly* the same certificate appears 8493 in the store: it was previously impossible to trust a 8494 single self signed certificate. This means that: 8495 openssl verify ss.pem 8496 now gives a warning about a self signed certificate but 8497 openssl verify -CAfile ss.pem ss.pem 8498 is OK. 8499 [Steve Henson] 8500 8501 *) For servers, store verify_result in SSL_SESSION data structure 8502 (and add it to external session representation). 8503 This is needed when client certificate verifications fails, 8504 but an application-provided verification callback (set by 8505 SSL_CTX_set_cert_verify_callback) allows accepting the session 8506 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 8507 but returns 1): When the session is reused, we have to set 8508 ssl->verify_result to the appropriate error code to avoid 8509 security holes. 8510 [Bodo Moeller, problem pointed out by Lutz Jaenicke] 8511 8512 *) Fix a bug in the new PKCS#7 code: it didn't consider the 8513 case in PKCS7_dataInit() where the signed PKCS7 structure 8514 didn't contain any existing data because it was being created. 8515 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 8516 8517 *) Add a salt to the key derivation routines in enc.c. This 8518 forms the first 8 bytes of the encrypted file. Also add a 8519 -S option to allow a salt to be input on the command line. 8520 [Steve Henson] 8521 8522 *) New function X509_cmp(). Oddly enough there wasn't a function 8523 to compare two certificates. We do this by working out the SHA1 8524 hash and comparing that. X509_cmp() will be needed by the trust 8525 code. 8526 [Steve Henson] 8527 8528 *) SSL_get1_session() is like SSL_get_session(), but increments 8529 the reference count in the SSL_SESSION returned. 8530 [Geoff Thorpe <geoff@eu.c2.net>] 8531 8532 *) Fix for 'req': it was adding a null to request attributes. 8533 Also change the X509_LOOKUP and X509_INFO code to handle 8534 certificate auxiliary information. 8535 [Steve Henson] 8536 8537 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 8538 the 'enc' command. 8539 [Steve Henson] 8540 8541 *) Add the possibility to add extra information to the memory leak 8542 detecting output, to form tracebacks, showing from where each 8543 allocation was originated: CRYPTO_push_info("constant string") adds 8544 the string plus current file name and line number to a per-thread 8545 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 8546 is like calling CYRPTO_pop_info() until the stack is empty. 8547 Also updated memory leak detection code to be multi-thread-safe. 8548 [Richard Levitte] 8549 8550 *) Add options -text and -noout to pkcs7 utility and delete the 8551 encryption options which never did anything. Update docs. 8552 [Steve Henson] 8553 8554 *) Add options to some of the utilities to allow the pass phrase 8555 to be included on either the command line (not recommended on 8556 OSes like Unix) or read from the environment. Update the 8557 manpages and fix a few bugs. 8558 [Steve Henson] 8559 8560 *) Add a few manpages for some of the openssl commands. 8561 [Steve Henson] 8562 8563 *) Fix the -revoke option in ca. It was freeing up memory twice, 8564 leaking and not finding already revoked certificates. 8565 [Steve Henson] 8566 8567 *) Extensive changes to support certificate auxiliary information. 8568 This involves the use of X509_CERT_AUX structure and X509_AUX 8569 functions. An X509_AUX function such as PEM_read_X509_AUX() 8570 can still read in a certificate file in the usual way but it 8571 will also read in any additional "auxiliary information". By 8572 doing things this way a fair degree of compatibility can be 8573 retained: existing certificates can have this information added 8574 using the new 'x509' options. 8575 8576 Current auxiliary information includes an "alias" and some trust 8577 settings. The trust settings will ultimately be used in enhanced 8578 certificate chain verification routines: currently a certificate 8579 can only be trusted if it is self signed and then it is trusted 8580 for all purposes. 8581 [Steve Henson] 8582 8583 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 8584 The problem was that one of the replacement routines had not been working 8585 since SSLeay releases. For now the offending routine has been replaced 8586 with non-optimised assembler. Even so, this now gives around 95% 8587 performance improvement for 1024 bit RSA signs. 8588 [Mark Cox] 8589 8590 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 8591 handling. Most clients have the effective key size in bits equal to 8592 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 8593 A few however don't do this and instead use the size of the decrypted key 8594 to determine the RC2 key length and the AlgorithmIdentifier to determine 8595 the effective key length. In this case the effective key length can still 8596 be 40 bits but the key length can be 168 bits for example. This is fixed 8597 by manually forcing an RC2 key into the EVP_PKEY structure because the 8598 EVP code can't currently handle unusual RC2 key sizes: it always assumes 8599 the key length and effective key length are equal. 8600 [Steve Henson] 8601 8602 *) Add a bunch of functions that should simplify the creation of 8603 X509_NAME structures. Now you should be able to do: 8604 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 8605 and have it automatically work out the correct field type and fill in 8606 the structures. The more adventurous can try: 8607 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 8608 and it will (hopefully) work out the correct multibyte encoding. 8609 [Steve Henson] 8610 8611 *) Change the 'req' utility to use the new field handling and multibyte 8612 copy routines. Before the DN field creation was handled in an ad hoc 8613 way in req, ca, and x509 which was rather broken and didn't support 8614 BMPStrings or UTF8Strings. Since some software doesn't implement 8615 BMPStrings or UTF8Strings yet, they can be enabled using the config file 8616 using the dirstring_type option. See the new comment in the default 8617 openssl.cnf for more info. 8618 [Steve Henson] 8619 8620 *) Make crypto/rand/md_rand.c more robust: 8621 - Assure unique random numbers after fork(). 8622 - Make sure that concurrent threads access the global counter and 8623 md serializably so that we never lose entropy in them 8624 or use exactly the same state in multiple threads. 8625 Access to the large state is not always serializable because 8626 the additional locking could be a performance killer, and 8627 md should be large enough anyway. 8628 [Bodo Moeller] 8629 8630 *) New file apps/app_rand.c with commonly needed functionality 8631 for handling the random seed file. 8632 8633 Use the random seed file in some applications that previously did not: 8634 ca, 8635 dsaparam -genkey (which also ignored its '-rand' option), 8636 s_client, 8637 s_server, 8638 x509 (when signing). 8639 Except on systems with /dev/urandom, it is crucial to have a random 8640 seed file at least for key creation, DSA signing, and for DH exchanges; 8641 for RSA signatures we could do without one. 8642 8643 gendh and gendsa (unlike genrsa) used to read only the first byte 8644 of each file listed in the '-rand' option. The function as previously 8645 found in genrsa is now in app_rand.c and is used by all programs 8646 that support '-rand'. 8647 [Bodo Moeller] 8648 8649 *) In RAND_write_file, use mode 0600 for creating files; 8650 don't just chmod when it may be too late. 8651 [Bodo Moeller] 8652 8653 *) Report an error from X509_STORE_load_locations 8654 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 8655 [Bill Perry] 8656 8657 *) New function ASN1_mbstring_copy() this copies a string in either 8658 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 8659 into an ASN1_STRING type. A mask of permissible types is passed 8660 and it chooses the "minimal" type to use or an error if not type 8661 is suitable. 8662 [Steve Henson] 8663 8664 *) Add function equivalents to the various macros in asn1.h. The old 8665 macros are retained with an M_ prefix. Code inside the library can 8666 use the M_ macros. External code (including the openssl utility) 8667 should *NOT* in order to be "shared library friendly". 8668 [Steve Henson] 8669 8670 *) Add various functions that can check a certificate's extensions 8671 to see if it usable for various purposes such as SSL client, 8672 server or S/MIME and CAs of these types. This is currently 8673 VERY EXPERIMENTAL but will ultimately be used for certificate chain 8674 verification. Also added a -purpose flag to x509 utility to 8675 print out all the purposes. 8676 [Steve Henson] 8677 8678 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 8679 functions. 8680 [Steve Henson] 8681 8682 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 8683 for, obtain and decode and extension and obtain its critical flag. 8684 This allows all the necessary extension code to be handled in a 8685 single function call. 8686 [Steve Henson] 8687 8688 *) RC4 tune-up featuring 30-40% performance improvement on most RISC 8689 platforms. See crypto/rc4/rc4_enc.c for further details. 8690 [Andy Polyakov] 8691 8692 *) New -noout option to asn1parse. This causes no output to be produced 8693 its main use is when combined with -strparse and -out to extract data 8694 from a file (which may not be in ASN.1 format). 8695 [Steve Henson] 8696 8697 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 8698 when producing the local key id. 8699 [Richard Levitte <levitte@stacken.kth.se>] 8700 8701 *) New option -dhparam in s_server. This allows a DH parameter file to be 8702 stated explicitly. If it is not stated then it tries the first server 8703 certificate file. The previous behaviour hard coded the filename 8704 "server.pem". 8705 [Steve Henson] 8706 8707 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 8708 a public key to be input or output. For example: 8709 openssl rsa -in key.pem -pubout -out pubkey.pem 8710 Also added necessary DSA public key functions to handle this. 8711 [Steve Henson] 8712 8713 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 8714 in the message. This was handled by allowing 8715 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 8716 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 8717 8718 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 8719 to the end of the strings whereas this didn't. This would cause problems 8720 if strings read with d2i_ASN1_bytes() were later modified. 