CHANGES revision 306196
1 2 OpenSSL CHANGES 3 _______________ 4 5 Changes between 1.0.1t and 1.0.1u [22 Sep 2016] 6 7 *) OCSP Status Request extension unbounded memory growth 8 9 A malicious client can send an excessively large OCSP Status Request 10 extension. If that client continually requests renegotiation, sending a 11 large OCSP Status Request extension each time, then there will be unbounded 12 memory growth on the server. This will eventually lead to a Denial Of 13 Service attack through memory exhaustion. Servers with a default 14 configuration are vulnerable even if they do not support OCSP. Builds using 15 the "no-ocsp" build time option are not affected. 16 17 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 18 (CVE-2016-6304) 19 [Matt Caswell] 20 21 *) In order to mitigate the SWEET32 attack, the DES ciphers were moved from 22 HIGH to MEDIUM. 23 24 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan 25 Leurent (INRIA) 26 (CVE-2016-2183) 27 [Rich Salz] 28 29 *) OOB write in MDC2_Update() 30 31 An overflow can occur in MDC2_Update() either if called directly or 32 through the EVP_DigestUpdate() function using MDC2. If an attacker 33 is able to supply very large amounts of input data after a previous 34 call to EVP_EncryptUpdate() with a partial block then a length check 35 can overflow resulting in a heap corruption. 36 37 The amount of data needed is comparable to SIZE_MAX which is impractical 38 on most platforms. 39 40 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 41 (CVE-2016-6303) 42 [Stephen Henson] 43 44 *) Malformed SHA512 ticket DoS 45 46 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a 47 DoS attack where a malformed ticket will result in an OOB read which will 48 ultimately crash. 49 50 The use of SHA512 in TLS session tickets is comparatively rare as it requires 51 a custom server callback and ticket lookup mechanism. 52 53 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 54 (CVE-2016-6302) 55 [Stephen Henson] 56 57 *) OOB write in BN_bn2dec() 58 59 The function BN_bn2dec() does not check the return value of BN_div_word(). 60 This can cause an OOB write if an application uses this function with an 61 overly large BIGNUM. This could be a problem if an overly large certificate 62 or CRL is printed out from an untrusted source. TLS is not affected because 63 record limits will reject an oversized certificate before it is parsed. 64 65 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 66 (CVE-2016-2182) 67 [Stephen Henson] 68 69 *) OOB read in TS_OBJ_print_bio() 70 71 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is 72 the total length the OID text representation would use and not the amount 73 of data written. This will result in OOB reads when large OIDs are 74 presented. 75 76 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 77 (CVE-2016-2180) 78 [Stephen Henson] 79 80 *) Pointer arithmetic undefined behaviour 81 82 Avoid some undefined pointer arithmetic 83 84 A common idiom in the codebase is to check limits in the following manner: 85 "p + len > limit" 86 87 Where "p" points to some malloc'd data of SIZE bytes and 88 limit == p + SIZE 89 90 "len" here could be from some externally supplied data (e.g. from a TLS 91 message). 92 93 The rules of C pointer arithmetic are such that "p + len" is only well 94 defined where len <= SIZE. Therefore the above idiom is actually 95 undefined behaviour. 96 97 For example this could cause problems if some malloc implementation 98 provides an address for "p" such that "p + len" actually overflows for 99 values of len that are too big and therefore p + len < limit. 100 101 This issue was reported to OpenSSL by Guido Vranken 102 (CVE-2016-2177) 103 [Matt Caswell] 104 105 *) Constant time flag not preserved in DSA signing 106 107 Operations in the DSA signing algorithm should run in constant time in 108 order to avoid side channel attacks. A flaw in the OpenSSL DSA 109 implementation means that a non-constant time codepath is followed for 110 certain operations. This has been demonstrated through a cache-timing 111 attack to be sufficient for an attacker to recover the private DSA key. 112 113 This issue was reported by C��sar Pereida (Aalto University), Billy Brumley 114 (Tampere University of Technology), and Yuval Yarom (The University of 115 Adelaide and NICTA). 116 (CVE-2016-2178) 117 [C��sar Pereida] 118 119 *) DTLS buffered message DoS 120 121 In a DTLS connection where handshake messages are delivered out-of-order 122 those messages that OpenSSL is not yet ready to process will be buffered 123 for later use. Under certain circumstances, a flaw in the logic means that 124 those messages do not get removed from the buffer even though the handshake 125 has been completed. An attacker could force up to approx. 15 messages to 126 remain in the buffer when they are no longer required. These messages will 127 be cleared when the DTLS connection is closed. The default maximum size for 128 a message is 100k. Therefore the attacker could force an additional 1500k 129 to be consumed per connection. By opening many simulataneous connections an 130 attacker could cause a DoS attack through memory exhaustion. 131 132 This issue was reported to OpenSSL by Quan Luo. 133 (CVE-2016-2179) 134 [Matt Caswell] 135 136 *) DTLS replay protection DoS 137 138 A flaw in the DTLS replay attack protection mechanism means that records 139 that arrive for future epochs update the replay protection "window" before 140 the MAC for the record has been validated. This could be exploited by an 141 attacker by sending a record for the next epoch (which does not have to 142 decrypt or have a valid MAC), with a very large sequence number. This means 143 that all subsequent legitimate packets are dropped causing a denial of 144 service for a specific DTLS connection. 145 146 This issue was reported to OpenSSL by the OCAP audit team. 147 (CVE-2016-2181) 148 [Matt Caswell] 149 150 *) Certificate message OOB reads 151 152 In OpenSSL 1.0.2 and earlier some missing message length checks can result 153 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a 154 theoretical DoS risk but this has not been observed in practice on common 155 platforms. 156 157 The messages affected are client certificate, client certificate request 158 and server certificate. As a result the attack can only be performed 159 against a client or a server which enables client authentication. 160 161 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 162 (CVE-2016-6306) 163 [Stephen Henson] 164 165 Changes between 1.0.1s and 1.0.1t [3 May 2016] 166 167 *) Prevent padding oracle in AES-NI CBC MAC check 168 169 A MITM attacker can use a padding oracle attack to decrypt traffic 170 when the connection uses an AES CBC cipher and the server support 171 AES-NI. 172 173 This issue was introduced as part of the fix for Lucky 13 padding 174 attack (CVE-2013-0169). The padding check was rewritten to be in 175 constant time by making sure that always the same bytes are read and 176 compared against either the MAC or padding bytes. But it no longer 177 checked that there was enough data to have both the MAC and padding 178 bytes. 179 180 This issue was reported by Juraj Somorovsky using TLS-Attacker. 181 (CVE-2016-2107) 182 [Kurt Roeckx] 183 184 *) Fix EVP_EncodeUpdate overflow 185 186 An overflow can occur in the EVP_EncodeUpdate() function which is used for 187 Base64 encoding of binary data. If an attacker is able to supply very large 188 amounts of input data then a length check can overflow resulting in a heap 189 corruption. 190 191 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by 192 the PEM_write_bio* family of functions. These are mainly used within the 193 OpenSSL command line applications, so any application which processes data 194 from an untrusted source and outputs it as a PEM file should be considered 195 vulnerable to this issue. User applications that call these APIs directly 196 with large amounts of untrusted data may also be vulnerable. 197 198 This issue was reported by Guido Vranken. 199 (CVE-2016-2105) 200 [Matt Caswell] 201 202 *) Fix EVP_EncryptUpdate overflow 203 204 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker 205 is able to supply very large amounts of input data after a previous call to 206 EVP_EncryptUpdate() with a partial block then a length check can overflow 207 resulting in a heap corruption. Following an analysis of all OpenSSL 208 internal usage of the EVP_EncryptUpdate() function all usage is one of two 209 forms. The first form is where the EVP_EncryptUpdate() call is known to be 210 the first called function after an EVP_EncryptInit(), and therefore that 211 specific call must be safe. The second form is where the length passed to 212 EVP_EncryptUpdate() can be seen from the code to be some small value and 213 therefore there is no possibility of an overflow. Since all instances are 214 one of these two forms, it is believed that there can be no overflows in 215 internal code due to this problem. It should be noted that 216 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths. 217 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances 218 of these calls have also been analysed too and it is believed there are no 219 instances in internal usage where an overflow could occur. 220 221 This issue was reported by Guido Vranken. 222 (CVE-2016-2106) 223 [Matt Caswell] 224 225 *) Prevent ASN.1 BIO excessive memory allocation 226 227 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio() 228 a short invalid encoding can casuse allocation of large amounts of memory 229 potentially consuming excessive resources or exhausting memory. 230 231 Any application parsing untrusted data through d2i BIO functions is 232 affected. The memory based functions such as d2i_X509() are *not* affected. 233 Since the memory based functions are used by the TLS library, TLS 234 applications are not affected. 235 236 This issue was reported by Brian Carpenter. 237 (CVE-2016-2109) 238 [Stephen Henson] 239 240 *) EBCDIC overread 241 242 ASN1 Strings that are over 1024 bytes can cause an overread in applications 243 using the X509_NAME_oneline() function on EBCDIC systems. This could result 244 in arbitrary stack data being returned in the buffer. 245 246 This issue was reported by Guido Vranken. 247 (CVE-2016-2176) 248 [Matt Caswell] 249 250 *) Modify behavior of ALPN to invoke callback after SNI/servername 251 callback, such that updates to the SSL_CTX affect ALPN. 252 [Todd Short] 253 254 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the 255 default. 256 [Kurt Roeckx] 257 258 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the 259 methods are enabled and ssl2 is disabled the methods return NULL. 260 [Kurt Roeckx] 261 262 Changes between 1.0.1r and 1.0.1s [1 Mar 2016] 263 264 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL. 265 Builds that are not configured with "enable-weak-ssl-ciphers" will not 266 provide any "EXPORT" or "LOW" strength ciphers. 267 [Viktor Dukhovni] 268 269 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2 270 is by default disabled at build-time. Builds that are not configured with 271 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used, 272 users who want to negotiate SSLv2 via the version-flexible SSLv23_method() 273 will need to explicitly call either of: 274 275 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2); 276 or 277 SSL_clear_options(ssl, SSL_OP_NO_SSLv2); 278 279 as appropriate. Even if either of those is used, or the application 280 explicitly uses the version-specific SSLv2_method() or its client and 281 server variants, SSLv2 ciphers vulnerable to exhaustive search key 282 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT 283 ciphers, and SSLv2 56-bit DES are no longer available. 284 (CVE-2016-0800) 285 [Viktor Dukhovni] 286 287 *) Fix a double-free in DSA code 288 289 A double free bug was discovered when OpenSSL parses malformed DSA private 290 keys and could lead to a DoS attack or memory corruption for applications 291 that receive DSA private keys from untrusted sources. This scenario is 292 considered rare. 293 294 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using 295 libFuzzer. 296 (CVE-2016-0705) 297 [Stephen Henson] 298 299 *) Disable SRP fake user seed to address a server memory leak. 300 301 Add a new method SRP_VBASE_get1_by_user that handles the seed properly. 302 303 SRP_VBASE_get_by_user had inconsistent memory management behaviour. 304 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user 305 was changed to ignore the "fake user" SRP seed, even if the seed 306 is configured. 307 308 Users should use SRP_VBASE_get1_by_user instead. Note that in 309 SRP_VBASE_get1_by_user, caller must free the returned value. Note 310 also that even though configuring the SRP seed attempts to hide 311 invalid usernames by continuing the handshake with fake 312 credentials, this behaviour is not constant time and no strong 313 guarantees are made that the handshake is indistinguishable from 314 that of a valid user. 315 (CVE-2016-0798) 316 [Emilia K��sper] 317 318 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption 319 320 In the BN_hex2bn function the number of hex digits is calculated using an 321 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For 322 large values of |i| this can result in |bn_expand| not allocating any 323 memory because |i * 4| is negative. This can leave the internal BIGNUM data 324 field as NULL leading to a subsequent NULL ptr deref. For very large values 325 of |i|, the calculation |i * 4| could be a positive value smaller than |i|. 326 In this case memory is allocated to the internal BIGNUM data field, but it 327 is insufficiently sized leading to heap corruption. A similar issue exists 328 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn 329 is ever called by user applications with very large untrusted hex/dec data. 330 This is anticipated to be a rare occurrence. 331 332 All OpenSSL internal usage of these functions use data that is not expected 333 to be untrusted, e.g. config file data or application command line 334 arguments. If user developed applications generate config file data based 335 on untrusted data then it is possible that this could also lead to security 336 consequences. This is also anticipated to be rare. 337 338 This issue was reported to OpenSSL by Guido Vranken. 339 (CVE-2016-0797) 340 [Matt Caswell] 341 342 *) Fix memory issues in BIO_*printf functions 343 344 The internal |fmtstr| function used in processing a "%s" format string in 345 the BIO_*printf functions could overflow while calculating the length of a 346 string and cause an OOB read when printing very long strings. 347 348 Additionally the internal |doapr_outch| function can attempt to write to an 349 OOB memory location (at an offset from the NULL pointer) in the event of a 350 memory allocation failure. In 1.0.2 and below this could be caused where 351 the size of a buffer to be allocated is greater than INT_MAX. E.g. this 352 could be in processing a very long "%s" format string. Memory leaks can 353 also occur. 354 355 The first issue may mask the second issue dependent on compiler behaviour. 356 These problems could enable attacks where large amounts of untrusted data 357 is passed to the BIO_*printf functions. If applications use these functions 358 in this way then they could be vulnerable. OpenSSL itself uses these 359 functions when printing out human-readable dumps of ASN.1 data. Therefore 360 applications that print this data could be vulnerable if the data is from 361 untrusted sources. OpenSSL command line applications could also be 362 vulnerable where they print out ASN.1 data, or if untrusted data is passed 363 as command line arguments. 364 365 Libssl is not considered directly vulnerable. Additionally certificates etc 366 received via remote connections via libssl are also unlikely to be able to 367 trigger these issues because of message size limits enforced within libssl. 368 369 This issue was reported to OpenSSL Guido Vranken. 370 (CVE-2016-0799) 371 [Matt Caswell] 372 373 *) Side channel attack on modular exponentiation 374 375 A side-channel attack was found which makes use of cache-bank conflicts on 376 the Intel Sandy-Bridge microarchitecture which could lead to the recovery 377 of RSA keys. The ability to exploit this issue is limited as it relies on 378 an attacker who has control of code in a thread running on the same 379 hyper-threaded core as the victim thread which is performing decryptions. 380 381 This issue was reported to OpenSSL by Yuval Yarom, The University of 382 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and 383 Nadia Heninger, University of Pennsylvania with more information at 384 http://cachebleed.info. 385 (CVE-2016-0702) 386 [Andy Polyakov] 387 388 *) Change the req app to generate a 2048-bit RSA/DSA key by default, 389 if no keysize is specified with default_bits. This fixes an 390 omission in an earlier change that changed all RSA/DSA key generation 391 apps to use 2048 bits by default. 392 [Emilia K��sper] 393 394 Changes between 1.0.1q and 1.0.1r [28 Jan 2016] 395 396 *) Protection for DH small subgroup attacks 397 398 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been 399 switched on by default and cannot be disabled. This could have some 400 performance impact. 401 [Matt Caswell] 402 403 *) SSLv2 doesn't block disabled ciphers 404 405 A malicious client can negotiate SSLv2 ciphers that have been disabled on 406 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have 407 been disabled, provided that the SSLv2 protocol was not also disabled via 408 SSL_OP_NO_SSLv2. 409 410 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram 411 and Sebastian Schinzel. 412 (CVE-2015-3197) 413 [Viktor Dukhovni] 414 415 *) Reject DH handshakes with parameters shorter than 1024 bits. 416 [Kurt Roeckx] 417 418 Changes between 1.0.1p and 1.0.1q [3 Dec 2015] 419 420 *) Certificate verify crash with missing PSS parameter 421 422 The signature verification routines will crash with a NULL pointer 423 dereference if presented with an ASN.1 signature using the RSA PSS 424 algorithm and absent mask generation function parameter. Since these 425 routines are used to verify certificate signature algorithms this can be 426 used to crash any certificate verification operation and exploited in a 427 DoS attack. Any application which performs certificate verification is 428 vulnerable including OpenSSL clients and servers which enable client 429 authentication. 430 431 This issue was reported to OpenSSL by Lo��c Jonas Etienne (Qnective AG). 432 (CVE-2015-3194) 433 [Stephen Henson] 434 435 *) X509_ATTRIBUTE memory leak 436 437 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak 438 memory. This structure is used by the PKCS#7 and CMS routines so any 439 application which reads PKCS#7 or CMS data from untrusted sources is 440 affected. SSL/TLS is not affected. 441 442 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using 443 libFuzzer. 444 (CVE-2015-3195) 445 [Stephen Henson] 446 447 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs. 448 This changes the decoding behaviour for some invalid messages, 449 though the change is mostly in the more lenient direction, and 450 legacy behaviour is preserved as much as possible. 451 [Emilia K��sper] 452 453 *) In DSA_generate_parameters_ex, if the provided seed is too short, 454 use a random seed, as already documented. 455 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>] 456 457 Changes between 1.0.1o and 1.0.1p [9 Jul 2015] 458 459 *) Alternate chains certificate forgery 460 461 During certificate verfification, OpenSSL will attempt to find an 462 alternative certificate chain if the first attempt to build such a chain 463 fails. An error in the implementation of this logic can mean that an 464 attacker could cause certain checks on untrusted certificates to be 465 bypassed, such as the CA flag, enabling them to use a valid leaf 466 certificate to act as a CA and "issue" an invalid certificate. 467 468 This issue was reported to OpenSSL by Adam Langley/David Benjamin 469 (Google/BoringSSL). 470 (CVE-2015-1793) 471 [Matt Caswell] 472 473 *) Race condition handling PSK identify hint 474 475 If PSK identity hints are received by a multi-threaded client then 476 the values are wrongly updated in the parent SSL_CTX structure. This can 477 result in a race condition potentially leading to a double free of the 478 identify hint data. 479 (CVE-2015-3196) 480 [Stephen Henson] 481 482 Changes between 1.0.1n and 1.0.1o [12 Jun 2015] 483 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI 484 incompatibility in the handling of HMAC. The previous ABI has now been 485 restored. 486 487 Changes between 1.0.1m and 1.0.1n [11 Jun 2015] 488 489 *) Malformed ECParameters causes infinite loop 490 491 When processing an ECParameters structure OpenSSL enters an infinite loop 492 if the curve specified is over a specially malformed binary polynomial 493 field. 494 495 This can be used to perform denial of service against any 496 system which processes public keys, certificate requests or 497 certificates. This includes TLS clients and TLS servers with 498 client authentication enabled. 499 500 This issue was reported to OpenSSL by Joseph Barr-Pixton. 501 (CVE-2015-1788) 502 [Andy Polyakov] 503 504 *) Exploitable out-of-bounds read in X509_cmp_time 505 506 X509_cmp_time does not properly check the length of the ASN1_TIME 507 string and can read a few bytes out of bounds. In addition, 508 X509_cmp_time accepts an arbitrary number of fractional seconds in the 509 time string. 510 511 An attacker can use this to craft malformed certificates and CRLs of 512 various sizes and potentially cause a segmentation fault, resulting in 513 a DoS on applications that verify certificates or CRLs. TLS clients 514 that verify CRLs are affected. TLS clients and servers with client 515 authentication enabled may be affected if they use custom verification 516 callbacks. 517 518 This issue was reported to OpenSSL by Robert Swiecki (Google), and 519 independently by Hanno B��ck. 520 (CVE-2015-1789) 521 [Emilia K��sper] 522 523 *) PKCS7 crash with missing EnvelopedContent 524 525 The PKCS#7 parsing code does not handle missing inner EncryptedContent 526 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs 527 with missing content and trigger a NULL pointer dereference on parsing. 528 529 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7 530 structures from untrusted sources are affected. OpenSSL clients and 531 servers are not affected. 532 533 This issue was reported to OpenSSL by Michal Zalewski (Google). 534 (CVE-2015-1790) 535 [Emilia K��sper] 536 537 *) CMS verify infinite loop with unknown hash function 538 539 When verifying a signedData message the CMS code can enter an infinite loop 540 if presented with an unknown hash function OID. This can be used to perform 541 denial of service against any system which verifies signedData messages using 542 the CMS code. 543 This issue was reported to OpenSSL by Johannes Bauer. 544 (CVE-2015-1792) 545 [Stephen Henson] 546 547 *) Race condition handling NewSessionTicket 548 549 If a NewSessionTicket is received by a multi-threaded client when attempting to 550 reuse a previous ticket then a race condition can occur potentially leading to 551 a double free of the ticket data. 552 (CVE-2015-1791) 553 [Matt Caswell] 554 555 *) Reject DH handshakes with parameters shorter than 768 bits. 556 [Kurt Roeckx and Emilia Kasper] 557 558 *) dhparam: generate 2048-bit parameters by default. 559 [Kurt Roeckx and Emilia Kasper] 560 561 Changes between 1.0.1l and 1.0.1m [19 Mar 2015] 562 563 *) Segmentation fault in ASN1_TYPE_cmp fix 564 565 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is 566 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check 567 certificate signature algorithm consistency this can be used to crash any 568 certificate verification operation and exploited in a DoS attack. Any 569 application which performs certificate verification is vulnerable including 570 OpenSSL clients and servers which enable client authentication. 571 (CVE-2015-0286) 572 [Stephen Henson] 573 574 *) ASN.1 structure reuse memory corruption fix 575 576 Reusing a structure in ASN.1 parsing may allow an attacker to cause 577 memory corruption via an invalid write. Such reuse is and has been 578 strongly discouraged and is believed to be rare. 579 580 Applications that parse structures containing CHOICE or ANY DEFINED BY 581 components may be affected. Certificate parsing (d2i_X509 and related 582 functions) are however not affected. OpenSSL clients and servers are 583 not affected. 584 (CVE-2015-0287) 585 [Stephen Henson] 586 587 *) PKCS7 NULL pointer dereferences fix 588 589 The PKCS#7 parsing code does not handle missing outer ContentInfo 590 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with 591 missing content and trigger a NULL pointer dereference on parsing. 592 593 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or 594 otherwise parse PKCS#7 structures from untrusted sources are 595 affected. OpenSSL clients and servers are not affected. 596 597 This issue was reported to OpenSSL by Michal Zalewski (Google). 598 (CVE-2015-0289) 599 [Emilia K��sper] 600 601 *) DoS via reachable assert in SSLv2 servers fix 602 603 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in 604 servers that both support SSLv2 and enable export cipher suites by sending 605 a specially crafted SSLv2 CLIENT-MASTER-KEY message. 606 607 This issue was discovered by Sean Burford (Google) and Emilia K��sper 608 (OpenSSL development team). 609 (CVE-2015-0293) 610 [Emilia K��sper] 611 612 *) Use After Free following d2i_ECPrivatekey error fix 613 614 A malformed EC private key file consumed via the d2i_ECPrivateKey function 615 could cause a use after free condition. This, in turn, could cause a double 616 free in several private key parsing functions (such as d2i_PrivateKey 617 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption 618 for applications that receive EC private keys from untrusted 619 sources. This scenario is considered rare. 620 621 This issue was discovered by the BoringSSL project and fixed in their 622 commit 517073cd4b. 623 (CVE-2015-0209) 624 [Matt Caswell] 625 626 *) X509_to_X509_REQ NULL pointer deref fix 627 628 The function X509_to_X509_REQ will crash with a NULL pointer dereference if 629 the certificate key is invalid. This function is rarely used in practice. 630 631 This issue was discovered by Brian Carpenter. 632 (CVE-2015-0288) 633 [Stephen Henson] 634 635 *) Removed the export ciphers from the DEFAULT ciphers 636 [Kurt Roeckx] 637 638 Changes between 1.0.1k and 1.0.1l [15 Jan 2015] 639 640 *) Build fixes for the Windows and OpenVMS platforms 641 [Matt Caswell and Richard Levitte] 642 643 Changes between 1.0.1j and 1.0.1k [8 Jan 2015] 644 645 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS 646 message can cause a segmentation fault in OpenSSL due to a NULL pointer 647 dereference. This could lead to a Denial Of Service attack. Thanks to 648 Markus Stenberg of Cisco Systems, Inc. for reporting this issue. 649 (CVE-2014-3571) 650 [Steve Henson] 651 652 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the 653 dtls1_buffer_record function under certain conditions. In particular this 654 could occur if an attacker sent repeated DTLS records with the same 655 sequence number but for the next epoch. The memory leak could be exploited 656 by an attacker in a Denial of Service attack through memory exhaustion. 657 Thanks to Chris Mueller for reporting this issue. 658 (CVE-2015-0206) 659 [Matt Caswell] 660 661 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is 662 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl 663 method would be set to NULL which could later result in a NULL pointer 664 dereference. Thanks to Frank Schmirler for reporting this issue. 665 (CVE-2014-3569) 666 [Kurt Roeckx] 667 668 *) Abort handshake if server key exchange message is omitted for ephemeral 669 ECDH ciphersuites. 670 671 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for 672 reporting this issue. 673 (CVE-2014-3572) 674 [Steve Henson] 675 676 *) Remove non-export ephemeral RSA code on client and server. This code 677 violated the TLS standard by allowing the use of temporary RSA keys in 678 non-export ciphersuites and could be used by a server to effectively 679 downgrade the RSA key length used to a value smaller than the server 680 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at 681 INRIA or reporting this issue. 682 (CVE-2015-0204) 683 [Steve Henson] 684 685 *) Fixed issue where DH client certificates are accepted without verification. 686 An OpenSSL server will accept a DH certificate for client authentication 687 without the certificate verify message. This effectively allows a client to 688 authenticate without the use of a private key. This only affects servers 689 which trust a client certificate authority which issues certificates 690 containing DH keys: these are extremely rare and hardly ever encountered. 691 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting 692 this issue. 693 (CVE-2015-0205) 694 [Steve Henson] 695 696 *) Ensure that the session ID context of an SSL is updated when its 697 SSL_CTX is updated via SSL_set_SSL_CTX. 698 699 The session ID context is typically set from the parent SSL_CTX, 700 and can vary with the CTX. 701 [Adam Langley] 702 703 *) Fix various certificate fingerprint issues. 704 705 By using non-DER or invalid encodings outside the signed portion of a 706 certificate the fingerprint can be changed without breaking the signature. 707 Although no details of the signed portion of the certificate can be changed 708 this can cause problems with some applications: e.g. those using the 709 certificate fingerprint for blacklists. 710 711 1. Reject signatures with non zero unused bits. 712 713 If the BIT STRING containing the signature has non zero unused bits reject 714 the signature. All current signature algorithms require zero unused bits. 715 716 2. Check certificate algorithm consistency. 717 718 Check the AlgorithmIdentifier inside TBS matches the one in the 719 certificate signature. NB: this will result in signature failure 720 errors for some broken certificates. 721 722 Thanks to Konrad Kraszewski from Google for reporting this issue. 723 724 3. Check DSA/ECDSA signatures use DER. 725 726 Reencode DSA/ECDSA signatures and compare with the original received 727 signature. Return an error if there is a mismatch. 728 729 This will reject various cases including garbage after signature 730 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS 731 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs 732 (negative or with leading zeroes). 733 734 Further analysis was conducted and fixes were developed by Stephen Henson 735 of the OpenSSL core team. 736 737 (CVE-2014-8275) 738 [Steve Henson] 739 740 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect 741 results on some platforms, including x86_64. This bug occurs at random 742 with a very low probability, and is not known to be exploitable in any 743 way, though its exact impact is difficult to determine. Thanks to Pieter 744 Wuille (Blockstream) who reported this issue and also suggested an initial 745 fix. Further analysis was conducted by the OpenSSL development team and 746 Adam Langley of Google. The final fix was developed by Andy Polyakov of 747 the OpenSSL core team. 748 (CVE-2014-3570) 749 [Andy Polyakov] 750 751 *) Do not resume sessions on the server if the negotiated protocol 752 version does not match the session's version. Resuming with a different 753 version, while not strictly forbidden by the RFC, is of questionable 754 sanity and breaks all known clients. 755 [David Benjamin, Emilia K��sper] 756 757 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject 758 early CCS messages during renegotiation. (Note that because 759 renegotiation is encrypted, this early CCS was not exploitable.) 760 [Emilia K��sper] 761 762 *) Tighten client-side session ticket handling during renegotiation: 763 ensure that the client only accepts a session ticket if the server sends 764 the extension anew in the ServerHello. Previously, a TLS client would 765 reuse the old extension state and thus accept a session ticket if one was 766 announced in the initial ServerHello. 767 768 Similarly, ensure that the client requires a session ticket if one 769 was advertised in the ServerHello. Previously, a TLS client would 770 ignore a missing NewSessionTicket message. 771 [Emilia K��sper] 772 773 Changes between 1.0.1i and 1.0.1j [15 Oct 2014] 774 775 *) SRTP Memory Leak. 776 777 A flaw in the DTLS SRTP extension parsing code allows an attacker, who 778 sends a carefully crafted handshake message, to cause OpenSSL to fail 779 to free up to 64k of memory causing a memory leak. This could be 780 exploited in a Denial Of Service attack. This issue affects OpenSSL 781 1.0.1 server implementations for both SSL/TLS and DTLS regardless of 782 whether SRTP is used or configured. Implementations of OpenSSL that 783 have been compiled with OPENSSL_NO_SRTP defined are not affected. 784 785 The fix was developed by the OpenSSL team. 786 (CVE-2014-3513) 787 [OpenSSL team] 788 789 *) Session Ticket Memory Leak. 790 791 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the 792 integrity of that ticket is first verified. In the event of a session 793 ticket integrity check failing, OpenSSL will fail to free memory 794 causing a memory leak. By sending a large number of invalid session 795 tickets an attacker could exploit this issue in a Denial Of Service 796 attack. 797 (CVE-2014-3567) 798 [Steve Henson] 799 800 *) Build option no-ssl3 is incomplete. 801 802 When OpenSSL is configured with "no-ssl3" as a build option, servers 803 could accept and complete a SSL 3.0 handshake, and clients could be 804 configured to send them. 805 (CVE-2014-3568) 806 [Akamai and the OpenSSL team] 807 808 *) Add support for TLS_FALLBACK_SCSV. 809 Client applications doing fallback retries should call 810 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV). 811 (CVE-2014-3566) 812 [Adam Langley, Bodo Moeller] 813 814 *) Add additional DigestInfo checks. 815 816 Reencode DigestInto in DER and check against the original when 817 verifying RSA signature: this will reject any improperly encoded 818 DigestInfo structures. 819 820 Note: this is a precautionary measure and no attacks are currently known. 821 822 [Steve Henson] 823 824 Changes between 1.0.1h and 1.0.1i [6 Aug 2014] 825 826 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the 827 SRP code can be overrun an internal buffer. Add sanity check that 828 g, A, B < N to SRP code. 829 830 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC 831 Group for discovering this issue. 832 (CVE-2014-3512) 833 [Steve Henson] 834 835 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate 836 TLS 1.0 instead of higher protocol versions when the ClientHello message 837 is badly fragmented. This allows a man-in-the-middle attacker to force a 838 downgrade to TLS 1.0 even if both the server and the client support a 839 higher protocol version, by modifying the client's TLS records. 840 841 Thanks to David Benjamin and Adam Langley (Google) for discovering and 842 researching this issue. 843 (CVE-2014-3511) 844 [David Benjamin] 845 846 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject 847 to a denial of service attack. A malicious server can crash the client 848 with a null pointer dereference (read) by specifying an anonymous (EC)DH 849 ciphersuite and sending carefully crafted handshake messages. 850 851 Thanks to Felix Gr��bert (Google) for discovering and researching this 852 issue. 