8721 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 8722 8723 *) Fix for base64 decode bug. When a base64 bio reads only one line of 8724 data and it contains EOF it will end up returning an error. This is 8725 caused by input 46 bytes long. The cause is due to the way base64 8726 BIOs find the start of base64 encoded data. They do this by trying a 8727 trial decode on each line until they find one that works. When they 8728 do a flag is set and it starts again knowing it can pass all the 8729 data directly through the decoder. Unfortunately it doesn't reset 8730 the context it uses. This means that if EOF is reached an attempt 8731 is made to pass two EOFs through the context and this causes the 8732 resulting error. This can also cause other problems as well. As is 8733 usual with these problems it takes *ages* to find and the fix is 8734 trivial: move one line. 8735 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 8736 8737 *) Ugly workaround to get s_client and s_server working under Windows. The 8738 old code wouldn't work because it needed to select() on sockets and the 8739 tty (for keypresses and to see if data could be written). Win32 only 8740 supports select() on sockets so we select() with a 1s timeout on the 8741 sockets and then see if any characters are waiting to be read, if none 8742 are present then we retry, we also assume we can always write data to 8743 the tty. This isn't nice because the code then blocks until we've 8744 received a complete line of data and it is effectively polling the 8745 keyboard at 1s intervals: however it's quite a bit better than not 8746 working at all :-) A dedicated Windows application might handle this 8747 with an event loop for example. 8748 [Steve Henson] 8749 8750 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 8751 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 8752 will be called when RSA_sign() and RSA_verify() are used. This is useful 8753 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 8754 For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 8755 should *not* be used: RSA_sign() and RSA_verify() must be used instead. 8756 This necessitated the support of an extra signature type NID_md5_sha1 8757 for SSL signatures and modifications to the SSL library to use it instead 8758 of calling RSA_public_decrypt() and RSA_private_encrypt(). 8759 [Steve Henson] 8760 8761 *) Add new -verify -CAfile and -CApath options to the crl program, these 8762 will lookup a CRL issuers certificate and verify the signature in a 8763 similar way to the verify program. Tidy up the crl program so it 8764 no longer accesses structures directly. Make the ASN1 CRL parsing a bit 8765 less strict. It will now permit CRL extensions even if it is not 8766 a V2 CRL: this will allow it to tolerate some broken CRLs. 8767 [Steve Henson] 8768 8769 *) Initialize all non-automatic variables each time one of the openssl 8770 sub-programs is started (this is necessary as they may be started 8771 multiple times from the "OpenSSL>" prompt). 8772 [Lennart Bang, Bodo Moeller] 8773 8774 *) Preliminary compilation option RSA_NULL which disables RSA crypto without 8775 removing all other RSA functionality (this is what NO_RSA does). This 8776 is so (for example) those in the US can disable those operations covered 8777 by the RSA patent while allowing storage and parsing of RSA keys and RSA 8778 key generation. 8779 [Steve Henson] 8780 8781 *) Non-copying interface to BIO pairs. 8782 (still largely untested) 8783 [Bodo Moeller] 8784 8785 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 8786 ASCII string. This was handled independently in various places before. 8787 [Steve Henson] 8788 8789 *) New functions UTF8_getc() and UTF8_putc() that parse and generate 8790 UTF8 strings a character at a time. 8791 [Steve Henson] 8792 8793 *) Use client_version from client hello to select the protocol 8794 (s23_srvr.c) and for RSA client key exchange verification 8795 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 8796 [Bodo Moeller] 8797 8798 *) Add various utility functions to handle SPKACs, these were previously 8799 handled by poking round in the structure internals. Added new function 8800 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 8801 print, verify and generate SPKACs. Based on an original idea from 8802 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 8803 [Steve Henson] 8804 8805 *) RIPEMD160 is operational on all platforms and is back in 'make test'. 8806 [Andy Polyakov] 8807 8808 *) Allow the config file extension section to be overwritten on the 8809 command line. Based on an original idea from Massimiliano Pala 8810 <madwolf@comune.modena.it>. The new option is called -extensions 8811 and can be applied to ca, req and x509. Also -reqexts to override 8812 the request extensions in req and -crlexts to override the crl extensions 8813 in ca. 8814 [Steve Henson] 8815 8816 *) Add new feature to the SPKAC handling in ca. Now you can include 8817 the same field multiple times by preceding it by "XXXX." for example: 8818 1.OU="Unit name 1" 8819 2.OU="Unit name 2" 8820 this is the same syntax as used in the req config file. 8821 [Steve Henson] 8822 8823 *) Allow certificate extensions to be added to certificate requests. These 8824 are specified in a 'req_extensions' option of the req section of the 8825 config file. They can be printed out with the -text option to req but 8826 are otherwise ignored at present. 8827 [Steve Henson] 8828 8829 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 8830 data read consists of only the final block it would not decrypted because 8831 EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 8832 A misplaced 'break' also meant the decrypted final block might not be 8833 copied until the next read. 8834 [Steve Henson] 8835 8836 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 8837 a few extra parameters to the DH structure: these will be useful if 8838 for example we want the value of 'q' or implement X9.42 DH. 8839 [Steve Henson] 8840 8841 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 8842 provides hooks that allow the default DSA functions or functions on a 8843 "per key" basis to be replaced. This allows hardware acceleration and 8844 hardware key storage to be handled without major modification to the 8845 library. Also added low level modexp hooks and CRYPTO_EX structure and 8846 associated functions. 8847 [Steve Henson] 8848 8849 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 8850 as "read only": it can't be written to and the buffer it points to will 8851 not be freed. Reading from a read only BIO is much more efficient than 8852 a normal memory BIO. This was added because there are several times when 8853 an area of memory needs to be read from a BIO. The previous method was 8854 to create a memory BIO and write the data to it, this results in two 8855 copies of the data and an O(n^2) reading algorithm. There is a new 8856 function BIO_new_mem_buf() which creates a read only memory BIO from 8857 an area of memory. Also modified the PKCS#7 routines to use read only 8858 memory BIOs. 8859 [Steve Henson] 8860 8861 *) Bugfix: ssl23_get_client_hello did not work properly when called in 8862 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 8863 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 8864 but a retry condition occured while trying to read the rest. 8865 [Bodo Moeller] 8866 8867 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 8868 NID_pkcs7_encrypted by default: this was wrong since this should almost 8869 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 8870 the encrypted data type: this is a more sensible place to put it and it 8871 allows the PKCS#12 code to be tidied up that duplicated this 8872 functionality. 8873 [Steve Henson] 8874 8875 *) Changed obj_dat.pl script so it takes its input and output files on 8876 the command line. This should avoid shell escape redirection problems 8877 under Win32. 8878 [Steve Henson] 8879 8880 *) Initial support for certificate extension requests, these are included 8881 in things like Xenroll certificate requests. Included functions to allow 8882 extensions to be obtained and added. 8883 [Steve Henson] 8884 8885 *) -crlf option to s_client and s_server for sending newlines as 8886 CRLF (as required by many protocols). 8887 [Bodo Moeller] 8888 8889 Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 8890 8891 *) Install libRSAglue.a when OpenSSL is built with RSAref. 8892 [Ralf S. Engelschall] 8893 8894 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 8895 [Andrija Antonijevic <TheAntony2@bigfoot.com>] 8896 8897 *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 8898 program. 8899 [Steve Henson] 8900 8901 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 8902 DH parameters/keys (q is lost during that conversion, but the resulting 8903 DH parameters contain its length). 8904 8905 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 8906 much faster than DH_generate_parameters (which creates parameters 8907 where p = 2*q + 1), and also the smaller q makes DH computations 8908 much more efficient (160-bit exponentiation instead of 1024-bit 8909 exponentiation); so this provides a convenient way to support DHE 8910 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 8911 utter importance to use 8912 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 8913 or 8914 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 8915 when such DH parameters are used, because otherwise small subgroup 8916 attacks may become possible! 8917 [Bodo Moeller] 8918 8919 *) Avoid memory leak in i2d_DHparams. 8920 [Bodo Moeller] 8921 8922 *) Allow the -k option to be used more than once in the enc program: 8923 this allows the same encrypted message to be read by multiple recipients. 8924 [Steve Henson] 8925 8926 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 8927 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 8928 it will always use the numerical form of the OID, even if it has a short 8929 or long name. 8930 [Steve Henson] 8931 8932 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 8933 method only got called if p,q,dmp1,dmq1,iqmp components were present, 8934 otherwise bn_mod_exp was called. In the case of hardware keys for example 8935 no private key components need be present and it might store extra data 8936 in the RSA structure, which cannot be accessed from bn_mod_exp. 8937 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 8938 private key operations. 8939 [Steve Henson] 8940 8941 *) Added support for SPARC Linux. 8942 [Andy Polyakov] 8943 8944 *) pem_password_cb function type incompatibly changed from 8945 typedef int pem_password_cb(char *buf, int size, int rwflag); 8946 to 8947 ....