853 (CVE-2014-3510) 854 [Emilia K��sper] 855 856 *) By sending carefully crafted DTLS packets an attacker could cause openssl 857 to leak memory. This can be exploited through a Denial of Service attack. 858 Thanks to Adam Langley for discovering and researching this issue. 859 (CVE-2014-3507) 860 [Adam Langley] 861 862 *) An attacker can force openssl to consume large amounts of memory whilst 863 processing DTLS handshake messages. This can be exploited through a 864 Denial of Service attack. 865 Thanks to Adam Langley for discovering and researching this issue. 866 (CVE-2014-3506) 867 [Adam Langley] 868 869 *) An attacker can force an error condition which causes openssl to crash 870 whilst processing DTLS packets due to memory being freed twice. This 871 can be exploited through a Denial of Service attack. 872 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching 873 this issue. 874 (CVE-2014-3505) 875 [Adam Langley] 876 877 *) If a multithreaded client connects to a malicious server using a resumed 878 session and the server sends an ec point format extension it could write 879 up to 255 bytes to freed memory. 880 881 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this 882 issue. 883 (CVE-2014-3509) 884 [Gabor Tyukasz] 885 886 *) A malicious server can crash an OpenSSL client with a null pointer 887 dereference (read) by specifying an SRP ciphersuite even though it was not 888 properly negotiated with the client. This can be exploited through a 889 Denial of Service attack. 890 891 Thanks to Joonas Kuorilehto and Riku Hietam��ki (Codenomicon) for 892 discovering and researching this issue. 893 (CVE-2014-5139) 894 [Steve Henson] 895 896 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as 897 X509_name_oneline, X509_name_print_ex et al. to leak some information 898 from the stack. Applications may be affected if they echo pretty printing 899 output to the attacker. 900 901 Thanks to Ivan Fratric (Google) for discovering this issue. 902 (CVE-2014-3508) 903 [Emilia K��sper, and Steve Henson] 904 905 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) 906 for corner cases. (Certain input points at infinity could lead to 907 bogus results, with non-infinity inputs mapped to infinity too.) 908 [Bodo Moeller] 909 910 Changes between 1.0.1g and 1.0.1h [5 Jun 2014] 911 912 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted 913 handshake can force the use of weak keying material in OpenSSL 914 SSL/TLS clients and servers. 915 916 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and 917 researching this issue. (CVE-2014-0224) 918 [KIKUCHI Masashi, Steve Henson] 919 920 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an 921 OpenSSL DTLS client the code can be made to recurse eventually crashing 922 in a DoS attack. 923 924 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue. 925 (CVE-2014-0221) 926 [Imre Rad, Steve Henson] 927 928 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can 929 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS 930 client or server. This is potentially exploitable to run arbitrary 931 code on a vulnerable client or server. 932 933 Thanks to J��ri Aedla for reporting this issue. (CVE-2014-0195) 934 [J��ri Aedla, Steve Henson] 935 936 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites 937 are subject to a denial of service attack. 938 939 Thanks to Felix Gr��bert and Ivan Fratric at Google for discovering 940 this issue. (CVE-2014-3470) 941 [Felix Gr��bert, Ivan Fratric, Steve Henson] 942 943 *) Harmonize version and its documentation. -f flag is used to display 944 compilation flags. 945 [mancha <mancha1@zoho.com>] 946 947 *) Fix eckey_priv_encode so it immediately returns an error upon a failure 948 in i2d_ECPrivateKey. 949 [mancha <mancha1@zoho.com>] 950 951 *) Fix some double frees. These are not thought to be exploitable. 952 [mancha <mancha1@zoho.com>] 953 954 Changes between 1.0.1f and 1.0.1g [7 Apr 2014] 955 956 *) A missing bounds check in the handling of the TLS heartbeat extension 957 can be used to reveal up to 64k of memory to a connected client or 958 server. 959 960 Thanks for Neel Mehta of Google Security for discovering this bug and to 961 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for 962 preparing the fix (CVE-2014-0160) 963 [Adam Langley, Bodo Moeller] 964 965 *) Fix for the attack described in the paper "Recovering OpenSSL 966 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack" 967 by Yuval Yarom and Naomi Benger. Details can be obtained from: 968 http://eprint.iacr.org/2014/140 969 970 Thanks to Yuval Yarom and Naomi Benger for discovering this 971 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076) 972 [Yuval Yarom and Naomi Benger] 973 974 *) TLS pad extension: draft-agl-tls-padding-03 975 976 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the 977 TLS client Hello record length value would otherwise be > 255 and 978 less that 512 pad with a dummy extension containing zeroes so it 979 is at least 512 bytes long. 980 981 [Adam Langley, Steve Henson] 982 983 Changes between 1.0.1e and 1.0.1f [6 Jan 2014] 984 985 *) Fix for TLS record tampering bug. A carefully crafted invalid 986 handshake could crash OpenSSL with a NULL pointer exception. 987 Thanks to Anton Johansson for reporting this issues. 988 (CVE-2013-4353) 989 990 *) Keep original DTLS digest and encryption contexts in retransmission 991 structures so we can use the previous session parameters if they need 992 to be resent. (CVE-2013-6450) 993 [Steve Henson] 994 995 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which 996 avoids preferring ECDHE-ECDSA ciphers when the client appears to be 997 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for 998 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug 999 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing 1000 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer. 1001 [Rob Stradling, Adam Langley] 1002 1003 Changes between 1.0.1d and 1.0.1e [11 Feb 2013] 1004 1005 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI 1006 supporting platforms or when small records were transferred. 1007 [Andy Polyakov, Steve Henson] 1008 1009 Changes between 1.0.1c and 1.0.1d [5 Feb 2013] 1010 1011 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 1012 1013 This addresses the flaw in CBC record processing discovered by 1014 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 1015 at: http://www.isg.rhul.ac.uk/tls/ 1016 1017 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 1018 Security Group at Royal Holloway, University of London 1019 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 1020 Emilia K��sper for the initial patch. 1021 (CVE-2013-0169) 1022 [Emilia K��sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 1023 1024 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode 1025 ciphersuites which can be exploited in a denial of service attack. 1026 Thanks go to and to Adam Langley <agl@chromium.org> for discovering 1027 and detecting this bug and to Wolfgang Ettlinger 1028 <wolfgang.ettlinger@gmail.com> for independently discovering this issue. 1029 (CVE-2012-2686) 1030 [Adam Langley] 1031 1032 *) Return an error when checking OCSP signatures when key is NULL. 1033 This fixes a DoS attack. (CVE-2013-0166) 1034 [Steve Henson] 1035 1036 *) Make openssl verify return errors. 1037 [Chris Palmer <palmer@google.com> and Ben Laurie] 1038 1039 *) Call OCSP Stapling callback after ciphersuite has been chosen, so 1040 the right response is stapled. Also change SSL_get_certificate() 1041 so it returns the certificate actually sent. 1042 See http://rt.openssl.org/Ticket/Display.html?id=2836. 1043 [Rob Stradling <rob.stradling@comodo.com>] 1044 1045 *) Fix possible deadlock when decoding public keys. 1046 [Steve Henson] 1047 1048 *) Don't use TLS 1.0 record version number in initial client hello 1049 if renegotiating. 1050 [Steve Henson] 1051 1052 Changes between 1.0.1b and 1.0.1c [10 May 2012] 1053 1054 *) Sanity check record length before skipping explicit IV in TLS 1055 1.2, 1.1 and DTLS to fix DoS attack. 1056 1057 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 1058 fuzzing as a service testing platform. 1059 (CVE-2012-2333) 1060 [Steve Henson] 1061 1062 *) Initialise tkeylen properly when encrypting CMS messages. 1063 Thanks to Solar Designer of Openwall for reporting this issue. 1064 [Steve Henson] 1065 1066 *) In FIPS mode don't try to use composite ciphers as they are not 1067 approved. 1068 [Steve Henson] 1069 1070 Changes between 1.0.1a and 1.0.1b [26 Apr 2012] 1071 1072 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and 1073 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately 1074 mean any application compiled against OpenSSL 1.0.0 headers setting 1075 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng 1076 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to 1077 0x10000000L Any application which was previously compiled against 1078 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1 1079 will need to be recompiled as a result. Letting be results in 1080 inability to disable specifically TLS 1.1 and in client context, 1081 in unlike event, limit maximum offered version to TLS 1.0 [see below]. 1082 [Steve Henson] 1083 1084 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not 1085 disable just protocol X, but all protocols above X *if* there are 1086 protocols *below* X still enabled. In more practical terms it means 1087 that if application wants to disable TLS1.0 in favor of TLS1.1 and 1088 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass 1089 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to 1090 client side. 1091 [Andy Polyakov] 1092 1093 Changes between 1.0.1 and 1.0.1a [19 Apr 2012] 1094 1095 *) Check for potentially exploitable overflows in asn1_d2i_read_bio 1096 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 1097 in CRYPTO_realloc_clean. 1098 1099 Thanks to Tavis Ormandy, Google Security Team, for discovering this 1100 issue and to Adam Langley <agl@chromium.org> for fixing it. 1101 (CVE-2012-2110) 1102 [Adam Langley (Google), Tavis Ormandy, Google Security Team] 1103 1104 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections. 1105 [Adam Langley] 1106 1107 *) Workarounds for some broken servers that "hang" if a client hello 1108 record length exceeds 255 bytes. 1109 1110 1. Do not use record version number > TLS 1.0 in initial client 1111 hello: some (but not all) hanging servers will now work. 1112 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate 1113 the number of ciphers sent in the client hello. This should be 1114 set to an even number, such as 50, for example by passing: 1115 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure. 1116 Most broken servers should now work. 1117 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable 1118 TLS 1.2 client support entirely. 1119 [Steve Henson] 1120 1121 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH. 1122 [Andy Polyakov] 1123 1124 Changes between 1.0.0h and 1.0.1 [14 Mar 2012] 1125 1126 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET 1127 STRING form instead of a DigestInfo. 1128 [Steve Henson] 1129 1130 *) The format used for MDC2 RSA signatures is inconsistent between EVP 1131 and the RSA_sign/RSA_verify functions. This was made more apparent when 1132 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular 1133 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect 1134 the correct format in RSA_verify so both forms transparently work. 1135 [Steve Henson] 1136 1137 *) Some servers which support TLS 1.0 can choke if we initially indicate 1138 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA 1139 encrypted premaster secret. As a workaround use the maximum pemitted 1140 client version in client hello, this should keep such servers happy 1141 and still work with previous versions of OpenSSL. 1142 [Steve Henson] 1143 1144 *) Add support for TLS/DTLS heartbeats. 1145 [Robin Seggelmann <seggelmann@fh-muenster.de>] 1146 1147 *) Add support for SCTP. 1148 [Robin Seggelmann <seggelmann@fh-muenster.de>] 1149 1150 *) Improved PRNG seeding for VOS. 1151 [Paul Green <Paul.Green@stratus.com>] 1152 1153 *) Extensive assembler packs updates, most notably: 1154 1155 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support; 1156 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES); 1157 - x86_64: bit-sliced AES implementation; 1158 - ARM: NEON support, contemporary platforms optimizations; 1159 - s390x: z196 support; 1160 - *: GHASH and GF(2^m) multiplication implementations; 1161 1162 [Andy Polyakov] 1163 1164 *) Make TLS-SRP code conformant with RFC 5054 API cleanup 1165 (removal of unnecessary code) 1166 [Peter Sylvester <peter.sylvester@edelweb.fr>] 1167 1168 *) Add TLS key material exporter from RFC 5705. 1169 [Eric Rescorla] 1170 1171 *) Add DTLS-SRTP negotiation from RFC 5764. 1172 [Eric Rescorla] 1173 1174 *) Add Next Protocol Negotiation, 1175 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be 1176 disabled with a no-npn flag to config or Configure. Code donated 1177 by Google. 1178 [Adam Langley <agl@google.com> and Ben Laurie] 1179 1180 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224, 1181 NIST-P256, NIST-P521, with constant-time single point multiplication on 1182 typical inputs. Compiler support for the nonstandard type __uint128_t is 1183 required to use this (present in gcc 4.4 and later, for 64-bit builds). 1184 Code made available under Apache License version 2.0. 1185 1186 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command 1187 line to include this in your build of OpenSSL, and run "make depend" (or 1188 "make update"). This enables the following EC_METHODs: 1189 1190 EC_GFp_nistp224_method() 1191 EC_GFp_nistp256_method() 1192 EC_GFp_nistp521_method() 1193 1194 EC_GROUP_new_by_curve_name() will automatically use these (while 1195 EC_GROUP_new_curve_GFp() currently prefers the more flexible 1196 implementations). 1197 [Emilia K��sper, Adam Langley, Bodo Moeller (Google)] 1198 1199 *) Use type ossl_ssize_t instad of ssize_t which isn't available on 1200 all platforms. Move ssize_t definition from e_os.h to the public 1201 header file e_os2.h as it now appears in public header file cms.h 1202 [Steve Henson] 1203 1204 *) New -sigopt option to the ca, req and x509 utilities. Additional 1205 signature parameters can be passed using this option and in 1206 particular PSS. 1207 [Steve Henson] 1208 1209 *) Add RSA PSS signing function. This will generate and set the 1210 appropriate AlgorithmIdentifiers for PSS based on those in the 1211 corresponding EVP_MD_CTX structure. No application support yet. 1212 [Steve Henson] 1213 1214 *) Support for companion algorithm specific ASN1 signing routines. 1215 New function ASN1_item_sign_ctx() signs a pre-initialised 1216 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on 1217 the appropriate parameters. 1218 [Steve Henson] 1219 1220 *) Add new algorithm specific ASN1 verification initialisation function 1221 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1 1222 handling will be the same no matter what EVP_PKEY_METHOD is used. 1223 Add a PSS handler to support verification of PSS signatures: checked 1224 against a number of sample certificates. 1225 [Steve Henson] 1226 1227 *) Add signature printing for PSS. Add PSS OIDs. 1228 [Steve Henson, Martin Kaiser <lists@kaiser.cx>] 1229 1230 *) Add algorithm specific signature printing. An individual ASN1 method 1231 can now print out signatures instead of the standard hex dump. 1232 1233 More complex signatures (e.g. PSS) can print out more meaningful 1234 information. Include DSA version that prints out the signature 1235 parameters r, s. 1236 [Steve Henson] 1237 1238 *) Password based recipient info support for CMS library: implementing 1239 RFC3211. 1240 [Steve Henson] 1241 1242 *) Split password based encryption into PBES2 and PBKDF2 functions. This 1243 neatly separates the code into cipher and PBE sections and is required 1244 for some algorithms that split PBES2 into separate pieces (such as 1245 password based CMS). 1246 [Steve Henson] 1247 1248 *) Session-handling fixes: 1249 - Fix handling of connections that are resuming with a session ID, 1250 but also support Session Tickets. 1251 - Fix a bug that suppressed issuing of a new ticket if the client 1252 presented a ticket with an expired session. 1253 - Try to set the ticket lifetime hint to something reasonable. 1254 - Make tickets shorter by excluding irrelevant information. 1255 - On the client side, don't ignore renewed tickets. 1256 [Adam Langley, Bodo Moeller (Google)] 1257 1258 *) Fix PSK session representation. 1259 [Bodo Moeller] 1260 1261 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations. 1262 1263 This work was sponsored by Intel. 1264 [Andy Polyakov] 1265 1266 *) Add GCM support to TLS library. Some custom code is needed to split 1267 the IV between the fixed (from PRF) and explicit (from TLS record) 1268 portions. This adds all GCM ciphersuites supported by RFC5288 and 1269 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and 1270 add a special AESGCM string for GCM only. 1271 [Steve Henson] 1272 1273 *) Expand range of ctrls for AES GCM. Permit setting invocation 1274 field on decrypt and retrieval of invocation field only on encrypt. 1275 [Steve Henson] 1276 1277 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support. 1278 As required by RFC5289 these ciphersuites cannot be used if for 1279 versions of TLS earlier than 1.2. 1280 [Steve Henson] 1281 1282 *) For FIPS capable OpenSSL interpret a NULL default public key method 1283 as unset and return the appopriate default but do *not* set the default. 1284 This means we can return the appopriate method in applications that 1285 swicth between FIPS and non-FIPS modes. 1286 [Steve Henson] 1287 1288 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an 1289 ENGINE is used then we cannot handle that in the FIPS module so we 1290 keep original code iff non-FIPS operations are allowed. 