(char *buf, int size, int rwflag, void *userdata); 8948 so that applications can pass data to their callbacks: 8949 The PEM[_ASN1]_{read,write}... functions and macros now take an 8950 additional void * argument, which is just handed through whenever 8951 the password callback is called. 8952 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 8953 8954 New function SSL_CTX_set_default_passwd_cb_userdata. 8955 8956 Compatibility note: As many C implementations push function arguments 8957 onto the stack in reverse order, the new library version is likely to 8958 interoperate with programs that have been compiled with the old 8959 pem_password_cb definition (PEM_whatever takes some data that 8960 happens to be on the stack as its last argument, and the callback 8961 just ignores this garbage); but there is no guarantee whatsoever that 8962 this will work. 8963 8964 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 8965 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 8966 problems not only on Windows, but also on some Unix platforms. 8967 To avoid problematic command lines, these definitions are now in an 8968 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 8969 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 8970 [Bodo Moeller] 8971 8972 *) MIPS III/IV assembler module is reimplemented. 8973 [Andy Polyakov] 8974 8975 *) More DES library cleanups: remove references to srand/rand and 8976 delete an unused file. 8977 [Ulf M��ller] 8978 8979 *) Add support for the the free Netwide assembler (NASM) under Win32, 8980 since not many people have MASM (ml) and it can be hard to obtain. 8981 This is currently experimental but it seems to work OK and pass all 8982 the tests. Check out INSTALL.W32 for info. 8983 [Steve Henson] 8984 8985 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 8986 without temporary keys kept an extra copy of the server key, 8987 and connections with temporary keys did not free everything in case 8988 of an error. 8989 [Bodo Moeller] 8990 8991 *) New function RSA_check_key and new openssl rsa option -check 8992 for verifying the consistency of RSA keys. 8993 [Ulf Moeller, Bodo Moeller] 8994 8995 *) Various changes to make Win32 compile work: 8996 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 8997 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 8998 comparison" warnings. 8999 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 9000 [Steve Henson] 9001 9002 *) Add a debugging option to PKCS#5 v2 key generation function: when 9003 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 9004 derived keys are printed to stderr. 9005 [Steve Henson] 9006 9007 *) Copy the flags in ASN1_STRING_dup(). 9008 [Roman E. Pavlov <pre@mo.msk.ru>] 9009 9010 *) The x509 application mishandled signing requests containing DSA 9011 keys when the signing key was also DSA and the parameters didn't match. 9012 9013 It was supposed to omit the parameters when they matched the signing key: 9014 the verifying software was then supposed to automatically use the CA's 9015 parameters if they were absent from the end user certificate. 9016 9017 Omitting parameters is no longer recommended. The test was also 9018 the wrong way round! This was probably due to unusual behaviour in 9019 EVP_cmp_parameters() which returns 1 if the parameters match. 9020 This meant that parameters were omitted when they *didn't* match and 9021 the certificate was useless. Certificates signed with 'ca' didn't have 9022 this bug. 9023 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 9024 9025 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 9026 The interface is as follows: 9027 Applications can use 9028 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 9029 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 9030 "off" is now the default. 9031 The library internally uses 9032 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 9033 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 9034 to disable memory-checking temporarily. 9035 9036 Some inconsistent states that previously were possible (and were 9037 even the default) are now avoided. 9038 9039 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 9040 with each memory chunk allocated; this is occasionally more helpful 9041 than just having a counter. 9042 9043 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 9044 9045 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 9046 extensions. 9047 [Bodo Moeller] 9048 9049 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 9050 which largely parallels "options", but is for changing API behaviour, 9051 whereas "options" are about protocol behaviour. 9052 Initial "mode" flags are: 9053 9054 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 9055 a single record has been written. 9056 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 9057 retries use the same buffer location. 9058 (But all of the contents must be 9059 copied!) 9060 [Bodo Moeller] 9061 9062 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 9063 worked. 9064 9065 *) Fix problems with no-hmac etc. 9066 [Ulf M��ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 9067 9068 *) New functions RSA_get_default_method(), RSA_set_method() and 9069 RSA_get_method(). These allows replacement of RSA_METHODs without having 9070 to mess around with the internals of an RSA structure. 9071 [Steve Henson] 9072 9073 *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 9074 Also really enable memory leak checks in openssl.c and in some 9075 test programs. 9076 [Chad C. Mulligan, Bodo Moeller] 9077 9078 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 9079 up the length of negative integers. This has now been simplified to just 9080 store the length when it is first determined and use it later, rather 9081 than trying to keep track of where data is copied and updating it to 9082 point to the end. 9083 [Steve Henson, reported by Brien Wheeler 9084 <bwheeler@authentica-security.com>] 9085 9086 *) Add a new function PKCS7_signatureVerify. This allows the verification 9087 of a PKCS#7 signature but with the signing certificate passed to the 9088 function itself. This contrasts with PKCS7_dataVerify which assumes the 9089 certificate is present in the PKCS#7 structure. This isn't always the 9090 case: certificates can be omitted from a PKCS#7 structure and be 9091 distributed by "out of band" means (such as a certificate database). 9092 [Steve Henson] 9093 9094 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 9095 function prototypes in pem.h, also change util/mkdef.pl to add the 9096 necessary function names. 9097 [Steve Henson] 9098 9099 *) mk1mf.pl (used by Windows builds) did not properly read the 9100 options set by Configure in the top level Makefile, and Configure 9101 was not even able to write more than one option correctly. 9102 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 9103 [Bodo Moeller] 9104 9105 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 9106 file to be loaded from a BIO or FILE pointer. The BIO version will 9107 for example allow memory BIOs to contain config info. 9108 [Steve Henson] 9109 9110 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 9111 Whoever hopes to achieve shared-library compatibility across versions 9112 must use this, not the compile-time macro. 9113 (Exercise 0.9.4: Which is the minimum library version required by 9114 such programs?) 9115 Note: All this applies only to multi-threaded programs, others don't 9116 need locks. 9117 [Bodo Moeller] 9118 9119 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 9120 through a BIO pair triggered the default case, i.e. 9121 SSLerr(...,SSL_R_UNKNOWN_STATE). 9122 [Bodo Moeller] 9123 9124 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 9125 can use the SSL library even if none of the specific BIOs is 9126 appropriate. 9127 [Bodo Moeller] 9128 9129 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 9130 for the encoded length. 9131 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 9132 9133 *) Add initial documentation of the X509V3 functions. 9134 [Steve Henson] 9135 9136 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 9137 PEM_write_bio_PKCS8PrivateKey() that are equivalent to 9138 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 9139 secure PKCS#8 private key format with a high iteration count. 9140 [Steve Henson] 9141 9142 *) Fix determination of Perl interpreter: A perl or perl5 9143 _directory_ in $PATH was also accepted as the interpreter. 9144 [Ralf S. Engelschall] 9145 9146 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 9147 wrong with it but it was very old and did things like calling 9148 PEM_ASN1_read() directly and used MD5 for the hash not to mention some 9149 unusual formatting. 9150 [Steve Henson] 9151 9152 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 9153 to use the new extension code. 9154 [Steve Henson] 9155 9156 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 9157 with macros. This should make it easier to change their form, add extra 9158 arguments etc. Fix a few PEM prototypes which didn't have cipher as a 9159 constant. 9160 [Steve Henson] 9161 9162 *) Add to configuration table a new entry that can specify an alternative 9163 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 9164 according to Mark Crispin <MRC@Panda.COM>. 9165 [Bodo Moeller] 9166 9167#if 0 9168 *) DES CBC did not update the IV. Weird. 9169 [Ben Laurie] 9170#else 9171 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 9172 Changing the behaviour of the former might break existing programs -- 9173 where IV updating is needed, des_ncbc_encrypt can be used. 9174#endif 9175 9176 *) When bntest is run from "make test" it drives bc to check its 9177 calculations, as well as internally checking them. If an internal check 9178 fails, it needs to cause bc to give a non-zero result or make test carries 9179 on without noticing the failure. Fixed. 9180 [Ben Laurie] 9181 9182 *) DES library cleanups. 9183 [Ulf M��ller] 9184 9185 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 9186 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 9187 ciphers. NOTE: although the key derivation function has been verified 9188 against some published test vectors it has not been extensively tested 9189 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 9190 of v2.0. 9191 [Steve Henson] 9192 9193 *) Instead of "mkdir -p", which is not fully portable, use new 9194 Perl script "util/mkdir-p.pl". 9195 [Bodo Moeller] 9196 9197 *) Rewrite the way password based encryption (PBE) is handled. It used to 9198 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 9199 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 9200 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 9201 the 'parameter' field of the AlgorithmIdentifier is passed to the 9202 underlying key generation function so it must do its own ASN1 parsing. 9203 This has also changed the EVP_PBE_CipherInit() function which now has a 9204 'parameter' argument instead of literal salt and iteration count values 9205 and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 9206 [Steve Henson] 9207 9208 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 9209 and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 9210 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 9211 KEY" because this clashed with PKCS#8 unencrypted string. Since this 9212 value was just used as a "magic string" and not used directly its 9213 value doesn't matter. 9214 [Steve Henson] 9215 9216 *) Introduce some semblance of const correctness to BN. Shame C doesn't 9217 support mutable. 9218 [Ben Laurie] 9219 9220 *) "linux-sparc64" configuration (ultrapenguin). 9221 [Ray Miller <ray.miller@oucs.ox.ac.uk>] 9222 "linux-sparc" configuration. 9223 [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 9224 9225 *) config now generates no-xxx options for missing ciphers. 9226 [Ulf M��ller] 9227 9228 *) Support the EBCDIC character set (work in progress). 9229 File ebcdic.c not yet included because it has a different license. 9230 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 9231 9232 *) Support BS2000/OSD-POSIX. 9233 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 9234 9235 *) Make callbacks for key generation use void * instead of char *. 9236 [Ben Laurie] 9237 9238 *) Make S/MIME samples compile (not yet tested). 9239 [Ben Laurie] 9240 9241 *) Additional typesafe stacks. 9242 [Ben Laurie] 9243 9244 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 9245 [Bodo Moeller] 9246 9247 9248 Changes between 0.9.3 and 0.9.3a [29 May 1999] 9249 9250 *) New configuration variant "sco5-gcc". 9251 9252 *) Updated some demos. 9253 [Sean O Riordain, Wade Scholine] 9254 9255 *) Add missing BIO_free at exit of pkcs12 application. 9256 [Wu Zhigang] 9257 9258 *) Fix memory leak in conf.c. 9259 [Steve Henson] 9260 9261 *) Updates for Win32 to assembler version of MD5. 9262 [Steve Henson] 9263 9264 *) Set #! path to perl in apps/der_chop to where we found it 9265 instead of using a fixed path. 9266 [Bodo Moeller] 9267 9268 *) SHA library changes for irix64-mips4-cc. 9269 [Andy Polyakov] 9270 9271 *) Improvements for VMS support. 9272 [Richard Levitte] 9273 9274 9275 Changes between 0.9.2b and 0.9.3 [24 May 1999] 9276 9277 *) Bignum library bug fix. IRIX 6 passes "make test" now! 9278 This also avoids the problems with SC4.2 and unpatched SC5. 9279 [Andy Polyakov <appro@fy.chalmers.se>] 9280 9281 *) New functions sk_num, sk_value and sk_set to replace the previous macros. 9282 These are required because of the typesafe stack would otherwise break 9283 existing code. If old code used a structure member which used to be STACK 9284 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 9285 sk_num or sk_value it would produce an error because the num, data members 9286 are not present in STACK_OF. Now it just produces a warning. sk_set 9287 replaces the old method of assigning a value to sk_value 9288 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 9289 that does this will no longer work (and should use sk_set instead) but 9290 this could be regarded as a "questionable" behaviour anyway. 9291 [Steve Henson] 9292 9293 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 9294 correctly handle encrypted S/MIME data. 9295 [Steve Henson] 9296 9297 *) Change type of various DES function arguments from des_cblock 9298 (which means, in function argument declarations, pointer to char) 9299 to des_cblock * (meaning pointer to array with 8 char elements), 9300 which allows the compiler to do more typechecking; it was like 9301 that back in SSLeay, but with lots of ugly casts. 9302 9303 Introduce new type const_des_cblock. 9304 [Bodo Moeller] 9305 9306 *) Reorganise the PKCS#7 library and get rid of some of the more obvious 9307 problems: find RecipientInfo structure that matches recipient certificate 9308 and initialise the ASN1 structures properly based on passed cipher. 9309 [Steve Henson] 9310 9311 *) Belatedly make the BN tests actually check the results. 9312 [Ben Laurie] 9313 9314 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 9315 to and from BNs: it was completely broken. New compilation option 9316 NEG_PUBKEY_BUG to allow for some broken certificates that encode public 9317 key elements as negative integers. 9318 [Steve Henson] 9319 9320 *) Reorganize and speed up MD5. 9321 [Andy Polyakov <appro@fy.chalmers.se>] 9322 9323 *) VMS support. 9324 [Richard Levitte <richard@levitte.org>] 9325 9326 *) New option -out to asn1parse to allow the parsed structure to be 9327 output to a file. This is most useful when combined with the -strparse 9328 option to examine the output of things like OCTET STRINGS. 9329 [Steve Henson] 9330 9331 *) Make SSL library a little more fool-proof by not requiring any longer 9332 that SSL_set_{accept,connect}_state be called before 9333 SSL_{accept,connect} may be used (SSL_set_..._state is omitted 9334 in many applications because usually everything *appeared* to work as 9335 intended anyway -- now it really works as intended). 9336 [Bodo Moeller] 9337 9338 *) Move openssl.cnf out of lib/. 9339 [Ulf M��ller] 9340 9341 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 9342 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 9343 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 9344 [Ralf S. Engelschall] 9345 9346 *) Various fixes to the EVP and PKCS#7 code. It may now be able to 9347 handle PKCS#7 enveloped data properly. 9348 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 9349 9350 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 9351 copying pointers. The cert_st handling is changed by this in 9352 various ways (and thus what used to be known as ctx->default_cert 9353 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 9354 any longer when s->cert does not give us what we need). 9355 ssl_cert_instantiate becomes obsolete by this change. 9356 As soon as we've got the new code right (possibly it already is?), 9357 we have solved a couple of bugs of the earlier code where s->cert 9358 was used as if it could not have been shared with other SSL structures. 9359 9360 Note that using the SSL API in certain dirty ways now will result 9361 in different behaviour than observed with earlier library versions: 9362 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 9363 does not influence s as it used to. 9364 9365 In order to clean up things more thoroughly, inside SSL_SESSION 9366 we don't use CERT any longer, but a new structure SESS_CERT 9367 that holds per-session data (if available); currently, this is 9368 the peer's certificate chain and, for clients, the server's certificate 9369 and temporary key. CERT holds only those values that can have 9370 meaningful defaults in an SSL_CTX. 9371 [Bodo Moeller] 9372 9373 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 9374 from the internal representation. Various PKCS#7 fixes: remove some 9375 evil casts and set the enc_dig_alg field properly based on the signing 9376 key type. 9377 [Steve Henson] 9378 9379 *) Allow PKCS#12 password to be set from the command line or the 9380 environment. Let 'ca' get its config file name from the environment 9381 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 9382 and 'x509'). 9383 [Steve Henson] 9384 9385 *) Allow certificate policies extension to use an IA5STRING for the 9386 organization field. This is contrary to the PKIX definition but 9387 VeriSign uses it and IE5 only recognises this form. Document 'x509' 9388 extension option. 9389 [Steve Henson] 9390 9391 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 9392 without disallowing inline assembler and the like for non-pedantic builds. 9393 [Ben Laurie] 9394 9395 *) Support Borland C++ builder. 9396 [Janez Jere <jj@void.si>, modified by Ulf M��ller] 9397 9398 *) Support Mingw32. 9399 [Ulf M��ller] 9400 9401 *) SHA-1 cleanups and performance enhancements. 9402 [Andy Polyakov <appro@fy.chalmers.se>] 9403 9404 *) Sparc v8plus assembler for the bignum library. 9405 [Andy Polyakov <appro@fy.chalmers.se>] 9406 9407 *) Accept any -xxx and +xxx compiler options in Configure. 9408 [Ulf M��ller] 9409 9410 *) Update HPUX configuration. 9411 [Anonymous] 9412 9413 *) Add missing sk_<type>_unshift() function to safestack.h 9414 [Ralf S. Engelschall] 9415 9416 *) New function SSL_CTX_use_certificate_chain_file that sets the 9417 "extra_cert"s in addition to the certificate. (This makes sense 9418 only for "PEM" format files, as chains as a whole are not 9419 DER-encoded.) 9420 [Bodo Moeller] 9421 9422 *) Support verify_depth from the SSL API. 9423 x509_vfy.c had what can be considered an off-by-one-error: 9424 Its depth (which was not part of the external interface) 9425 was actually counting the number of certificates in a chain; 9426 now it really counts the depth. 9427 [Bodo Moeller] 9428 9429 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 9430 instead of X509err, which often resulted in confusing error 9431 messages since the error codes are not globally unique 9432 (e.g. an alleged error in ssl3_accept when a certificate 9433 didn't match the private key). 9434 9435 *) New function SSL_CTX_set_session_id_context that allows to set a default 9436 value (so that you don't need SSL_set_session_id_context for each 9437 connection using the SSL_CTX). 9438 [Bodo Moeller] 9439 9440 *) OAEP decoding bug fix. 9441 [Ulf M��ller] 9442 9443 *) Support INSTALL_PREFIX for package builders, as proposed by 9444 David Harris. 9445 [Bodo Moeller] 9446 9447 *) New Configure options "threads" and "no-threads". For systems 9448 where the proper compiler options are known (currently Solaris 9449 and Linux), "threads" is the default. 9450 [Bodo Moeller] 9451 9452 *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 9453 [Bodo Moeller] 9454 9455 *) Install various scripts to $(OPENSSLDIR)/misc, not to 9456 $(INSTALLTOP)/bin -- they shouldn't clutter directories 9457 such as /usr/local/bin. 9458 [Bodo Moeller] 9459 9460 *) "make linux-shared" to build shared libraries. 9461 [Niels Poppe <niels@netbox.org>] 9462 9463 *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 9464 [Ulf M��ller] 9465 9466 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 9467 extension adding in x509 utility. 