1291 [Steve Henson] 1292 1293 *) Add -attime option to openssl utilities. 1294 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson] 1295 1296 *) Redirect DSA and DH operations to FIPS module in FIPS mode. 1297 [Steve Henson] 1298 1299 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use 1300 FIPS EC methods unconditionally for now. 1301 [Steve Henson] 1302 1303 *) New build option no-ec2m to disable characteristic 2 code. 1304 [Steve Henson] 1305 1306 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not 1307 all cases can be covered as some introduce binary incompatibilities. 1308 [Steve Henson] 1309 1310 *) Redirect RSA operations to FIPS module including keygen, 1311 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods. 1312 [Steve Henson] 1313 1314 *) Add similar low level API blocking to ciphers. 1315 [Steve Henson] 1316 1317 *) Low level digest APIs are not approved in FIPS mode: any attempt 1318 to use these will cause a fatal error. Applications that *really* want 1319 to use them can use the private_* version instead. 1320 [Steve Henson] 1321 1322 *) Redirect cipher operations to FIPS module for FIPS builds. 1323 [Steve Henson] 1324 1325 *) Redirect digest operations to FIPS module for FIPS builds. 1326 [Steve Henson] 1327 1328 *) Update build system to add "fips" flag which will link in fipscanister.o 1329 for static and shared library builds embedding a signature if needed. 1330 [Steve Henson] 1331 1332 *) Output TLS supported curves in preference order instead of numerical 1333 order. This is currently hardcoded for the highest order curves first. 1334 This should be configurable so applications can judge speed vs strength. 1335 [Steve Henson] 1336 1337 *) Add TLS v1.2 server support for client authentication. 1338 [Steve Henson] 1339 1340 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers 1341 and enable MD5. 1342 [Steve Henson] 1343 1344 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying 1345 FIPS modules versions. 1346 [Steve Henson] 1347 1348 *) Add TLS v1.2 client side support for client authentication. Keep cache 1349 of handshake records longer as we don't know the hash algorithm to use 1350 until after the certificate request message is received. 1351 [Steve Henson] 1352 1353 *) Initial TLS v1.2 client support. Add a default signature algorithms 1354 extension including all the algorithms we support. Parse new signature 1355 format in client key exchange. Relax some ECC signing restrictions for 1356 TLS v1.2 as indicated in RFC5246. 1357 [Steve Henson] 1358 1359 *) Add server support for TLS v1.2 signature algorithms extension. Switch 1360 to new signature format when needed using client digest preference. 1361 All server ciphersuites should now work correctly in TLS v1.2. No client 1362 support yet and no support for client certificates. 1363 [Steve Henson] 1364 1365 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch 1366 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based 1367 ciphersuites. At present only RSA key exchange ciphersuites work with 1368 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete 1369 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods 1370 and version checking. 1371 [Steve Henson] 1372 1373 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled 1374 with this defined it will not be affected by any changes to ssl internal 1375 structures. Add several utility functions to allow openssl application 1376 to work with OPENSSL_NO_SSL_INTERN defined. 1377 [Steve Henson] 1378 1379 *) Add SRP support. 1380 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie] 1381 1382 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id. 1383 [Steve Henson] 1384 1385 *) Permit abbreviated handshakes when renegotiating using the function 1386 SSL_renegotiate_abbreviated(). 1387 [Robin Seggelmann <seggelmann@fh-muenster.de>] 1388 1389 *) Add call to ENGINE_register_all_complete() to 1390 ENGINE_load_builtin_engines(), so some implementations get used 1391 automatically instead of needing explicit application support. 1392 [Steve Henson] 1393 1394 *) Add support for TLS key exporter as described in RFC5705. 1395 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson] 1396 1397 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only 1398 a few changes are required: 1399 1400 Add SSL_OP_NO_TLSv1_1 flag. 1401 Add TLSv1_1 methods. 1402 Update version checking logic to handle version 1.1. 1403 Add explicit IV handling (ported from DTLS code). 1404 Add command line options to s_client/s_server. 1405 [Steve Henson] 1406 1407 Changes between 1.0.0g and 1.0.0h [12 Mar 2012] 1408 1409 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 1410 in CMS and PKCS7 code. When RSA decryption fails use a random key for 1411 content decryption and always return the same error. Note: this attack 1412 needs on average 2^20 messages so it only affects automated senders. The 1413 old behaviour can be reenabled in the CMS code by setting the 1414 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 1415 an MMA defence is not necessary. 1416 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 1417 this issue. (CVE-2012-0884) 1418 [Steve Henson] 1419 1420 *) Fix CVE-2011-4619: make sure we really are receiving a 1421 client hello before rejecting multiple SGC restarts. Thanks to 1422 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 1423 [Steve Henson] 1424 1425 Changes between 1.0.0f and 1.0.0g [18 Jan 2012] 1426 1427 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 1428 Thanks to Antonio Martin, Enterprise Secure Access Research and 1429 Development, Cisco Systems, Inc. for discovering this bug and 1430 preparing a fix. (CVE-2012-0050) 1431 [Antonio Martin] 1432 1433 Changes between 1.0.0e and 1.0.0f [4 Jan 2012] 1434 1435 *) Nadhem Alfardan and Kenny Paterson have discovered an extension 1436 of the Vaudenay padding oracle attack on CBC mode encryption 1437 which enables an efficient plaintext recovery attack against 1438 the OpenSSL implementation of DTLS. Their attack exploits timing 1439 differences arising during decryption processing. A research 1440 paper describing this attack can be found at: 1441 http://www.isg.rhul.ac.uk/~kp/dtls.pdf 1442 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 1443 Security Group at Royal Holloway, University of London 1444 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 1445 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 1446 for preparing the fix. (CVE-2011-4108) 1447 [Robin Seggelmann, Michael Tuexen] 1448 1449 *) Clear bytes used for block padding of SSL 3.0 records. 1450 (CVE-2011-4576) 1451 [Adam Langley (Google)] 1452 1453 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 1454 Kadianakis <desnacked@gmail.com> for discovering this issue and 1455 Adam Langley for preparing the fix. (CVE-2011-4619) 1456 [Adam Langley (Google)] 1457 1458 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027) 1459 [Andrey Kulikov <amdeich@gmail.com>] 1460 1461 *) Prevent malformed RFC3779 data triggering an assertion failure. 1462 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 1463 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 1464 [Rob Austein <sra@hactrn.net>] 1465 1466 *) Improved PRNG seeding for VOS. 1467 [Paul Green <Paul.Green@stratus.com>] 1468 1469 *) Fix ssl_ciph.c set-up race. 1470 [Adam Langley (Google)] 1471 1472 *) Fix spurious failures in ecdsatest.c. 1473 [Emilia K��sper (Google)] 1474 1475 *) Fix the BIO_f_buffer() implementation (which was mixing different 1476 interpretations of the '..._len' fields). 1477 [Adam Langley (Google)] 1478 1479 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 1480 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 1481 threads won't reuse the same blinding coefficients. 1482 1483 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 1484 lock to call BN_BLINDING_invert_ex, and avoids one use of 1485 BN_BLINDING_update for each BN_BLINDING structure (previously, 1486 the last update always remained unused). 1487 [Emilia K��sper (Google)] 1488 1489 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf. 1490 [Bob Buckholz (Google)] 1491 1492 Changes between 1.0.0d and 1.0.0e [6 Sep 2011] 1493 1494 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted 1495 by initialising X509_STORE_CTX properly. (CVE-2011-3207) 1496 [Kaspar Brand <ossl@velox.ch>] 1497 1498 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 1499 for multi-threaded use of ECDH. (CVE-2011-3210) 1500 [Adam Langley (Google)] 1501 1502 *) Fix x509_name_ex_d2i memory leak on bad inputs. 1503 [Bodo Moeller] 1504 1505 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check 1506 signature public key algorithm by using OID xref utilities instead. 1507 Before this you could only use some ECC ciphersuites with SHA1 only. 1508 [Steve Henson] 1509 1510 *) Add protection against ECDSA timing attacks as mentioned in the paper 1511 by Billy Bob Brumley and Nicola Tuveri, see: 1512 1513 http://eprint.iacr.org/2011/232.pdf 1514 1515 [Billy Bob Brumley and Nicola Tuveri] 1516 1517 Changes between 1.0.0c and 1.0.0d [8 Feb 2011] 1518 1519 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 1520 [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 1521 1522 *) Fix bug in string printing code: if *any* escaping is enabled we must 1523 escape the escape character (backslash) or the resulting string is 1524 ambiguous. 1525 [Steve Henson] 1526 1527 Changes between 1.0.0b and 1.0.0c [2 Dec 2010] 1528 1529 *) Disable code workaround for ancient and obsolete Netscape browsers 1530 and servers: an attacker can use it in a ciphersuite downgrade attack. 1531 Thanks to Martin Rex for discovering this bug. CVE-2010-4180 1532 [Steve Henson] 1533 1534 *) Fixed J-PAKE implementation error, originally discovered by 1535 Sebastien Martini, further info and confirmation from Stefan 1536 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 1537 [Ben Laurie] 1538 1539 Changes between 1.0.0a and 1.0.0b [16 Nov 2010] 1540 1541 *) Fix extension code to avoid race conditions which can result in a buffer 1542 overrun vulnerability: resumed sessions must not be modified as they can 1543 be shared by multiple threads. CVE-2010-3864 1544 [Steve Henson] 1545 1546 *) Fix WIN32 build system to correctly link an ENGINE directory into 1547 a DLL. 1548 [Steve Henson] 1549 1550 Changes between 1.0.0 and 1.0.0a [01 Jun 2010] 1551 1552 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover 1553 (CVE-2010-1633) 1554 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>] 1555 1556 Changes between 0.9.8n and 1.0.0 [29 Mar 2010] 1557 1558 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher 1559 context. The operation can be customised via the ctrl mechanism in 1560 case ENGINEs want to include additional functionality. 1561 [Steve Henson] 1562 1563 *) Tolerate yet another broken PKCS#8 key format: private key value negative. 1564 [Steve Henson] 1565 1566 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to 1567 output hashes compatible with older versions of OpenSSL. 1568 [Willy Weisz <weisz@vcpc.univie.ac.at>] 1569 1570 *) Fix compression algorithm handling: if resuming a session use the 1571 compression algorithm of the resumed session instead of determining 1572 it from client hello again. Don't allow server to change algorithm. 1573 [Steve Henson] 1574 1575 *) Add load_crls() function to apps tidying load_certs() too. Add option 1576 to verify utility to allow additional CRLs to be included. 1577 [Steve Henson] 1578 1579 *) Update OCSP request code to permit adding custom headers to the request: 1580 some responders need this. 1581 [Steve Henson] 1582 1583 *) The function EVP_PKEY_sign() returns <=0 on error: check return code 1584 correctly. 1585 [Julia Lawall <julia@diku.dk>] 1586 1587 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it 1588 needlessly dereferenced structures, used obsolete functions and 1589 didn't handle all updated verify codes correctly. 1590 [Steve Henson] 1591 1592 *) Disable MD2 in the default configuration. 1593 [Steve Henson] 1594 1595 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to 1596 indicate the initial BIO being pushed or popped. This makes it possible 1597 to determine whether the BIO is the one explicitly called or as a result 1598 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so 1599 it handles reference counts correctly and doesn't zero out the I/O bio 1600 when it is not being explicitly popped. WARNING: applications which 1601 included workarounds for the old buggy behaviour will need to be modified 1602 or they could free up already freed BIOs. 1603 [Steve Henson] 1604 1605 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni 1606 renaming to all platforms (within the 0.9.8 branch, this was 1607 done conditionally on Netware platforms to avoid a name clash). 1608 [Guenter <lists@gknw.net>] 1609 1610 *) Add ECDHE and PSK support to DTLS. 1611 [Michael Tuexen <tuexen@fh-muenster.de>] 1612 1613 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't 1614 be used on C++. 1615 [Steve Henson] 1616 1617 *) Add "missing" function EVP_MD_flags() (without this the only way to 1618 retrieve a digest flags is by accessing the structure directly. Update 1619 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest 1620 or cipher is registered as in the "from" argument. Print out all 1621 registered digests in the dgst usage message instead of manually 1622 attempting to work them out. 1623 [Steve Henson] 1624 1625 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello: 1626 this allows the use of compression and extensions. Change default cipher 1627 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2 1628 by default unless an application cipher string requests it. 1629 [Steve Henson] 1630 1631 *) Alter match criteria in PKCS12_parse(). It used to try to use local 1632 key ids to find matching certificates and keys but some PKCS#12 files 1633 don't follow the (somewhat unwritten) rules and this strategy fails. 1634 Now just gather all certificates together and the first private key 1635 then look for the first certificate that matches the key. 1636 [Steve Henson] 1637 1638 *) Support use of registered digest and cipher names for dgst and cipher 1639 commands instead of having to add each one as a special case. So now 1640 you can do: 1641 1642 openssl sha256 foo 1643 1644 as well as: 1645 1646 openssl dgst -sha256 foo 1647 1648 and this works for ENGINE based algorithms too. 1649 1650 [Steve Henson] 1651 1652 *) Update Gost ENGINE to support parameter files. 1653 [Victor B. Wagner <vitus@cryptocom.ru>] 1654 1655 *) Support GeneralizedTime in ca utility. 1656 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson] 1657 1658 *) Enhance the hash format used for certificate directory links. The new 1659 form uses the canonical encoding (meaning equivalent names will work 1660 even if they aren't identical) and uses SHA1 instead of MD5. This form 1661 is incompatible with the older format and as a result c_rehash should 1662 be used to rebuild symbolic links. 1663 [Steve Henson] 1664 1665 *) Make PKCS#8 the default write format for private keys, replacing the 1666 traditional format. This form is standardised, more secure and doesn't 1667 include an implicit MD5 dependency. 1668 [Steve Henson] 1669 1670 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code 1671 committed to OpenSSL should pass this lot as a minimum. 1672 [Steve Henson] 1673 1674 *) Add session ticket override functionality for use by EAP-FAST. 1675 [Jouni Malinen <j@w1.fi>] 1676 1677 *) Modify HMAC functions to return a value. Since these can be implemented 1678 in an ENGINE errors can occur. 1679 [Steve Henson] 1680 1681 *) Type-checked OBJ_bsearch_ex. 1682 [Ben Laurie] 1683 1684 *) Type-checked OBJ_bsearch. Also some constification necessitated 1685 by type-checking. Still to come: TXT_DB, bsearch(?), 1686 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING, 1687 CONF_VALUE. 1688 [Ben Laurie] 1689 1690 *) New function OPENSSL_gmtime_adj() to add a specific number of days and 1691 seconds to a tm structure directly, instead of going through OS 1692 specific date routines. This avoids any issues with OS routines such 1693 as the year 2038 bug. New *_adj() functions for ASN1 time structures 1694 and X509_time_adj_ex() to cover the extended range. The existing 1695 X509_time_adj() is still usable and will no longer have any date issues. 1696 [Steve Henson] 1697 1698 *) Delta CRL support. New use deltas option which will attempt to locate 1699 and search any appropriate delta CRLs available. 1700 1701 This work was sponsored by Google. 1702 [Steve Henson] 1703 1704 *) Support for CRLs partitioned by reason code. Reorganise CRL processing 1705 code and add additional score elements. Validate alternate CRL paths 1706 as part of the CRL checking and indicate a new error "CRL path validation 1707 error" in this case. Applications wanting additional details can use 1708 the verify callback and check the new "parent" field. If this is not 1709 NULL CRL path validation is taking place. Existing applications wont 1710 see this because it requires extended CRL support which is off by 1711 default. 1712 1713 This work was sponsored by Google. 1714 [Steve Henson] 1715 1716 *) Support for freshest CRL extension. 1717 1718 This work was sponsored by Google. 1719 [Steve Henson] 1720 1721 *) Initial indirect CRL support. Currently only supported in the CRLs 1722 passed directly and not via lookup. Process certificate issuer 1723 CRL entry extension and lookup CRL entries by bother issuer name 1724 and serial number. Check and process CRL issuer entry in IDP extension. 1725 1726 This work was sponsored by Google. 1727 [Steve Henson] 1728 1729 *) Add support for distinct certificate and CRL paths. The CRL issuer 1730 certificate is validated separately in this case. Only enabled if 1731 an extended CRL support flag is set: this flag will enable additional 1732 CRL functionality in future. 1733 1734 This work was sponsored by Google. 1735 [Steve Henson] 1736 1737 *) Add support for policy mappings extension. 1738 1739 This work was sponsored by Google. 1740 [Steve Henson] 1741 1742 *) Fixes to pathlength constraint, self issued certificate handling, 1743 policy processing to align with RFC3280 and PKITS tests. 1744 1745 This work was sponsored by Google. 1746 [Steve Henson] 1747 1748 *) Support for name constraints certificate extension. DN, email, DNS 1749 and URI types are currently supported. 1750 1751 This work was sponsored by Google. 1752 [Steve Henson] 1753 1754 *) To cater for systems that provide a pointer-based thread ID rather 1755 than numeric, deprecate the current numeric thread ID mechanism and 1756 replace it with a structure and associated callback type. This 1757 mechanism allows a numeric "hash" to be extracted from a thread ID in 1758 either case, and on platforms where pointers are larger than 'long', 1759 mixing is done to help ensure the numeric 'hash' is usable even if it 1760 can't be guaranteed unique. The default mechanism is to use "&errno" 1761 as a pointer-based thread ID to distinguish between threads. 1762 1763 Applications that want to provide their own thread IDs should now use 1764 CRYPTO_THREADID_set_callback() to register a callback that will call 1765 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer(). 1766 1767 Note that ERR_remove_state() is now deprecated, because it is tied 1768 to the assumption that thread IDs are numeric. ERR_remove_state(0) 1769 to free the current thread's error state should be replaced by 1770 ERR_remove_thread_state(NULL). 1771 1772 (This new approach replaces the functions CRYPTO_set_idptr_callback(), 1773 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in 1774 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an 1775 application was previously providing a numeric thread callback that 1776 was inappropriate for distinguishing threads, then uniqueness might 1777 have been obtained with &errno that happened immediately in the 1778 intermediate development versions of OpenSSL; this is no longer the 1779 case, the numeric thread callback will now override the automatic use 1780 of &errno.) 1781 [Geoff Thorpe, with help from Bodo Moeller] 1782 1783 *) Initial support for different CRL issuing certificates. This covers a 1784 simple case where the self issued certificates in the chain exist and 1785 the real CRL issuer is higher in the existing chain. 1786 1787 This work was sponsored by Google. 1788 [Steve Henson] 1789 1790 *) Removed effectively defunct crypto/store from the build. 1791 [Ben Laurie] 1792 1793 *) Revamp of STACK to provide stronger type-checking. Still to come: 1794 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE, 1795 ASN1_STRING, CONF_VALUE. 1796 [Ben Laurie] 1797 1798 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer 1799 RAM on SSL connections. This option can save about 34k per idle SSL. 1800 [Nick Mathewson] 1801 1802 *) Revamp of LHASH to provide stronger type-checking. Still to come: 1803 STACK, TXT_DB, bsearch, qsort. 1804 [Ben Laurie] 1805 1806 *) Initial support for Cryptographic Message Syntax (aka CMS) based 1807 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility, 1808 support for data, signedData, compressedData, digestedData and 1809 encryptedData, envelopedData types included. Scripts to check against 1810 RFC4134 examples draft and interop and consistency checks of many 1811 content types and variants. 1812 [Steve Henson] 1813 1814 *) Add options to enc utility to support use of zlib compression BIO. 1815 [Steve Henson] 1816 1817 *) Extend mk1mf to support importing of options and assembly language 1818 files from Configure script, currently only included in VC-WIN32. 1819 The assembly language rules can now optionally generate the source 1820 files from the associated perl scripts. 1821 [Steve Henson] 1822 1823 *) Implement remaining functionality needed to support GOST ciphersuites. 1824 Interop testing has been performed using CryptoPro implementations. 1825 [Victor B. Wagner <vitus@cryptocom.ru>] 1826 1827 *) s390x assembler pack. 1828 [Andy Polyakov] 1829 1830 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU 1831 "family." 1832 [Andy Polyakov] 1833 1834 *) Implement Opaque PRF Input TLS extension as specified in 1835 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an 1836 official specification yet and no extension type assignment by 1837 IANA exists, this extension (for now) will have to be explicitly 1838 enabled when building OpenSSL by providing the extension number 1839 to use. For example, specify an option 1840 1841 -DTLSEXT_TYPE_opaque_prf_input=0x9527 1842 1843 to the "config" or "Configure" script to enable the extension, 1844 assuming extension number 0x9527 (which is a completely arbitrary 1845 and unofficial assignment based on the MD5 hash of the Internet 1846 Draft). Note that by doing so, you potentially lose 1847 interoperability with other TLS implementations since these might 1848 be using the same extension number for other purposes. 1849 1850 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the 1851 opaque PRF input value to use in the handshake. This will create 1852 an interal copy of the length-'len' string at 'src', and will 1853 return non-zero for success. 1854 1855 To get more control and flexibility, provide a callback function 1856 by using 1857 1858 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb) 1859 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg) 1860 1861 where 1862 1863 int (*cb)(SSL *, void *peerinput, size_t len, void *arg); 1864 void *arg; 1865 1866 Callback function 'cb' will be called in handshakes, and is 1867 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate. 1868 Argument 'arg' is for application purposes (the value as given to 1869 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly 1870 be provided to the callback function). The callback function 1871 has to return non-zero to report success: usually 1 to use opaque 1872 PRF input just if possible, or 2 to enforce use of the opaque PRF 1873 input. In the latter case, the library will abort the handshake 1874 if opaque PRF input is not successfully negotiated. 1875 1876 Arguments 'peerinput' and 'len' given to the callback function 1877 will always be NULL and 0 in the case of a client. A server will 1878 see the client's opaque PRF input through these variables if 1879 available (NULL and 0 otherwise). Note that if the server 1880 provides an opaque PRF input, the length must be the same as the 1881 length of the client's opaque PRF input. 1882 1883 Note that the callback function will only be called when creating 1884 a new session (session resumption can resume whatever was 1885 previously negotiated), and will not be called in SSL 2.0 1886 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or 1887 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended 1888 for applications that need to enforce opaque PRF input. 1889 1890 [Bodo Moeller] 1891 1892 *) Update ssl code to support digests other than SHA1+MD5 for handshake 1893 MAC. 1894 1895 [Victor B. Wagner <vitus@cryptocom.ru>] 1896 1897 *) Add RFC4507 support to OpenSSL. This includes the corrections in 1898 RFC4507bis. The encrypted ticket format is an encrypted encoded 1899 SSL_SESSION structure, that way new session features are automatically 1900 supported. 1901 1902 If a client application caches session in an SSL_SESSION structure 1903 support is transparent because tickets are now stored in the encoded 1904 SSL_SESSION. 1905 1906 The SSL_CTX structure automatically generates keys for ticket 1907 protection in servers so again support should be possible 1908 with no application modification. 1909 1910 If a client or server wishes to disable RFC4507 support then the option 1911 SSL_OP_NO_TICKET can be set. 1912 1913 Add a TLS extension debugging callback to allow the contents of any client 1914 or server extensions to be examined. 1915 1916 This work was sponsored by Google. 1917 [Steve Henson] 1918 1919 *) Final changes to avoid use of pointer pointer casts in OpenSSL. 1920 OpenSSL should now compile cleanly on gcc 4.2 1921 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson] 1922 1923 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC 1924 support including streaming MAC support: this is required for GOST 1925 ciphersuite support. 1926 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson] 1927 1928 *) Add option -stream to use PKCS#7 streaming in smime utility. New 1929 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream() 1930 to output in BER and PEM format. 1931 [Steve Henson] 1932 1933 *) Experimental support for use of HMAC via EVP_PKEY interface. This 1934 allows HMAC to be handled via the EVP_DigestSign*() interface. The 1935 EVP_PKEY "key" in this case is the HMAC key, potentially allowing 1936 ENGINE support for HMAC keys which are unextractable. New -mac and 1937 -macopt options to dgst utility. 1938 [Steve Henson] 1939 1940 *) New option -sigopt to dgst utility. Update dgst to use 1941 EVP_Digest{Sign,Verify}*. These two changes make it possible to use 1942 alternative signing paramaters such as X9.31 or PSS in the dgst 1943 utility. 1944 [Steve Henson] 1945 1946 *) Change ssl_cipher_apply_rule(), the internal function that does 1947 the work each time a ciphersuite string requests enabling 1948 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or 1949 removing ("!foo+bar") a class of ciphersuites: Now it maintains 1950 the order of disabled ciphersuites such that those ciphersuites 1951 that most recently went from enabled to disabled not only stay 1952 in order with respect to each other, but also have higher priority 1953 than other disabled ciphersuites the next time ciphersuites are 1954 enabled again. 1955 1956 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable 1957 the same ciphersuites as with "HIGH" alone, but in a specific 1958 order where the PSK ciphersuites come first (since they are the 1959 most recently disabled ciphersuites when "HIGH" is parsed). 1960 1961 Also, change ssl_create_cipher_list() (using this new 1962 funcionality) such that between otherwise identical 1963 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in 1964 the default order. 1965 [Bodo Moeller] 1966 1967 *) Change ssl_create_cipher_list() so that it automatically 1968 arranges the ciphersuites in reasonable order before starting 1969 to process the rule string. Thus, the definition for "DEFAULT" 1970 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but 1971 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH". 1972 This makes it much easier to arrive at a reasonable default order 1973 in applications for which anonymous ciphers are OK (meaning 1974 that you can't actually use DEFAULT). 1975 [Bodo Moeller; suggested by Victor Duchovni] 1976 1977 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string 1978 processing) into multiple integers instead of setting 1979 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK", 1980 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer. 1981 (These masks as well as the individual bit definitions are hidden 1982 away into the non-exported interface ssl/ssl_locl.h, so this 1983 change to the definition of the SSL_CIPHER structure shouldn't 1984 affect applications.) This give us more bits for each of these 1985 categories, so there is no longer a need to coagulate AES128 and 1986 AES256 into a single algorithm bit, and to coagulate Camellia128 1987 and Camellia256 into a single algorithm bit, which has led to all 1988 kinds of kludges. 1989 1990 Thus, among other things, the kludge introduced in 0.9.7m and 1991 0.9.8e for masking out AES256 independently of AES128 or masking 1992 out Camellia256 independently of AES256 is not needed here in 0.9.9. 1993 1994 With the change, we also introduce new ciphersuite aliases that 1995 so far were missing: "AES128", "AES256", "CAMELLIA128", and 1996 "CAMELLIA256". 1997 [Bodo Moeller] 1998 1999 *) Add support for dsa-with-SHA224 and dsa-with-SHA256. 2000 Use the leftmost N bytes of the signature input if the input is 2001 larger than the prime q (with N being the size in bytes of q). 2002 [Nils Larsch] 2003 2004 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses 2005 it yet and it is largely untested. 2006 [Steve Henson] 2007 2008 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types. 2009 [Nils Larsch] 2010 2011 *) Initial incomplete changes to avoid need for function casts in OpenSSL 2012 some compilers (gcc 4.2 and later) reject their use. Safestack is 2013 reimplemented. Update ASN1 to avoid use of legacy functions. 2014 [Steve Henson] 2015 2016 *) Win32/64 targets are linked with Winsock2. 2017 [Andy Polyakov] 2018 2019 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected 2020 to external functions. This can be used to increase CRL handling 2021 efficiency especially when CRLs are very large by (for example) storing 2022 the CRL revoked certificates in a database. 2023 [Steve Henson] 2024 2025 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so 2026 new CRLs added to a directory can be used. New command line option 2027 -verify_return_error to s_client and s_server. This causes real errors 2028 to be returned by the verify callback instead of carrying on no matter 2029 what. This reflects the way a "real world" verify callback would behave. 2030 [Steve Henson] 2031 2032 *) GOST engine, supporting several GOST algorithms and public key formats. 2033 Kindly donated by Cryptocom. 2034 [Cryptocom] 2035 2036 *) Partial support for Issuing Distribution Point CRL extension. CRLs 2037 partitioned by DP are handled but no indirect CRL or reason partitioning 2038 (yet). Complete overhaul of CRL handling: now the most suitable CRL is 2039 selected via a scoring technique which handles IDP and AKID in CRLs. 2040 [Steve Henson] 2041 2042 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which 2043 will ultimately be used for all verify operations: this will remove the 2044 X509_STORE dependency on certificate verification and allow alternative 2045 lookup methods. X509_STORE based implementations of these two callbacks. 2046 [Steve Henson] 2047 2048 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names. 2049 Modify get_crl() to find a valid (unexpired) CRL if possible. 2050 [Steve Henson] 2051 2052 *) New function X509_CRL_match() to check if two CRLs are identical. Normally 2053 this would be called X509_CRL_cmp() but that name is already used by 2054 a function that just compares CRL issuer names. Cache several CRL 2055 extensions in X509_CRL structure and cache CRLDP in X509. 2056 [Steve Henson] 2057 2058 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name) 2059 this maps equivalent X509_NAME structures into a consistent structure. 2060 Name comparison can then be performed rapidly using memcmp(). 2061 [Steve Henson] 2062 2063 *) Non-blocking OCSP request processing. Add -timeout option to ocsp 2064 utility. 2065 [Steve Henson] 2066 2067 *) Allow digests to supply their own micalg string for S/MIME type using 2068 the ctrl EVP_MD_CTRL_MICALG. 2069 [Steve Henson] 2070 2071 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the 2072 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN 2073 ctrl. It can then customise the structure before and/or after signing 2074 if necessary. 2075 [Steve Henson] 2076 2077 *) New function OBJ_add_sigid() to allow application defined signature OIDs 2078 to be added to OpenSSLs internal tables. New function OBJ_sigid_free() 2079 to free up any added signature OIDs. 2080 [Steve Henson] 2081 2082 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(), 2083 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal 2084 digest and cipher tables. New options added to openssl utility: 2085 list-message-digest-algorithms and list-cipher-algorithms. 2086 [Steve Henson] 2087 2088 *) Change the array representation of binary polynomials: the list 2089 of degrees of non-zero coefficients is now terminated with -1. 2090 Previously it was terminated with 0, which was also part of the 2091 value; thus, the array representation was not applicable to 2092 polynomials where t^0 has coefficient zero. This change makes 2093 the array representation useful in a more general context. 2094 [Douglas Stebila] 2095 2096 *) Various modifications and fixes to SSL/TLS cipher string 2097 handling. For ECC, the code now distinguishes between fixed ECDH 2098 with RSA certificates on the one hand and with ECDSA certificates 2099 on the other hand, since these are separate ciphersuites. The 2100 unused code for Fortezza ciphersuites has been removed. 2101 2102 For consistency with EDH, ephemeral ECDH is now called "EECDH" 2103 (not "ECDHE"). For consistency with the code for DH 2104 certificates, use of ECDH certificates is now considered ECDH 2105 authentication, not RSA or ECDSA authentication (the latter is 2106 merely the CA's signing algorithm and not actively used in the 2107 protocol). 2108 2109 The temporary ciphersuite alias "ECCdraft" is no longer 2110 available, and ECC ciphersuites are no longer excluded from "ALL" 2111 and "DEFAULT". The following aliases now exist for RFC 4492 2112 ciphersuites, most of these by analogy with the DH case: 2113 2114 kECDHr - ECDH cert, signed with RSA 2115 kECDHe - ECDH cert, signed with ECDSA 2116 kECDH - ECDH cert (signed with either RSA or ECDSA) 2117 kEECDH - ephemeral ECDH 2118 ECDH - ECDH cert or ephemeral ECDH 2119 2120 aECDH - ECDH cert 2121 aECDSA - ECDSA cert 2122 ECDSA - ECDSA cert 2123 2124 AECDH - anonymous ECDH 2125 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH") 2126 2127 [Bodo Moeller] 2128 2129 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported. 2130 Use correct micalg parameters depending on digest(s) in signed message. 