9468 [Steve Henson] 9469 9470 *) Remove NOPROTO sections and error code comments. 9471 [Ulf M��ller] 9472 9473 *) Partial rewrite of the DEF file generator to now parse the ANSI 9474 prototypes. 9475 [Steve Henson] 9476 9477 *) New Configure options --prefix=DIR and --openssldir=DIR. 9478 [Ulf M��ller] 9479 9480 *) Complete rewrite of the error code script(s). It is all now handled 9481 by one script at the top level which handles error code gathering, 9482 header rewriting and C source file generation. It should be much better 9483 than the old method: it now uses a modified version of Ulf's parser to 9484 read the ANSI prototypes in all header files (thus the old K&R definitions 9485 aren't needed for error creation any more) and do a better job of 9486 translating function codes into names. The old 'ASN1 error code imbedded 9487 in a comment' is no longer necessary and it doesn't use .err files which 9488 have now been deleted. Also the error code call doesn't have to appear all 9489 on one line (which resulted in some large lines...). 9490 [Steve Henson] 9491 9492 *) Change #include filenames from <foo.h> to <openssl/foo.h>. 9493 [Bodo Moeller] 9494 9495 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 9496 0 (which usually indicates a closed connection), but continue reading. 9497 [Bodo Moeller] 9498 9499 *) Fix some race conditions. 9500 [Bodo Moeller] 9501 9502 *) Add support for CRL distribution points extension. Add Certificate 9503 Policies and CRL distribution points documentation. 9504 [Steve Henson] 9505 9506 *) Move the autogenerated header file parts to crypto/opensslconf.h. 9507 [Ulf M��ller] 9508 9509 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 9510 8 of keying material. Merlin has also confirmed interop with this fix 9511 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 9512 [Merlin Hughes <merlin@baltimore.ie>] 9513 9514 *) Fix lots of warnings. 9515 [Richard Levitte <levitte@stacken.kth.se>] 9516 9517 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 9518 the directory spec didn't end with a LIST_SEPARATOR_CHAR. 9519 [Richard Levitte <levitte@stacken.kth.se>] 9520 9521 *) Fix problems with sizeof(long) == 8. 9522 [Andy Polyakov <appro@fy.chalmers.se>] 9523 9524 *) Change functions to ANSI C. 9525 [Ulf M��ller] 9526 9527 *) Fix typos in error codes. 9528 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M��ller] 9529 9530 *) Remove defunct assembler files from Configure. 9531 [Ulf M��ller] 9532 9533 *) SPARC v8 assembler BIGNUM implementation. 9534 [Andy Polyakov <appro@fy.chalmers.se>] 9535 9536 *) Support for Certificate Policies extension: both print and set. 9537 Various additions to support the r2i method this uses. 9538 [Steve Henson] 9539 9540 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 9541 return a const string when you are expecting an allocated buffer. 9542 [Ben Laurie] 9543 9544 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 9545 types DirectoryString and DisplayText. 9546 [Steve Henson] 9547 9548 *) Add code to allow r2i extensions to access the configuration database, 9549 add an LHASH database driver and add several ctx helper functions. 9550 [Steve Henson] 9551 9552 *) Fix an evil bug in bn_expand2() which caused various BN functions to 9553 fail when they extended the size of a BIGNUM. 9554 [Steve Henson] 9555 9556 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 9557 support typesafe stack. 9558 [Steve Henson] 9559 9560 *) Fix typo in SSL_[gs]et_options(). 9561 [Nils Frostberg <nils@medcom.se>] 9562 9563 *) Delete various functions and files that belonged to the (now obsolete) 9564 old X509V3 handling code. 9565 [Steve Henson] 9566 9567 *) New Configure option "rsaref". 9568 [Ulf M��ller] 9569 9570 *) Don't auto-generate pem.h. 9571 [Bodo Moeller] 9572 9573 *) Introduce type-safe ASN.1 SETs. 9574 [Ben Laurie] 9575 9576 *) Convert various additional casted stacks to type-safe STACK_OF() variants. 9577 [Ben Laurie, Ralf S. Engelschall, Steve Henson] 9578 9579 *) Introduce type-safe STACKs. This will almost certainly break lots of code 9580 that links with OpenSSL (well at least cause lots of warnings), but fear 9581 not: the conversion is trivial, and it eliminates loads of evil casts. A 9582 few STACKed things have been converted already. Feel free to convert more. 9583 In the fullness of time, I'll do away with the STACK type altogether. 9584 [Ben Laurie] 9585 9586 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 9587 specified in <certfile> by updating the entry in the index.txt file. 9588 This way one no longer has to edit the index.txt file manually for 9589 revoking a certificate. The -revoke option does the gory details now. 9590 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 9591 9592 *) Fix `openssl crl -noout -text' combination where `-noout' killed the 9593 `-text' option at all and this way the `-noout -text' combination was 9594 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 9595 [Ralf S. Engelschall] 9596 9597 *) Make sure a corresponding plain text error message exists for the 9598 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 9599 verify callback function determined that a certificate was revoked. 9600 [Ralf S. Engelschall] 9601 9602 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 9603 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 9604 all available cipers including rc5, which was forgotten until now. 9605 In order to let the testing shell script know which algorithms 9606 are available, a new (up to now undocumented) command 9607 "openssl list-cipher-commands" is used. 9608 [Bodo Moeller] 9609 9610 *) Bugfix: s_client occasionally would sleep in select() when 9611 it should have checked SSL_pending() first. 9612 [Bodo Moeller] 9613 9614 *) New functions DSA_do_sign and DSA_do_verify to provide access to 9615 the raw DSA values prior to ASN.1 encoding. 9616 [Ulf M��ller] 9617 9618 *) Tweaks to Configure 9619 [Niels Poppe <niels@netbox.org>] 9620 9621 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 9622 yet... 9623 [Steve Henson] 9624 9625 *) New variables $(RANLIB) and $(PERL) in the Makefiles. 9626 [Ulf M��ller] 9627 9628 *) New config option to avoid instructions that are illegal on the 80386. 9629 The default code is faster, but requires at least a 486. 9630 [Ulf M��ller] 9631 9632 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 9633 SSL2_SERVER_VERSION (not used at all) macros, which are now the 9634 same as SSL2_VERSION anyway. 9635 [Bodo Moeller] 9636 9637 *) New "-showcerts" option for s_client. 9638 [Bodo Moeller] 9639 9640 *) Still more PKCS#12 integration. Add pkcs12 application to openssl 9641 application. Various cleanups and fixes. 9642 [Steve Henson] 9643 9644 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 9645 modify error routines to work internally. Add error codes and PBE init 9646 to library startup routines. 9647 [Steve Henson] 9648 9649 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 9650 packing functions to asn1 and evp. Changed function names and error 9651 codes along the way. 9652 [Steve Henson] 9653 9654 *) PKCS12 integration: and so it begins... First of several patches to 9655 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 9656 objects to objects.h 9657 [Steve Henson] 9658 9659 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 9660 and display support for Thawte strong extranet extension. 9661 [Steve Henson] 9662 9663 *) Add LinuxPPC support. 9664 [Jeff Dubrule <igor@pobox.org>] 9665 9666 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 9667 bn_div_words in alpha.s. 9668 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 9669 9670 *) Make sure the RSA OAEP test is skipped under -DRSAref because 9671 OAEP isn't supported when OpenSSL is built with RSAref. 9672 [Ulf Moeller <ulf@fitug.de>] 9673 9674 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 9675 so they no longer are missing under -DNOPROTO. 9676 [Soren S. Jorvang <soren@t.dk>] 9677 9678 9679 Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 9680 9681 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 9682 doesn't work when the session is reused. Coming soon! 9683 [Ben Laurie] 9684 9685 *) Fix a security hole, that allows sessions to be reused in the wrong 9686 context thus bypassing client cert protection! All software that uses 9687 client certs and session caches in multiple contexts NEEDS PATCHING to 9688 allow session reuse! A fuller solution is in the works. 9689 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 9690 9691 *) Some more source tree cleanups (removed obsolete files 9692 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 9693 permission on "config" script to be executable) and a fix for the INSTALL 9694 document. 9695 [Ulf Moeller <ulf@fitug.de>] 9696 9697 *) Remove some legacy and erroneous uses of malloc, free instead of 9698 Malloc, Free. 9699 [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 9700 9701 *) Make rsa_oaep_test return non-zero on error. 9702 [Ulf Moeller <ulf@fitug.de>] 9703 9704 *) Add support for native Solaris shared libraries. Configure 9705 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 9706 if someone would make that last step automatic. 9707 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 9708 9709 *) ctx_size was not built with the right compiler during "make links". Fixed. 9710 [Ben Laurie] 9711 9712 *) Change the meaning of 'ALL' in the cipher list. It now means "everything 9713 except NULL ciphers". This means the default cipher list will no longer 9714 enable NULL ciphers. They need to be specifically enabled e.g. with 9715 the string "DEFAULT:eNULL". 9716 [Steve Henson] 9717 9718 *) Fix to RSA private encryption routines: if p < q then it would 9719 occasionally produce an invalid result. This will only happen with 9720 externally generated keys because OpenSSL (and SSLeay) ensure p > q. 9721 [Steve Henson] 9722 9723 *) Be less restrictive and allow also `perl util/perlpath.pl 9724 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 9725 because this way one can also use an interpreter named `perl5' (which is 9726 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 9727 installed as `perl'). 9728 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 9729 9730 *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 9731 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 9732 9733 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 9734 advapi32.lib to Win32 build and change the pem test comparision 9735 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 9736 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 9737 and crypto/des/ede_cbcm_enc.c. 9738 [Steve Henson] 9739 9740 *) DES quad checksum was broken on big-endian architectures. Fixed. 