2131 [Steve Henson] 2132 2133 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process 2134 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code. 2135 [Steve Henson] 2136 2137 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit 2138 an engine to register a method. Add ENGINE lookups for methods and 2139 functional reference processing. 2140 [Steve Henson] 2141 2142 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of 2143 EVP_{Sign,Verify}* which allow an application to customise the signature 2144 process. 2145 [Steve Henson] 2146 2147 *) New -resign option to smime utility. This adds one or more signers 2148 to an existing PKCS#7 signedData structure. Also -md option to use an 2149 alternative message digest algorithm for signing. 2150 [Steve Henson] 2151 2152 *) Tidy up PKCS#7 routines and add new functions to make it easier to 2153 create PKCS7 structures containing multiple signers. Update smime 2154 application to support multiple signers. 2155 [Steve Henson] 2156 2157 *) New -macalg option to pkcs12 utility to allow setting of an alternative 2158 digest MAC. 2159 [Steve Henson] 2160 2161 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC. 2162 Reorganize PBE internals to lookup from a static table using NIDs, 2163 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl: 2164 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative 2165 PRF which will be automatically used with PBES2. 2166 [Steve Henson] 2167 2168 *) Replace the algorithm specific calls to generate keys in "req" with the 2169 new API. 2170 [Steve Henson] 2171 2172 *) Update PKCS#7 enveloped data routines to use new API. This is now 2173 supported by any public key method supporting the encrypt operation. A 2174 ctrl is added to allow the public key algorithm to examine or modify 2175 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is 2176 a no op. 2177 [Steve Henson] 2178 2179 *) Add a ctrl to asn1 method to allow a public key algorithm to express 2180 a default digest type to use. In most cases this will be SHA1 but some 2181 algorithms (such as GOST) need to specify an alternative digest. The 2182 return value indicates how strong the prefernce is 1 means optional and 2183 2 is mandatory (that is it is the only supported type). Modify 2184 ASN1_item_sign() to accept a NULL digest argument to indicate it should 2185 use the default md. Update openssl utilities to use the default digest 2186 type for signing if it is not explicitly indicated. 2187 [Steve Henson] 2188 2189 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New 2190 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant 2191 signing method from the key type. This effectively removes the link 2192 between digests and public key types. 2193 [Steve Henson] 2194 2195 *) Add an OID cross reference table and utility functions. Its purpose is to 2196 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1, 2197 rsaEncryption. This will allow some of the algorithm specific hackery 2198 needed to use the correct OID to be removed. 2199 [Steve Henson] 2200 2201 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO 2202 structures for PKCS7_sign(). They are now set up by the relevant public 2203 key ASN1 method. 2204 [Steve Henson] 2205 2206 *) Add provisional EC pkey method with support for ECDSA and ECDH. 2207 [Steve Henson] 2208 2209 *) Add support for key derivation (agreement) in the API, DH method and 2210 pkeyutl. 2211 [Steve Henson] 2212 2213 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support 2214 public and private key formats. As a side effect these add additional 2215 command line functionality not previously available: DSA signatures can be 2216 generated and verified using pkeyutl and DH key support and generation in 2217 pkey, genpkey. 2218 [Steve Henson] 2219 2220 *) BeOS support. 2221 [Oliver Tappe <zooey@hirschkaefer.de>] 2222 2223 *) New make target "install_html_docs" installs HTML renditions of the 2224 manual pages. 2225 [Oliver Tappe <zooey@hirschkaefer.de>] 2226 2227 *) New utility "genpkey" this is analagous to "genrsa" etc except it can 2228 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to 2229 support key and parameter generation and add initial key generation 2230 functionality for RSA. 2231 [Steve Henson] 2232 2233 *) Add functions for main EVP_PKEY_method operations. The undocumented 2234 functions EVP_PKEY_{encrypt,decrypt} have been renamed to 2235 EVP_PKEY_{encrypt,decrypt}_old. 2236 [Steve Henson] 2237 2238 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public 2239 key API, doesn't do much yet. 2240 [Steve Henson] 2241 2242 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about 2243 public key algorithms. New option to openssl utility: 2244 "list-public-key-algorithms" to print out info. 2245 [Steve Henson] 2246 2247 *) Implement the Supported Elliptic Curves Extension for 2248 ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 2249 [Douglas Stebila] 2250 2251 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or 2252 EVP_CIPHER structures to avoid later problems in EVP_cleanup(). 2253 [Steve Henson] 2254 2255 *) New utilities pkey and pkeyparam. These are similar to algorithm specific 2256 utilities such as rsa, dsa, dsaparam etc except they process any key 2257 type. 2258 [Steve Henson] 2259 2260 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New 2261 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(), 2262 EVP_PKEY_print_param() to print public key data from an EVP_PKEY 2263 structure. 2264 [Steve Henson] 2265 2266 *) Initial support for pluggable public key ASN1. 2267 De-spaghettify the public key ASN1 handling. Move public and private 2268 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate 2269 algorithm specific handling to a single module within the relevant 2270 algorithm directory. Add functions to allow (near) opaque processing 2271 of public and private key structures. 2272 [Steve Henson] 2273 2274 *) Implement the Supported Point Formats Extension for 2275 ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 2276 [Douglas Stebila] 2277 2278 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members 2279 for the psk identity [hint] and the psk callback functions to the 2280 SSL_SESSION, SSL and SSL_CTX structure. 2281 2282 New ciphersuites: 2283 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA, 2284 PSK-AES256-CBC-SHA 2285 2286 New functions: 2287 SSL_CTX_use_psk_identity_hint 2288 SSL_get_psk_identity_hint 2289 SSL_get_psk_identity 2290 SSL_use_psk_identity_hint 2291 2292 [Mika Kousa and Pasi Eronen of Nokia Corporation] 2293 2294 *) Add RFC 3161 compliant time stamp request creation, response generation 2295 and response verification functionality. 2296 [Zolt��n Gl��zik <zglozik@opentsa.org>, The OpenTSA Project] 2297 2298 *) Add initial support for TLS extensions, specifically for the server_name 2299 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 2300 have new members for a host name. The SSL data structure has an 2301 additional member SSL_CTX *initial_ctx so that new sessions can be 2302 stored in that context to allow for session resumption, even after the 2303 SSL has been switched to a new SSL_CTX in reaction to a client's 2304 server_name extension. 2305 2306 New functions (subject to change): 2307 2308 SSL_get_servername() 2309 SSL_get_servername_type() 2310 SSL_set_SSL_CTX() 2311 2312 New CTRL codes and macros (subject to change): 2313 2314 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 2315 - SSL_CTX_set_tlsext_servername_callback() 2316 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 2317 - SSL_CTX_set_tlsext_servername_arg() 2318 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 2319 2320 openssl s_client has a new '-servername ...' option. 2321 2322 openssl s_server has new options '-servername_host ...', '-cert2 ...', 2323 '-key2 ...', '-servername_fatal' (subject to change). This allows 2324 testing the HostName extension for a specific single host name ('-cert' 2325 and '-key' remain fallbacks for handshakes without HostName 2326 negotiation). If the unrecogninzed_name alert has to be sent, this by 2327 default is a warning; it becomes fatal with the '-servername_fatal' 2328 option. 2329 2330 [Peter Sylvester, Remy Allais, Christophe Renou] 2331 2332 *) Whirlpool hash implementation is added. 2333 [Andy Polyakov] 2334 2335 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to 2336 bn(64,32). Because of instruction set limitations it doesn't have 2337 any negative impact on performance. This was done mostly in order 2338 to make it possible to share assembler modules, such as bn_mul_mont 2339 implementations, between 32- and 64-bit builds without hassle. 2340 [Andy Polyakov] 2341 2342 *) Move code previously exiled into file crypto/ec/ec2_smpt.c 2343 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP 2344 macro. 2345 [Bodo Moeller] 2346 2347 *) New candidate for BIGNUM assembler implementation, bn_mul_mont, 2348 dedicated Montgomery multiplication procedure, is introduced. 2349 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher 2350 "64-bit" performance on certain 32-bit targets. 2351 [Andy Polyakov] 2352 2353 *) New option SSL_OP_NO_COMP to disable use of compression selectively 2354 in SSL structures. New SSL ctrl to set maximum send fragment size. 2355 Save memory by seeting the I/O buffer sizes dynamically instead of 2356 using the maximum available value. 2357 [Steve Henson] 2358 2359 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code 2360 in addition to the text details. 2361 [Bodo Moeller] 2362 2363 *) Very, very preliminary EXPERIMENTAL support for printing of general 2364 ASN1 structures. This currently produces rather ugly output and doesn't 2365 handle several customised structures at all. 2366 [Steve Henson] 2367 2368 *) Integrated support for PVK file format and some related formats such 2369 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support 2370 these in the 'rsa' and 'dsa' utilities. 2371 [Steve Henson] 2372 2373 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line. 2374 [Steve Henson] 2375 2376 *) Remove the ancient ASN1_METHOD code. This was only ever used in one 2377 place for the (very old) "NETSCAPE" format certificates which are now 2378 handled using new ASN1 code equivalents. 2379 [Steve Henson] 2380 2381 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD 2382 pointer and make the SSL_METHOD parameter in SSL_CTX_new, 2383 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'. 2384 [Nils Larsch] 2385 2386 *) Modify CRL distribution points extension code to print out previously 2387 unsupported fields. Enhance extension setting code to allow setting of 2388 all fields. 2389 [Steve Henson] 2390 2391 *) Add print and set support for Issuing Distribution Point CRL extension. 2392 [Steve Henson] 2393 2394 *) Change 'Configure' script to enable Camellia by default. 2395 [NTT] 2396 2397 Changes between 0.9.8m and 0.9.8n [24 Mar 2010] 2398 2399 *) When rejecting SSL/TLS records due to an incorrect version number, never 2400 update s->server with a new major version number. As of 2401 - OpenSSL 0.9.8m if 'short' is a 16-bit type, 2402 - OpenSSL 0.9.8f if 'short' is longer than 16 bits, 2403 the previous behavior could result in a read attempt at NULL when 2404 receiving specific incorrect SSL/TLS records once record payload 2405 protection is active. (CVE-2010-0740) 2406 [Bodo Moeller, Adam Langley <agl@chromium.org>] 2407 2408 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL 2409 could be crashed if the relevant tables were not present (e.g. chrooted). 2410 [Tomas Hoger <thoger@redhat.com>] 2411 2412 Changes between 0.9.8l and 0.9.8m [25 Feb 2010] 2413 2414 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245) 2415 [Martin Olsson, Neel Mehta] 2416 2417 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to 2418 accommodate for stack sorting, always a write lock!). 2419 [Bodo Moeller] 2420 2421 *) On some versions of WIN32 Heap32Next is very slow. This can cause 2422 excessive delays in the RAND_poll(): over a minute. As a workaround 2423 include a time check in the inner Heap32Next loop too. 2424 [Steve Henson] 2425 2426 *) The code that handled flushing of data in SSL/TLS originally used the 2427 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused 2428 the problem outlined in PR#1949. The fix suggested there however can 2429 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions 2430 of Apache). So instead simplify the code to flush unconditionally. 2431 This should be fine since flushing with no data to flush is a no op. 2432 [Steve Henson] 2433 2434 *) Handle TLS versions 2.0 and later properly and correctly use the 2435 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way 2436 off ancient servers have a habit of sticking around for a while... 2437 [Steve Henson] 2438 2439 *) Modify compression code so it frees up structures without using the 2440 ex_data callbacks. This works around a problem where some applications 2441 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when 2442 restarting) then use compression (e.g. SSL with compression) later. 2443 This results in significant per-connection memory leaks and 2444 has caused some security issues including CVE-2008-1678 and 2445 CVE-2009-4355. 2446 [Steve Henson] 2447 2448 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't 2449 change when encrypting or decrypting. 2450 [Bodo Moeller] 2451 2452 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to 2453 connect and renegotiate with servers which do not support RI. 2454 Until RI is more widely deployed this option is enabled by default. 2455 [Steve Henson] 2456 2457 *) Add "missing" ssl ctrls to clear options and mode. 2458 [Steve Henson] 2459 2460 *) If client attempts to renegotiate and doesn't support RI respond with 2461 a no_renegotiation alert as required by RFC5746. Some renegotiating 2462 TLS clients will continue a connection gracefully when they receive 2463 the alert. Unfortunately OpenSSL mishandled this alert and would hang 2464 waiting for a server hello which it will never receive. Now we treat a 2465 received no_renegotiation alert as a fatal error. This is because 2466 applications requesting a renegotiation might well expect it to succeed 2467 and would have no code in place to handle the server denying it so the 2468 only safe thing to do is to terminate the connection. 2469 [Steve Henson] 2470 2471 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if 2472 peer supports secure renegotiation and 0 otherwise. Print out peer 2473 renegotiation support in s_client/s_server. 2474 [Steve Henson] 2475 2476 *) Replace the highly broken and deprecated SPKAC certification method with 2477 the updated NID creation version. This should correctly handle UTF8. 2478 [Steve Henson] 2479 2480 *) Implement RFC5746. Re-enable renegotiation but require the extension 2481 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION 2482 turns out to be a bad idea. It has been replaced by 2483 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with 2484 SSL_CTX_set_options(). This is really not recommended unless you 2485 know what you are doing. 2486 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson] 2487 2488 *) Fixes to stateless session resumption handling. Use initial_ctx when 2489 issuing and attempting to decrypt tickets in case it has changed during 2490 servername handling. Use a non-zero length session ID when attempting 2491 stateless session resumption: this makes it possible to determine if 2492 a resumption has occurred immediately after receiving server hello 2493 (several places in OpenSSL subtly assume this) instead of later in 2494 the handshake. 2495 [Steve Henson] 2496 2497 *) The functions ENGINE_ctrl(), OPENSSL_isservice(), 2498 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error 2499 fixes for a few places where the return code is not checked 2500 correctly. 2501 [Julia Lawall <julia@diku.dk>] 2502 2503 *) Add --strict-warnings option to Configure script to include devteam 2504 warnings in other configurations. 2505 [Steve Henson] 2506 2507 *) Add support for --libdir option and LIBDIR variable in makefiles. This 2508 makes it possible to install openssl libraries in locations which 2509 have names other than "lib", for example "/usr/lib64" which some 2510 systems need. 2511 [Steve Henson, based on patch from Jeremy Utley] 2512 2513 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of 2514 X690 8.9.12 and can produce some misleading textual output of OIDs. 2515 [Steve Henson, reported by Dan Kaminsky] 2516 2517 *) Delete MD2 from algorithm tables. This follows the recommendation in 2518 several standards that it is not used in new applications due to 2519 several cryptographic weaknesses. For binary compatibility reasons 2520 the MD2 API is still compiled in by default. 2521 [Steve Henson] 2522 2523 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved 2524 and restored. 2525 [Steve Henson] 2526 2527 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and 2528 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name 2529 clash. 2530 [Guenter <lists@gknw.net>] 2531 2532 *) Fix the server certificate chain building code to use X509_verify_cert(), 2533 it used to have an ad-hoc builder which was unable to cope with anything 2534 other than a simple chain. 2535 [David Woodhouse <dwmw2@infradead.org>, Steve Henson] 2536 2537 *) Don't check self signed certificate signatures in X509_verify_cert() 2538 by default (a flag can override this): it just wastes time without 2539 adding any security. As a useful side effect self signed root CAs 2540 with non-FIPS digests are now usable in FIPS mode. 2541 [Steve Henson] 2542 2543 *) In dtls1_process_out_of_seq_message() the check if the current message 2544 is already buffered was missing. For every new message was memory 2545 allocated, allowing an attacker to perform an denial of service attack 2546 with sending out of seq handshake messages until there is no memory 2547 left. Additionally every future messege was buffered, even if the 2548 sequence number made no sense and would be part of another handshake. 2549 So only messages with sequence numbers less than 10 in advance will be 2550 buffered. (CVE-2009-1378) 2551 [Robin Seggelmann, discovered by Daniel Mentz] 2552 2553 *) Records are buffered if they arrive with a future epoch to be 2554 processed after finishing the corresponding handshake. There is 2555 currently no limitation to this buffer allowing an attacker to perform 2556 a DOS attack with sending records with future epochs until there is no 2557 memory left. This patch adds the pqueue_size() function to detemine 2558 the size of a buffer and limits the record buffer to 100 entries. 2559 (CVE-2009-1377) 2560 [Robin Seggelmann, discovered by Daniel Mentz] 2561 2562 *) Keep a copy of frag->msg_header.frag_len so it can be used after the 2563 parent structure is freed. (CVE-2009-1379) 2564 [Daniel Mentz] 2565 2566 *) Handle non-blocking I/O properly in SSL_shutdown() call. 2567 [Darryl Miles <darryl-mailinglists@netbauds.net>] 2568 2569 *) Add 2.5.4.* OIDs 2570 [Ilya O. <vrghost@gmail.com>] 2571 2572 Changes between 0.9.8k and 0.9.8l [5 Nov 2009] 2573 2574 *) Disable renegotiation completely - this fixes a severe security 2575 problem (CVE-2009-3555) at the cost of breaking all 2576 renegotiation. Renegotiation can be re-enabled by setting 2577 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at 2578 run-time. This is really not recommended unless you know what 2579 you're doing. 2580 [Ben Laurie] 2581 2582 Changes between 0.9.8j and 0.9.8k [25 Mar 2009] 2583 2584 *) Don't set val to NULL when freeing up structures, it is freed up by 2585 underlying code. If sizeof(void *) > sizeof(long) this can result in 2586 zeroing past the valid field. (CVE-2009-0789) 2587 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>] 2588 2589 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not 2590 checked correctly. This would allow some invalid signed attributes to 2591 appear to verify correctly. (CVE-2009-0591) 2592 [Ivan Nestlerode <inestlerode@us.ibm.com>] 2593 2594 *) Reject UniversalString and BMPString types with invalid lengths. This 2595 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have 2596 a legal length. (CVE-2009-0590) 2597 [Steve Henson] 2598 2599 *) Set S/MIME signing as the default purpose rather than setting it 2600 unconditionally. This allows applications to override it at the store 2601 level. 2602 [Steve Henson] 2603 2604 *) Permit restricted recursion of ASN1 strings. This is needed in practice 2605 to handle some structures. 2606 [Steve Henson] 2607 2608 *) Improve efficiency of mem_gets: don't search whole buffer each time 2609 for a '\n' 2610 [Jeremy Shapiro <jnshapir@us.ibm.com>] 2611 2612 *) New -hex option for openssl rand. 2613 [Matthieu Herrb] 2614 2615 *) Print out UTF8String and NumericString when parsing ASN1. 2616 [Steve Henson] 2617 2618 *) Support NumericString type for name components. 2619 [Steve Henson] 2620 2621 *) Allow CC in the environment to override the automatically chosen 2622 compiler. Note that nothing is done to ensure flags work with the 2623 chosen compiler. 2624 [Ben Laurie] 2625 2626 Changes between 0.9.8i and 0.9.8j [07 Jan 2009] 2627 2628 *) Properly check EVP_VerifyFinal() and similar return values 2629 (CVE-2008-5077). 2630 [Ben Laurie, Bodo Moeller, Google Security Team] 2631 2632 *) Enable TLS extensions by default. 2633 [Ben Laurie] 2634 2635 *) Allow the CHIL engine to be loaded, whether the application is 2636 multithreaded or not. (This does not release the developer from the 2637 obligation to set up the dynamic locking callbacks.) 2638 [Sander Temme <sander@temme.net>] 2639 2640 *) Use correct exit code if there is an error in dgst command. 2641 [Steve Henson; problem pointed out by Roland Dirlewanger] 2642 2643 *) Tweak Configure so that you need to say "experimental-jpake" to enable 2644 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. 2645 [Bodo Moeller] 2646 2647 *) Add experimental JPAKE support, including demo authentication in 2648 s_client and s_server. 2649 [Ben Laurie] 2650 2651 *) Set the comparison function in v3_addr_canonize(). 2652 [Rob Austein <sra@hactrn.net>] 2653 2654 *) Add support for XMPP STARTTLS in s_client. 2655 [Philip Paeps <philip@freebsd.org>] 2656 2657 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior 2658 to ensure that even with this option, only ciphersuites in the 2659 server's preference list will be accepted. (Note that the option 2660 applies only when resuming a session, so the earlier behavior was 2661 just about the algorithm choice for symmetric cryptography.) 2662 [Bodo Moeller] 2663 2664 Changes between 0.9.8h and 0.9.8i [15 Sep 2008] 2665 2666 *) Fix NULL pointer dereference if a DTLS server received 2667 ChangeCipherSpec as first record (CVE-2009-1386). 2668 [PR #1679] 2669 2670 *) Fix a state transitition in s3_srvr.c and d1_srvr.c 2671 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...). 2672 [Nagendra Modadugu] 2673 2674 *) The fix in 0.9.8c that supposedly got rid of unsafe 2675 double-checked locking was incomplete for RSA blinding, 2676 addressing just one layer of what turns out to have been 2677 doubly unsafe triple-checked locking. 2678 2679 So now fix this for real by retiring the MONT_HELPER macro 2680 in crypto/rsa/rsa_eay.c. 2681 2682 [Bodo Moeller; problem pointed out by Marius Schilder] 2683 2684 *) Various precautionary measures: 2685 2686 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). 2687 2688 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). 2689 (NB: This would require knowledge of the secret session ticket key 2690 to exploit, in which case you'd be SOL either way.) 2691 2692 - Change bn_nist.c so that it will properly handle input BIGNUMs 2693 outside the expected range. 2694 2695 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG 2696 builds. 2697 2698 [Neel Mehta, Bodo Moeller] 2699 2700 *) Allow engines to be "soft loaded" - i.e. optionally don't die if 2701 the load fails. Useful for distros. 2702 [Ben Laurie and the FreeBSD team] 2703 2704 *) Add support for Local Machine Keyset attribute in PKCS#12 files. 2705 [Steve Henson] 2706 2707 *) Fix BN_GF2m_mod_arr() top-bit cleanup code. 2708 [Huang Ying] 2709 2710 *) Expand ENGINE to support engine supplied SSL client certificate functions. 2711 2712 This work was sponsored by Logica. 2713 [Steve Henson] 2714 2715 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows 2716 keystores. Support for SSL/TLS client authentication too. 2717 Not compiled unless enable-capieng specified to Configure. 2718 2719 This work was sponsored by Logica. 2720 [Steve Henson] 2721 2722 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using 2723 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain 2724 attribute creation routines such as certifcate requests and PKCS#12 2725 files. 2726 [Steve Henson] 2727 2728 Changes between 0.9.8g and 0.9.8h [28 May 2008] 2729 2730 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS 2731 handshake which could lead to a cilent crash as found using the 2732 Codenomicon TLS test suite (CVE-2008-1672) 2733 [Steve Henson, Mark Cox] 2734 2735 *) Fix double free in TLS server name extensions which could lead to 2736 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891) 2737 [Joe Orton] 2738 2739 *) Clear error queue in SSL_CTX_use_certificate_chain_file() 2740 2741 Clear the error queue to ensure that error entries left from 2742 older function calls do not interfere with the correct operation. 2743 [Lutz Jaenicke, Erik de Castro Lopo] 2744 2745 *) Remove root CA certificates of commercial CAs: 2746 2747 The OpenSSL project does not recommend any specific CA and does not 2748 have any policy with respect to including or excluding any CA. 2749 Therefore it does not make any sense to ship an arbitrary selection 2750 of root CA certificates with the OpenSSL software. 2751 [Lutz Jaenicke] 2752 2753 *) RSA OAEP patches to fix two separate invalid memory reads. 2754 The first one involves inputs when 'lzero' is greater than 2755 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes 2756 before the beginning of from). The second one involves inputs where 2757 the 'db' section contains nothing but zeroes (there is a one-byte 2758 invalid read after the end of 'db'). 2759 [Ivan Nestlerode <inestlerode@us.ibm.com>] 2760 2761 *) Partial backport from 0.9.9-dev: 2762 2763 Introduce bn_mul_mont (dedicated Montgomery multiplication 2764 procedure) as a candidate for BIGNUM assembler implementation. 2765 While 0.9.9-dev uses assembler for various architectures, only 2766 x86_64 is available by default here in the 0.9.8 branch, and 2767 32-bit x86 is available through a compile-time setting. 2768 2769 To try the 32-bit x86 assembler implementation, use Configure 2770 option "enable-montasm" (which exists only for this backport). 2771 2772 As "enable-montasm" for 32-bit x86 disclaims code stability 2773 anyway, in this constellation we activate additional code 2774 backported from 0.9.9-dev for further performance improvements, 2775 namely BN_from_montgomery_word. (To enable this otherwise, 2776 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".) 2777 2778 [Andy Polyakov (backport partially by Bodo Moeller)] 2779 2780 *) Add TLS session ticket callback. This allows an application to set 2781 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed 2782 values. This is useful for key rollover for example where several key 2783 sets may exist with different names. 2784 [Steve Henson] 2785 2786 *) Reverse ENGINE-internal logic for caching default ENGINE handles. 2787 This was broken until now in 0.9.8 releases, such that the only way 2788 a registered ENGINE could be used (assuming it initialises 2789 successfully on the host) was to explicitly set it as the default 2790 for the relevant algorithms. This is in contradiction with 0.9.7 2791 behaviour and the documentation. With this fix, when an ENGINE is 2792 registered into a given algorithm's table of implementations, the 2793 'uptodate' flag is reset so that auto-discovery will be used next 2794 time a new context for that algorithm attempts to select an 2795 implementation. 2796 [Ian Lister (tweaked by Geoff Thorpe)] 2797 2798 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 2799 implemention in the following ways: 2800 2801 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be 2802 hard coded. 2803 2804 Lack of BER streaming support means one pass streaming processing is 2805 only supported if data is detached: setting the streaming flag is 2806 ignored for embedded content. 2807 2808 CMS support is disabled by default and must be explicitly enabled 2809 with the enable-cms configuration option. 2810 [Steve Henson] 2811 2812 *) Update the GMP engine glue to do direct copies between BIGNUM and 2813 mpz_t when openssl and GMP use the same limb size. Otherwise the 2814 existing "conversion via a text string export" trick is still used. 2815 [Paul Sheer <paulsheer@gmail.com>] 2816 2817 *) Zlib compression BIO. This is a filter BIO which compressed and 2818 uncompresses any data passed through it. 2819 [Steve Henson] 2820 2821 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement 2822 RFC3394 compatible AES key wrapping. 2823 [Steve Henson] 2824 2825 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): 2826 sets string data without copying. X509_ALGOR_set0() and 2827 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) 2828 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data 2829 from an X509_ATTRIBUTE structure optionally checking it occurs only 2830 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied 2831 data. 2832 [Steve Henson] 2833 2834 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() 2835 to get the expected BN_FLG_CONSTTIME behavior. 2836 [Bodo Moeller (Google)] 2837 2838 *) Netware support: 2839 2840 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets 2841 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) 2842 - added some more tests to do_tests.pl 2843 - fixed RunningProcess usage so that it works with newer LIBC NDKs too 2844 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency 2845 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, 2846 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc 2847 - various changes to netware.pl to enable gcc-cross builds on Win32 2848 platform 2849 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) 2850 - various changes to fix missing prototype warnings 2851 - fixed x86nasm.pl to create correct asm files for NASM COFF output 2852 - added AES, WHIRLPOOL and CPUID assembler code to build files 2853 - added missing AES assembler make rules to mk1mf.pl 2854 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply 2855 [Guenter Knauf <eflash@gmx.net>] 2856 2857 *) Implement certificate status request TLS extension defined in RFC3546. 2858 A client can set the appropriate parameters and receive the encoded 2859 OCSP response via a callback. A server can query the supplied parameters 2860 and set the encoded OCSP response in the callback. Add simplified examples 2861 to s_client and s_server. 2862 [Steve Henson] 2863 2864 Changes between 0.9.8f and 0.9.8g [19 Oct 2007] 2865 2866 *) Fix various bugs: 2867 + Binary incompatibility of ssl_ctx_st structure 2868 + DTLS interoperation with non-compliant servers 2869 + Don't call get_session_cb() without proposed session 2870 + Fix ia64 assembler code 2871 [Andy Polyakov, Steve Henson] 2872 2873 Changes between 0.9.8e and 0.9.8f [11 Oct 2007] 2874 2875 *) DTLS Handshake overhaul. There were longstanding issues with 2876 OpenSSL DTLS implementation, which were making it impossible for 2877 RFC 4347 compliant client to communicate with OpenSSL server. 2878 Unfortunately just fixing these incompatibilities would "cut off" 2879 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e 2880 server keeps tolerating non RFC compliant syntax. The opposite is 2881 not true, 0.9.8f client can not communicate with earlier server. 2882 This update even addresses CVE-2007-4995. 2883 [Andy Polyakov] 2884 2885 *) Changes to avoid need for function casts in OpenSSL: some compilers 2886 (gcc 4.2 and later) reject their use. 2887 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, 2888 Steve Henson] 2889 2890 *) Add RFC4507 support to OpenSSL. This includes the corrections in 2891 RFC4507bis. The encrypted ticket format is an encrypted encoded 2892 SSL_SESSION structure, that way new session features are automatically 2893 supported. 2894 2895 If a client application caches session in an SSL_SESSION structure 2896 support is transparent because tickets are now stored in the encoded 2897 SSL_SESSION. 2898 2899 The SSL_CTX structure automatically generates keys for ticket 2900 protection in servers so again support should be possible 2901 with no application modification. 2902 2903 If a client or server wishes to disable RFC4507 support then the option 2904 SSL_OP_NO_TICKET can be set. 2905 2906 Add a TLS extension debugging callback to allow the contents of any client 2907 or server extensions to be examined. 2908 2909 This work was sponsored by Google. 2910 [Steve Henson] 2911 2912 *) Add initial support for TLS extensions, specifically for the server_name 2913 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 2914 have new members for a host name. The SSL data structure has an 2915 additional member SSL_CTX *initial_ctx so that new sessions can be 2916 stored in that context to allow for session resumption, even after the 2917 SSL has been switched to a new SSL_CTX in reaction to a client's 2918 server_name extension. 2919 2920 New functions (subject to change): 2921 2922 SSL_get_servername() 2923 SSL_get_servername_type() 2924 SSL_set_SSL_CTX() 2925 2926 New CTRL codes and macros (subject to change): 2927 2928 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 2929 - SSL_CTX_set_tlsext_servername_callback() 2930 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 2931 - SSL_CTX_set_tlsext_servername_arg() 2932 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 2933 2934 openssl s_client has a new '-servername ...' option. 2935 2936 openssl s_server has new options '-servername_host ...', '-cert2 ...', 2937 '-key2 ...', '-servername_fatal' (subject to change). This allows 2938 testing the HostName extension for a specific single host name ('-cert' 2939 and '-key' remain fallbacks for handshakes without HostName 2940 negotiation). If the unrecogninzed_name alert has to be sent, this by 2941 default is a warning; it becomes fatal with the '-servername_fatal' 2942 option. 2943 2944 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson] 2945 2946 *) Add AES and SSE2 assembly language support to VC++ build. 2947 [Steve Henson] 2948 2949 *) Mitigate attack on final subtraction in Montgomery reduction. 2950 [Andy Polyakov] 2951 2952 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 2953 (which previously caused an internal error). 2954 [Bodo Moeller] 2955 2956 *) Squeeze another 10% out of IGE mode when in != out. 2957 [Ben Laurie] 2958 2959 *) AES IGE mode speedup. 2960 [Dean Gaudet (Google)] 2961 2962 *) Add the Korean symmetric 128-bit cipher SEED (see 2963 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and 2964 add SEED ciphersuites from RFC 4162: 2965 2966 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" 2967 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" 2968 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" 2969 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" 2970 2971 To minimize changes between patchlevels in the OpenSSL 0.9.8 2972 series, SEED remains excluded from compilation unless OpenSSL 2973 is configured with 'enable-seed'. 