9741 [Ben Laurie] 9742 9743 *) Comment out two functions in bio.h that aren't implemented. Fix up the 9744 Win32 test batch file so it (might) work again. The Win32 test batch file 9745 is horrible: I feel ill.... 9746 [Steve Henson] 9747 9748 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 9749 in e_os.h. Audit of header files to check ANSI and non ANSI 9750 sections: 10 functions were absent from non ANSI section and not exported 9751 from Windows DLLs. Fixed up libeay.num for new functions. 9752 [Steve Henson] 9753 9754 *) Make `openssl version' output lines consistent. 9755 [Ralf S. Engelschall] 9756 9757 *) Fix Win32 symbol export lists for BIO functions: Added 9758 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 9759 to ms/libeay{16,32}.def. 9760 [Ralf S. Engelschall] 9761 9762 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 9763 fine under Unix and passes some trivial tests I've now added. But the 9764 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 9765 added to make sure no one expects that this stuff really works in the 9766 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 9767 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 9768 openssl_bio.xs. 9769 [Ralf S. Engelschall] 9770 9771 *) Fix the generation of two part addresses in perl. 9772 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 9773 9774 *) Add config entry for Linux on MIPS. 9775 [John Tobey <jtobey@channel1.com>] 9776 9777 *) Make links whenever Configure is run, unless we are on Windoze. 9778 [Ben Laurie] 9779 9780 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 9781 Currently only issuerAltName and AuthorityKeyIdentifier make any sense 9782 in CRLs. 9783 [Steve Henson] 9784 9785 *) Add a useful kludge to allow package maintainers to specify compiler and 9786 other platforms details on the command line without having to patch the 9787 Configure script everytime: One now can use ``perl Configure 9788 <id>:<details>'', i.e. platform ids are allowed to have details appended 9789 to them (seperated by colons). This is treated as there would be a static 9790 pre-configured entry in Configure's %table under key <id> with value 9791 <details> and ``perl Configure <id>'' is called. So, when you want to 9792 perform a quick test-compile under FreeBSD 3.1 with pgcc and without 9793 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 9794 now, which overrides the FreeBSD-elf entry on-the-fly. 9795 [Ralf S. Engelschall] 9796 9797 *) Disable new TLS1 ciphersuites by default: they aren't official yet. 9798 [Ben Laurie] 9799 9800 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 9801 on the `perl Configure ...' command line. This way one can compile 9802 OpenSSL libraries with Position Independent Code (PIC) which is needed 9803 for linking it into DSOs. 9804 [Ralf S. Engelschall] 9805 9806 *) Remarkably, export ciphers were totally broken and no-one had noticed! 9807 Fixed. 9808 [Ben Laurie] 9809 9810 *) Cleaned up the LICENSE document: The official contact for any license 9811 questions now is the OpenSSL core team under openssl-core@openssl.org. 9812 And add a paragraph about the dual-license situation to make sure people 9813 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 9814 to the OpenSSL toolkit. 9815 [Ralf S. Engelschall] 9816 9817 *) General source tree makefile cleanups: Made `making xxx in yyy...' 9818 display consistent in the source tree and replaced `/bin/rm' by `rm'. 9819 Additonally cleaned up the `make links' target: Remove unnecessary 9820 semicolons, subsequent redundant removes, inline point.sh into mklink.sh 9821 to speed processing and no longer clutter the display with confusing 9822 stuff. Instead only the actually done links are displayed. 9823 [Ralf S. Engelschall] 9824 9825 *) Permit null encryption ciphersuites, used for authentication only. It used 9826 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 9827 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 9828 encryption. 9829 [Ben Laurie] 9830 9831 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 9832 signed attributes when verifying signatures (this would break them), 9833 the detached data encoding was wrong and public keys obtained using 9834 X509_get_pubkey() weren't freed. 9835 [Steve Henson] 9836 9837 *) Add text documentation for the BUFFER functions. Also added a work around 9838 to a Win95 console bug. This was triggered by the password read stuff: the 9839 last character typed gets carried over to the next fread(). If you were 9840 generating a new cert request using 'req' for example then the last 9841 character of the passphrase would be CR which would then enter the first 9842 field as blank. 9843 [Steve Henson] 9844 9845 *) Added the new `Includes OpenSSL Cryptography Software' button as 9846 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 9847 button and can be used by applications based on OpenSSL to show the 9848 relationship to the OpenSSL project. 9849 [Ralf S. Engelschall] 9850 9851 *) Remove confusing variables in function signatures in files 9852 ssl/ssl_lib.c and ssl/ssl.h. 9853 [Lennart Bong <lob@kulthea.stacken.kth.se>] 9854 9855 *) Don't install bss_file.c under PREFIX/include/ 9856 [Lennart Bong <lob@kulthea.stacken.kth.se>] 9857 9858 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 9859 functions that return function pointers and has support for NT specific 9860 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 9861 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 9862 unsigned to signed types: this was killing the Win32 compile. 9863 [Steve Henson] 9864 9865 *) Add new certificate file to stack functions, 9866 SSL_add_dir_cert_subjects_to_stack() and 9867 SSL_add_file_cert_subjects_to_stack(). These largely supplant 9868 SSL_load_client_CA_file(), and can be used to add multiple certs easily 9869 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 9870 This means that Apache-SSL and similar packages don't have to mess around 9871 to add as many CAs as they want to the preferred list. 9872 [Ben Laurie] 9873 9874 *) Experiment with doxygen documentation. Currently only partially applied to 9875 ssl/ssl_lib.c. 9876 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 9877 openssl.doxy as the configuration file. 9878 [Ben Laurie] 9879 9880 *) Get rid of remaining C++-style comments which strict C compilers hate. 9881 [Ralf S. Engelschall, pointed out by Carlos Amengual] 9882 9883 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 9884 compiled in by default: it has problems with large keys. 9885 [Steve Henson] 9886 9887 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 9888 DH private keys and/or callback functions which directly correspond to 9889 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 9890 is needed for applications which have to configure certificates on a 9891 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 9892 (e.g. s_server). 9893 For the RSA certificate situation is makes no difference, but 9894 for the DSA certificate situation this fixes the "no shared cipher" 9895 problem where the OpenSSL cipher selection procedure failed because the 9896 temporary keys were not overtaken from the context and the API provided 9897 no way to reconfigure them. 9898 The new functions now let applications reconfigure the stuff and they 9899 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 9900 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 9901 non-public-API function ssl_cert_instantiate() is used as a helper 9902 function and also to reduce code redundancy inside ssl_rsa.c. 9903 [Ralf S. Engelschall] 9904 9905 *) Move s_server -dcert and -dkey options out of the undocumented feature 9906 area because they are useful for the DSA situation and should be 9907 recognized by the users. 9908 [Ralf S. Engelschall] 9909 9910 *) Fix the cipher decision scheme for export ciphers: the export bits are 9911 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 9912 SSL_EXP_MASK. So, the original variable has to be used instead of the 9913 already masked variable. 9914 [Richard Levitte <levitte@stacken.kth.se>] 9915 9916 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 9917 [Richard Levitte <levitte@stacken.kth.se>] 9918 9919 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 9920 from `int' to `unsigned int' because it's a length and initialized by 9921 EVP_DigestFinal() which expects an `unsigned int *'. 9922 [Richard Levitte <levitte@stacken.kth.se>] 9923 9924 *) Don't hard-code path to Perl interpreter on shebang line of Configure 9925 script. Instead use the usual Shell->Perl transition trick. 9926 [Ralf S. Engelschall] 9927 9928 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 9929 (in addition to RSA certificates) to match the behaviour of `openssl dsa 9930 -noout -modulus' as it's already the case for `openssl rsa -noout 9931 -modulus'. For RSA the -modulus is the real "modulus" while for DSA 9932 currently the public key is printed (a decision which was already done by 9933 `openssl dsa -modulus' in the past) which serves a similar purpose. 9934 Additionally the NO_RSA no longer completely removes the whole -modulus 9935 option; it now only avoids using the RSA stuff. Same applies to NO_DSA 9936 now, too. 9937 [Ralf S. Engelschall] 9938 9939 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 9940 BIO. See the source (crypto/evp/bio_ok.c) for more info. 9941 [Arne Ansper <arne@ats.cyber.ee>] 9942 9943 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 9944 to be added. Now both 'req' and 'ca' can use new objects defined in the 9945 config file. 9946 [Steve Henson] 9947 9948 *) Add cool BIO that does syslog (or event log on NT). 9949 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 9950 9951 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 9952 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 9953 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 9954 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 9955 [Ben Laurie] 9956 9957 *) Add preliminary config info for new extension code. 9958 [Steve Henson] 9959 9960 *) Make RSA_NO_PADDING really use no padding. 9961 [Ulf Moeller <ulf@fitug.de>] 9962 9963 *) Generate errors when private/public key check is done. 9964 [Ben Laurie] 9965 9966 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 9967 for some CRL extensions and new objects added. 9968 [Steve Henson] 9969 9970 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 9971 key usage extension and fuller support for authority key id. 9972 [Steve Henson] 9973 9974 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 9975 padding method for RSA, which is recommended for new applications in PKCS 9976 #1 v2.0 (RFC 2437, October 1998). 9977 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 9978 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 9979 against Bleichbacher's attack on RSA. 9980 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 9981 Ben Laurie] 9982 9983 *) Updates to the new SSL compression code 9984 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 9985 9986 *) Fix so that the version number in the master secret, when passed 9987 via RSA, checks that if TLS was proposed, but we roll back to SSLv3 9988 (because the server will not accept higher), that the version number 9989 is 0x03,0x01, not 0x03,0x00 9990 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 9991 9992 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 9993 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 9994 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 9995 [Steve Henson] 9996 9997 *) Support for RAW extensions where an arbitrary extension can be 9998 created by including its DER encoding. See apps/openssl.cnf for 9999 an example. 10000 [Steve Henson] 10001 10002 *) Make sure latest Perl versions don't interpret some generated C array 10003 code as Perl array code in the crypto/err/err_genc.pl script. 10004 [Lars Weber <3weber@informatik.uni-hamburg.de>] 10005 10006 *) Modify ms/do_ms.bat to not generate assembly language makefiles since 10007 not many people have the assembler. Various Win32 compilation fixes and 10008 update to the INSTALL.W32 file with (hopefully) more accurate Win32 10009 build instructions. 10010 [Steve Henson] 10011 10012 *) Modify configure script 'Configure' to automatically create crypto/date.h 10013 file under Win32 and also build pem.h from pem.org. New script 10014 util/mkfiles.pl to create the MINFO file on environments that can't do a 10015 'make files': perl util/mkfiles.pl >MINFO should work. 10016 [Steve Henson] 10017 10018 *) Major rework of DES function declarations, in the pursuit of correctness 10019 and purity. As a result, many evil casts evaporated, and some weirdness, 10020 too. You may find this causes warnings in your code. Zapping your evil 10021 casts will probably fix them. Mostly. 10022 [Ben Laurie] 10023 10024 *) Fix for a typo in asn1.h. Bug fix to object creation script 10025 obj_dat.pl. It considered a zero in an object definition to mean 10026 "end of object": none of the objects in objects.h have any zeros 10027 so it wasn't spotted. 10028 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 10029 10030 *) Add support for Triple DES Cipher Block Chaining with Output Feedback 10031 Masking (CBCM). In the absence of test vectors, the best I have been able 10032 to do is check that the decrypt undoes the encrypt, so far. Send me test 10033 vectors if you have them. 10034 [Ben Laurie] 10035 10036 *) Correct calculation of key length for export ciphers (too much space was 10037 allocated for null ciphers). This has not been tested! 10038 [Ben Laurie] 10039 10040 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 10041 message is now correct (it understands "crypto" and "ssl" on its 10042 command line). There is also now an "update" option. This will update 10043 the util/ssleay.num and util/libeay.num files with any new functions. 10044 If you do a: 10045 perl util/mkdef.pl crypto ssl update 10046 it will update them. 10047 [Steve Henson] 10048 10049 *) Overhauled the Perl interface (perl/*): 10050 - ported BN stuff to OpenSSL's different BN library 10051 - made the perl/ source tree CVS-aware 10052 - renamed the package from SSLeay to OpenSSL (the files still contain 10053 their history because I've copied them in the repository) 10054 - removed obsolete files (the test scripts will be replaced 10055 by better Test::Harness variants in the future) 10056 [Ralf S. Engelschall] 10057 10058 *) First cut for a very conservative source tree cleanup: 10059 1. merge various obsolete readme texts into doc/ssleay.txt 10060 where we collect the old documents and readme texts. 10061 2. remove the first part of files where I'm already sure that we no 10062 longer need them because of three reasons: either they are just temporary 10063 files which were left by Eric or they are preserved original files where 10064 I've verified that the diff is also available in the CVS via "cvs diff 10065 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 10066 the crypto/md/ stuff). 10067 [Ralf S. Engelschall] 10068 10069 *) More extension code. Incomplete support for subject and issuer alt 10070 name, issuer and authority key id. Change the i2v function parameters 10071 and add an extra 'crl' parameter in the X509V3_CTX structure: guess 10072 what that's for :-) Fix to ASN1 macro which messed up 10073 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 10074 [Steve Henson] 10075 10076 *) Preliminary support for ENUMERATED type. This is largely copied from the 10077 INTEGER code. 10078 [Steve Henson] 10079 10080 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 10081 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 10082 10083 *) Make sure `make rehash' target really finds the `openssl' program. 10084 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 10085 10086 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 10087 like to hear about it if this slows down other processors. 10088 [Ben Laurie] 10089 10090 *) Add CygWin32 platform information to Configure script. 10091 [Alan Batie <batie@aahz.jf.intel.com>] 10092 10093 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 10094 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 10095 10096 *) New program nseq to manipulate netscape certificate sequences 10097 [Steve Henson] 10098 10099 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 10100 few typos. 10101 [Steve Henson] 10102 10103 *) Fixes to BN code. Previously the default was to define BN_RECURSION 10104 but the BN code had some problems that would cause failures when 10105 doing certificate verification and some other functions. 10106 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 10107 10108 *) Add ASN1 and PEM code to support netscape certificate sequences. 10109 [Steve Henson] 10110 10111 *) Add ASN1 and PEM code to support netscape certificate sequences. 10112 [Steve Henson] 10113 10114 *) Add several PKIX and private extended key usage OIDs. 10115 [Steve Henson] 10116 10117 *) Modify the 'ca' program to handle the new extension code. Modify 10118 openssl.cnf for new extension format, add comments. 10119 [Steve Henson] 10120 10121 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 10122 and add a sample to openssl.cnf so req -x509 now adds appropriate 10123 CA extensions. 10124 [Steve Henson] 10125 10126 *) Continued X509 V3 changes. Add to other makefiles, integrate with the 10127 error code, add initial support to X509_print() and x509 application. 10128 [Steve Henson] 10129 10130 *) Takes a deep breath and start addding X509 V3 extension support code. Add 10131 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 10132 stuff is currently isolated and isn't even compiled yet. 10133 [Steve Henson] 10134 10135 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 10136 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 10137 Removed the versions check from X509 routines when loading extensions: 10138 this allows certain broken certificates that don't set the version 10139 properly to be processed. 10140 [Steve Henson] 10141 10142 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 10143 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 10144 can still be regenerated with "make depend". 10145 [Ben Laurie] 10146 10147 *) Spelling mistake in C version of CAST-128. 10148 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 10149 10150 *) Changes to the error generation code. The perl script err-code.pl 10151 now reads in the old error codes and retains the old numbers, only 10152 adding new ones if necessary. It also only changes the .err files if new 10153 codes are added. The makefiles have been modified to only insert errors 10154 when needed (to avoid needlessly modifying header files). This is done 10155 by only inserting errors if the .err file is newer than the auto generated 10156 C file. To rebuild all the error codes from scratch (the old behaviour) 10157 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 10158 or delete all the .err files. 10159 [Steve Henson] 10160 10161 *) CAST-128 was incorrectly implemented for short keys. The C version has 10162 been fixed, but is untested. The assembler versions are also fixed, but 10163 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 10164 to regenerate it if needed. 10165 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 10166 Hagino <itojun@kame.net>] 10167 10168 *) File was opened incorrectly in randfile.c. 10169 [Ulf M��ller <ulf@fitug.de>] 10170 10171 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 10172 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 10173 GeneralizedTime. ASN1_TIME is the proper type used in certificates et 10174 al: it's just almost always a UTCTime. Note this patch adds new error 10175 codes so do a "make errors" if there are problems. 10176 [Steve Henson] 10177 10178 *) Correct Linux 1 recognition in config. 10179 [Ulf M��ller <ulf@fitug.de>] 10180 10181 *) Remove pointless MD5 hash when using DSA keys in ca. 10182 [Anonymous <nobody@replay.com>] 10183 10184 *) Generate an error if given an empty string as a cert directory. Also 10185 generate an error if handed NULL (previously returned 0 to indicate an 10186 error, but didn't set one). 10187 [Ben Laurie, reported by Anonymous <nobody@replay.com>] 10188 10189 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 10190 [Ben Laurie] 10191 10192 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 10193 parameters. This was causing a warning which killed off the Win32 compile. 10194 [Steve Henson] 10195 10196 *) Remove C++ style comments from crypto/bn/bn_local.h. 10197 [Neil Costigan <neil.costigan@celocom.com>] 10198 10199 *) The function OBJ_txt2nid was broken. It was supposed to return a nid 10200 based on a text string, looking up short and long names and finally 10201 "dot" format. The "dot" format stuff didn't work. Added new function 10202 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 10203 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 10204 OID is not part of the table. 10205 [Steve Henson] 10206 10207 *) Add prototypes to X509 lookup/verify methods, fixing a bug in 10208 X509_LOOKUP_by_alias(). 10209 [Ben Laurie] 10210 10211 *) Sort openssl functions by name. 10212 [Ben Laurie] 10213 10214 *) Get the gendsa program working (hopefully) and add it to app list. Remove 10215 encryption from sample DSA keys (in case anyone is interested the password 10216 was "1234"). 10217 [Steve Henson] 10218 10219 *) Make _all_ *_free functions accept a NULL pointer. 10220 [Frans Heymans <fheymans@isaserver.be>] 10221 10222 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 10223 NULL pointers. 10224 [Anonymous <nobody@replay.com>] 10225 10226 *) s_server should send the CAfile as acceptable CAs, not its own cert. 10227 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 10228 10229 *) Don't blow it for numeric -newkey arguments to apps/req. 10230 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 10231 10232 *) Temp key "for export" tests were wrong in s3_srvr.c. 10233 [Anonymous <nobody@replay.com>] 10234 10235 *) Add prototype for temp key callback functions 10236 SSL_CTX_set_tmp_{rsa,dh}_callback(). 10237 [Ben Laurie] 10238 10239 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 10240 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 10241 [Steve Henson] 10242 10243 *) X509_name_add_entry() freed the wrong thing after an error. 10244 [Arne Ansper <arne@ats.cyber.ee>] 10245 10246 *) rsa_eay.c would attempt to free a NULL context. 10247 [Arne Ansper <arne@ats.cyber.ee>] 10248 10249 *) BIO_s_socket() had a broken should_retry() on Windoze. 10250 [Arne Ansper <arne@ats.cyber.ee>] 10251 10252 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 10253 [Arne Ansper <arne@ats.cyber.ee>] 10254 10255 *) Make sure the already existing X509_STORE->depth variable is initialized 10256 in X509_STORE_new(), but document the fact that this variable is still 10257 unused in the certificate verification process. 10258 [Ralf S. Engelschall] 10259 10260 *) Fix the various library and apps files to free up pkeys obtained from 10261 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 10262 [Steve Henson] 10263 10264 *) Fix reference counting in X509_PUBKEY_get(). This makes 10265 demos/maurice/example2.c work, amongst others, probably. 10266 [Steve Henson and Ben Laurie] 10267 10268 *) First cut of a cleanup for apps/. First the `ssleay' program is now named 10269 `openssl' and second, the shortcut symlinks for the `openssl <command>' 10270 are no longer created. This way we have a single and consistent command 10271 line interface `openssl <command>', similar to `cvs <command>'. 10272 [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 10273 10274 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 10275 BIT STRING wrapper always have zero unused bits. 10276 [Steve Henson] 10277 10278 *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 10279 [Steve Henson] 10280 10281 *) Make the top-level INSTALL documentation easier to understand. 10282 [Paul Sutton] 10283 10284 *) Makefiles updated to exit if an error occurs in a sub-directory 10285 make (including if user presses ^C) [Paul Sutton] 10286 10287 *) Make Montgomery context stuff explicit in RSA data structure. 10288 [Ben Laurie] 10289 10290 *) Fix build order of pem and err to allow for generated pem.h. 10291 [Ben Laurie] 10292 10293 *) Fix renumbering bug in X509_NAME_delete_entry(). 10294 [Ben Laurie] 10295 10296 *) Enhanced the err-ins.pl script so it makes the error library number 10297 global and can add a library name. This is needed for external ASN1 and 10298 other error libraries. 10299 [Steve Henson] 10300 10301 *) Fixed sk_insert which never worked properly. 10302 [Steve Henson] 10303 10304 *) Fix ASN1 macros so they can handle indefinite length construted 10305 EXPLICIT tags. Some non standard certificates use these: they can now 10306 be read in. 10307 [Steve Henson] 10308 10309 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 10310 into a single doc/ssleay.txt bundle. This way the information is still 10311 preserved but no longer messes up this directory. Now it's new room for 10312 the new set of documenation files. 10313 [Ralf S. Engelschall] 10314 10315 *) SETs were incorrectly DER encoded. This was a major pain, because they 10316 shared code with SEQUENCEs, which aren't coded the same. This means that 10317 almost everything to do with SETs or SEQUENCEs has either changed name or 10318 number of arguments. 10319 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 10320 10321 *) Fix test data to work with the above. 10322 [Ben Laurie] 10323 10324 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 10325 was already fixed by Eric for 0.9.1 it seems. 10326 [Ben Laurie - pointed out by Ulf M��ller <ulf@fitug.de>] 10327 10328 *) Autodetect FreeBSD3. 10329 [Ben Laurie] 10330 10331 *) Fix various bugs in Configure. This affects the following platforms: 10332 nextstep 10333 ncr-scde 10334 unixware-2.0 10335 unixware-2.0-pentium 10336 sco5-cc. 10337 [Ben Laurie] 10338 10339 *) Eliminate generated files from CVS. Reorder tests to regenerate files 10340 before they are needed. 10341 [Ben Laurie] 10342 10343 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 10344 [Ben Laurie] 10345 10346 10347 Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 10348 10349 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 10350 changed SSLeay to OpenSSL in version strings. 10351 [Ralf S. Engelschall] 10352 10353 *) Some fixups to the top-level documents. 10354 [Paul Sutton] 10355 10356 *) Fixed the nasty bug where rsaref.h was not found under compile-time 10357 because the symlink to include/ was missing. 10358 [Ralf S. Engelschall] 10359 10360 *) Incorporated the popular no-RSA/DSA-only patches 10361 which allow to compile a RSA-free SSLeay. 10362 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 10363 10364 *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 10365 when "ssleay" is still not found. 10366 [Ralf S. Engelschall] 10367 10368 *) Added more platforms to Configure: Cray T3E, HPUX 11, 10369 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 10370 10371 *) Updated the README file. 10372 [Ralf S. Engelschall] 10373 10374 *) Added various .cvsignore files in the CVS repository subdirs 10375 to make a "cvs update" really silent. 10376 [Ralf S. Engelschall] 10377 10378 *) Recompiled the error-definition header files and added 10379 missing symbols to the Win32 linker tables. 10380 [Ralf S. Engelschall] 10381 10382 *) Cleaned up the top-level documents; 10383 o new files: CHANGES and LICENSE 10384 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 10385 o merged COPYRIGHT into LICENSE 10386 o removed obsolete TODO file 10387 o renamed MICROSOFT to INSTALL.W32 10388 [Ralf S. Engelschall] 10389 10390 *) Removed dummy files from the 0.9.1b source tree: 10391 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 10392 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 10393 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 10394 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 10395 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 10396 [Ralf S. Engelschall] 10397 10398 *) Added various platform portability fixes. 10399 [Mark J. Cox] 10400 10401 *) The Genesis of the OpenSSL rpject: 10402 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 10403 Young and Tim J. Hudson created while they were working for C2Net until 10404 summer 1998. 10405 [The OpenSSL Project] 10406 10407 10408 Changes between 0.9.0b and 0.9.1b [not released] 10409 10410 *) Updated a few CA certificates under certs/ 10411 [Eric A. Young] 10412 10413 *) Changed some BIGNUM api stuff. 10414 [Eric A. Young] 10415 10416 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 10417 DGUX x86, Linux Alpha, etc. 10418 [Eric A. Young] 10419 10420 *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 10421 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 10422 available). 10423 [Eric A. Young] 10424 10425 *) Add -strparse option to asn1pars program which parses nested 10426 binary structures 10427 [Dr Stephen Henson <shenson@bigfoot.com>] 10428 10429 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 10430 [Eric A. Young] 10431 10432 *) DSA fix for "ca" program. 10433 [Eric A. Young] 10434 10435 *) Added "-genkey" option to "dsaparam" program. 10436 [Eric A. Young] 10437 10438 *) Added RIPE MD160 (rmd160) message digest. 10439 [Eric A. Young] 10440 10441 *) Added -a (all) option to "ssleay version" command. 10442 [Eric A. Young] 10443 10444 *) Added PLATFORM define which is the id given to Configure. 10445 [Eric A. Young] 10446 10447 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 10448 [Eric A. Young] 10449 10450 *) Extended the ASN.1 parser routines. 10451 [Eric A. Young] 10452 10453 *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 10454 [Eric A. Young] 10455 10456 *) Added a BN_CTX to the BN library. 10457 [Eric A. Young] 10458 10459 *) Fixed the weak key values in DES library 10460 [Eric A. Young] 10461 10462 *) Changed API in EVP library for cipher aliases. 10463 [Eric A. Young] 10464 10465 *) Added support for RC2/64bit cipher. 10466 [Eric A. Young] 10467 10468 *) Converted the lhash library to the crypto/mem.c functions. 10469 [Eric A. Young] 10470 10471 *) Added more recognized ASN.1 object ids. 10472 [Eric A. Young] 10473 10474 *) Added more RSA padding checks for SSL/TLS. 10475 [Eric A. Young] 10476 10477 *) Added BIO proxy/filter functionality. 10478 [Eric A. Young] 10479 10480 *) Added extra_certs to SSL_CTX which can be used 10481 send extra CA certificates to the client in the CA cert chain sending 10482 process. It can be configured with SSL_CTX_add_extra_chain_cert(). 10483 [Eric A. Young] 10484 10485 *) Now Fortezza is denied in the authentication phase because 10486 this is key exchange mechanism is not supported by SSLeay at all. 10487 [Eric A. Young] 10488 10489 *) Additional PKCS1 checks. 10490 [Eric A. Young] 10491 10492 *) Support the string "TLSv1" for all TLS v1 ciphers. 10493 [Eric A. Young] 10494 10495 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 10496 ex_data index of the SSL context in the X509_STORE_CTX ex_data. 10497 [Eric A. Young] 10498 10499 *) Fixed a few memory leaks. 10500 [Eric A. Young] 10501 10502 *) Fixed various code and comment typos. 10503 [Eric A. Young] 10504 10505 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 10506 bytes sent in the client random. 10507 [Edward Bishop <ebishop@spyglass.com>] 10508 10509