2974 [KISA, Bodo Moeller] 2975 2976 *) Mitigate branch prediction attacks, which can be practical if a 2977 single processor is shared, allowing a spy process to extract 2978 information. For detailed background information, see 2979 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron, 2980 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL 2981 and Necessary Software Countermeasures"). The core of the change 2982 are new versions BN_div_no_branch() and 2983 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), 2984 respectively, which are slower, but avoid the security-relevant 2985 conditional branches. These are automatically called by BN_div() 2986 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one 2987 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to 2988 remove a conditional branch. 2989 2990 BN_FLG_CONSTTIME is the new name for the previous 2991 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just 2992 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag 2993 in the exponent causes BN_mod_exp_mont() to use the alternative 2994 implementation in BN_mod_exp_mont_consttime().) The old name 2995 remains as a deprecated alias. 2996 2997 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general 2998 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses 2999 constant-time implementations for more than just exponentiation. 3000 Here too the old name is kept as a deprecated alias. 3001 3002 BN_BLINDING_new() will now use BN_dup() for the modulus so that 3003 the BN_BLINDING structure gets an independent copy of the 3004 modulus. This means that the previous "BIGNUM *m" argument to 3005 BN_BLINDING_new() and to BN_BLINDING_create_param() now 3006 essentially becomes "const BIGNUM *m", although we can't actually 3007 change this in the header file before 0.9.9. It allows 3008 RSA_setup_blinding() to use BN_with_flags() on the modulus to 3009 enable BN_FLG_CONSTTIME. 3010 3011 [Matthew D Wood (Intel Corp)] 3012 3013 *) In the SSL/TLS server implementation, be strict about session ID 3014 context matching (which matters if an application uses a single 3015 external cache for different purposes). Previously, 3016 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was 3017 set. This did ensure strict client verification, but meant that, 3018 with applications using a single external cache for quite 3019 different requirements, clients could circumvent ciphersuite 3020 restrictions for a given session ID context by starting a session 3021 in a different context. 3022 [Bodo Moeller] 3023 3024 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 3025 a ciphersuite string such as "DEFAULT:RSA" cannot enable 3026 authentication-only ciphersuites. 3027 [Bodo Moeller] 3028 3029 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was 3030 not complete and could lead to a possible single byte overflow 3031 (CVE-2007-5135) [Ben Laurie] 3032 3033 Changes between 0.9.8d and 0.9.8e [23 Feb 2007] 3034 3035 *) Since AES128 and AES256 (and similarly Camellia128 and 3036 Camellia256) share a single mask bit in the logic of 3037 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 3038 kludge to work properly if AES128 is available and AES256 isn't 3039 (or if Camellia128 is available and Camellia256 isn't). 3040 [Victor Duchovni] 3041 3042 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c 3043 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): 3044 When a point or a seed is encoded in a BIT STRING, we need to 3045 prevent the removal of trailing zero bits to get the proper DER 3046 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case 3047 of a NamedBitList, for which trailing 0 bits need to be removed.) 3048 [Bodo Moeller] 3049 3050 *) Have SSL/TLS server implementation tolerate "mismatched" record 3051 protocol version while receiving ClientHello even if the 3052 ClientHello is fragmented. (The server can't insist on the 3053 particular protocol version it has chosen before the ServerHello 3054 message has informed the client about his choice.) 3055 [Bodo Moeller] 3056 3057 *) Add RFC 3779 support. 3058 [Rob Austein for ARIN, Ben Laurie] 3059 3060 *) Load error codes if they are not already present instead of using a 3061 static variable. This allows them to be cleanly unloaded and reloaded. 3062 Improve header file function name parsing. 3063 [Steve Henson] 3064 3065 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO 3066 or CAPABILITY handshake as required by RFCs. 3067 [Goetz Babin-Ebell] 3068 3069 Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 3070 3071 *) Introduce limits to prevent malicious keys being able to 3072 cause a denial of service. (CVE-2006-2940) 3073 [Steve Henson, Bodo Moeller] 3074 3075 *) Fix ASN.1 parsing of certain invalid structures that can result 3076 in a denial of service. (CVE-2006-2937) [Steve Henson] 3077 3078 *) Fix buffer overflow in SSL_get_shared_ciphers() function. 3079 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 3080 3081 *) Fix SSL client code which could crash if connecting to a 3082 malicious SSLv2 server. (CVE-2006-4343) 3083 [Tavis Ormandy and Will Drewry, Google Security Team] 3084 3085 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 3086 match only those. Before that, "AES256-SHA" would be interpreted 3087 as a pattern and match "AES128-SHA" too (since AES128-SHA got 3088 the same strength classification in 0.9.7h) as we currently only 3089 have a single AES bit in the ciphersuite description bitmap. 3090 That change, however, also applied to ciphersuite strings such as 3091 "RC4-MD5" that intentionally matched multiple ciphersuites -- 3092 namely, SSL 2.0 ciphersuites in addition to the more common ones 3093 from SSL 3.0/TLS 1.0. 3094 3095 So we change the selection algorithm again: Naming an explicit 3096 ciphersuite selects this one ciphersuite, and any other similar 3097 ciphersuite (same bitmap) from *other* protocol versions. 3098 Thus, "RC4-MD5" again will properly select both the SSL 2.0 3099 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 3100 3101 Since SSL 2.0 does not have any ciphersuites for which the 3102 128/256 bit distinction would be relevant, this works for now. 3103 The proper fix will be to use different bits for AES128 and 3104 AES256, which would have avoided the problems from the beginning; 3105 however, bits are scarce, so we can only do this in a new release 3106 (not just a patchlevel) when we can change the SSL_CIPHER 3107 definition to split the single 'unsigned long mask' bitmap into 3108 multiple values to extend the available space. 3109 3110 [Bodo Moeller] 3111 3112 Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 3113 3114 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 3115 (CVE-2006-4339) [Ben Laurie and Google Security Team] 3116 3117 *) Add AES IGE and biIGE modes. 3118 [Ben Laurie] 3119 3120 *) Change the Unix randomness entropy gathering to use poll() when 3121 possible instead of select(), since the latter has some 3122 undesirable limitations. 3123 [Darryl Miles via Richard Levitte and Bodo Moeller] 3124 3125 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special 3126 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 3127 cannot be implicitly activated as part of, e.g., the "AES" alias. 3128 However, please upgrade to OpenSSL 0.9.9[-dev] for 3129 non-experimental use of the ECC ciphersuites to get TLS extension 3130 support, which is required for curve and point format negotiation 3131 to avoid potential handshake problems. 3132 [Bodo Moeller] 3133 3134 *) Disable rogue ciphersuites: 3135 3136 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 3137 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 3138 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 3139 3140 The latter two were purportedly from 3141 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 3142 appear there. 3143 3144 Also deactivate the remaining ciphersuites from 3145 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 3146 unofficial, and the ID has long expired. 3147 [Bodo Moeller] 3148 3149 *) Fix RSA blinding Heisenbug (problems sometimes occured on 3150 dual-core machines) and other potential thread-safety issues. 3151 [Bodo Moeller] 3152 3153 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 3154 versions), which is now available for royalty-free use 3155 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). 3156 Also, add Camellia TLS ciphersuites from RFC 4132. 3157 3158 To minimize changes between patchlevels in the OpenSSL 0.9.8 3159 series, Camellia remains excluded from compilation unless OpenSSL 3160 is configured with 'enable-camellia'. 3161 [NTT] 3162 3163 *) Disable the padding bug check when compression is in use. The padding 3164 bug check assumes the first packet is of even length, this is not 3165 necessarily true if compresssion is enabled and can result in false 3166 positives causing handshake failure. The actual bug test is ancient 3167 code so it is hoped that implementations will either have fixed it by 3168 now or any which still have the bug do not support compression. 3169 [Steve Henson] 3170 3171 Changes between 0.9.8a and 0.9.8b [04 May 2006] 3172 3173 *) When applying a cipher rule check to see if string match is an explicit 3174 cipher suite and only match that one cipher suite if it is. 3175 [Steve Henson] 3176 3177 *) Link in manifests for VC++ if needed. 3178 [Austin Ziegler <halostatue@gmail.com>] 3179 3180 *) Update support for ECC-based TLS ciphersuites according to 3181 draft-ietf-tls-ecc-12.txt with proposed changes (but without 3182 TLS extensions, which are supported starting with the 0.9.9 3183 branch, not in the OpenSSL 0.9.8 branch). 3184 [Douglas Stebila] 3185 3186 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 3187 opaque EVP_CIPHER_CTX handling. 3188 [Steve Henson] 3189 3190 *) Fixes and enhancements to zlib compression code. We now only use 3191 "zlib1.dll" and use the default __cdecl calling convention on Win32 3192 to conform with the standards mentioned here: 3193 http://www.zlib.net/DLL_FAQ.txt 3194 Static zlib linking now works on Windows and the new --with-zlib-include 3195 --with-zlib-lib options to Configure can be used to supply the location 3196 of the headers and library. Gracefully handle case where zlib library 3197 can't be loaded. 3198 [Steve Henson] 3199 3200 *) Several fixes and enhancements to the OID generation code. The old code 3201 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 3202 handle numbers larger than ULONG_MAX, truncated printing and had a 3203 non standard OBJ_obj2txt() behaviour. 3204 [Steve Henson] 3205 3206 *) Add support for building of engines under engine/ as shared libraries 3207 under VC++ build system. 3208 [Steve Henson] 3209 3210 *) Corrected the numerous bugs in the Win32 path splitter in DSO. 3211 Hopefully, we will not see any false combination of paths any more. 3212 [Richard Levitte] 3213 3214 Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 3215 3216 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 3217 (part of SSL_OP_ALL). This option used to disable the 3218 countermeasure against man-in-the-middle protocol-version 3219 rollback in the SSL 2.0 server implementation, which is a bad 3220 idea. (CVE-2005-2969) 3221 3222 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 3223 for Information Security, National Institute of Advanced Industrial 3224 Science and Technology [AIST], Japan)] 3225 3226 *) Add two function to clear and return the verify parameter flags. 3227 [Steve Henson] 3228 3229 *) Keep cipherlists sorted in the source instead of sorting them at 3230 runtime, thus removing the need for a lock. 3231 [Nils Larsch] 3232 3233 *) Avoid some small subgroup attacks in Diffie-Hellman. 3234 [Nick Mathewson and Ben Laurie] 3235 3236 *) Add functions for well-known primes. 3237 [Nick Mathewson] 3238 3239 *) Extended Windows CE support. 3240 [Satoshi Nakamura and Andy Polyakov] 3241 3242 *) Initialize SSL_METHOD structures at compile time instead of during 3243 runtime, thus removing the need for a lock. 3244 [Steve Henson] 3245 3246 *) Make PKCS7_decrypt() work even if no certificate is supplied by 3247 attempting to decrypt each encrypted key in turn. Add support to 3248 smime utility. 3249 [Steve Henson] 3250 3251 Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 3252 3253 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 3254 OpenSSL 0.9.8.] 3255 3256 *) Add libcrypto.pc and libssl.pc for those who feel they need them. 3257 [Richard Levitte] 3258 3259 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 3260 key into the same file any more. 3261 [Richard Levitte] 3262 3263 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 3264 [Andy Polyakov] 3265 3266 *) Add -utf8 command line and config file option to 'ca'. 3267 [Stefan <stf@udoma.org] 3268 3269 *) Removed the macro des_crypt(), as it seems to conflict with some 3270 libraries. Use DES_crypt(). 3271 [Richard Levitte] 3272 3273 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 3274 involves renaming the source and generated shared-libs for 3275 both. The engines will accept the corrected or legacy ids 3276 ('ncipher' and '4758_cca' respectively) when binding. NB, 3277 this only applies when building 'shared'. 3278 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 3279 3280 *) Add attribute functions to EVP_PKEY structure. Modify 3281 PKCS12_create() to recognize a CSP name attribute and 3282 use it. Make -CSP option work again in pkcs12 utility. 3283 [Steve Henson] 3284 3285 *) Add new functionality to the bn blinding code: 3286 - automatic re-creation of the BN_BLINDING parameters after 3287 a fixed number of uses (currently 32) 3288 - add new function for parameter creation 3289 - introduce flags to control the update behaviour of the 3290 BN_BLINDING parameters 3291 - hide BN_BLINDING structure 3292 Add a second BN_BLINDING slot to the RSA structure to improve 3293 performance when a single RSA object is shared among several 3294 threads. 3295 [Nils Larsch] 3296 3297 *) Add support for DTLS. 3298 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 3299 3300 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 3301 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 3302 [Walter Goulet] 3303 3304 *) Remove buggy and incompletet DH cert support from 3305 ssl/ssl_rsa.c and ssl/s3_both.c 3306 [Nils Larsch] 3307 3308 *) Use SHA-1 instead of MD5 as the default digest algorithm for 3309 the apps/openssl applications. 3310 [Nils Larsch] 3311 3312 *) Compile clean with "-Wall -Wmissing-prototypes 3313 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 3314 DEBUG_SAFESTACK must also be set. 3315 [Ben Laurie] 3316 3317 *) Change ./Configure so that certain algorithms can be disabled by default. 3318 The new counterpiece to "no-xxx" is "enable-xxx". 3319 3320 The patented RC5 and MDC2 algorithms will now be disabled unless 3321 "enable-rc5" and "enable-mdc2", respectively, are specified. 3322 3323 (IDEA remains enabled despite being patented. This is because IDEA 3324 is frequently required for interoperability, and there is no license 3325 fee for non-commercial use. As before, "no-idea" can be used to 3326 avoid this algorithm.) 3327 3328 [Bodo Moeller] 3329 3330 *) Add processing of proxy certificates (see RFC 3820). This work was 3331 sponsored by KTH (The Royal Institute of Technology in Stockholm) and 3332 EGEE (Enabling Grids for E-science in Europe). 3333 [Richard Levitte] 3334 3335 *) RC4 performance overhaul on modern architectures/implementations, such 3336 as Intel P4, IA-64 and AMD64. 3337 [Andy Polyakov] 3338 3339 *) New utility extract-section.pl. This can be used specify an alternative 3340 section number in a pod file instead of having to treat each file as 3341 a separate case in Makefile. This can be done by adding two lines to the 3342 pod file: 3343 3344 =for comment openssl_section:XXX 3345 3346 The blank line is mandatory. 3347 3348 [Steve Henson] 3349 3350 *) New arguments -certform, -keyform and -pass for s_client and s_server 3351 to allow alternative format key and certificate files and passphrase 3352 sources. 3353 [Steve Henson] 3354 3355 *) New structure X509_VERIFY_PARAM which combines current verify parameters, 3356 update associated structures and add various utility functions. 3357 3358 Add new policy related verify parameters, include policy checking in 3359 standard verify code. Enhance 'smime' application with extra parameters 3360 to support policy checking and print out. 3361 [Steve Henson] 3362 3363 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 3364 Nehemiah processors. These extensions support AES encryption in hardware 3365 as well as RNG (though RNG support is currently disabled). 3366 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 3367 3368 *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 3369 [Geoff Thorpe] 3370 3371 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 3372 [Andy Polyakov and a number of other people] 3373 3374 *) Improved PowerPC platform support. Most notably BIGNUM assembler 3375 implementation contributed by IBM. 3376 [Suresh Chari, Peter Waltenberg, Andy Polyakov] 3377 3378 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 3379 exponent rather than 'unsigned long'. There is a corresponding change to 3380 the new 'rsa_keygen' element of the RSA_METHOD structure. 3381 [Jelte Jansen, Geoff Thorpe] 3382 3383 *) Functionality for creating the initial serial number file is now 3384 moved from CA.pl to the 'ca' utility with a new option -create_serial. 3385 3386 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 3387 number file to 1, which is bound to cause problems. To avoid 3388 the problems while respecting compatibility between different 0.9.7 3389 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 3390 CA.pl for serial number initialization. With the new release 0.9.8, 3391 we can fix the problem directly in the 'ca' utility.) 3392 [Steve Henson] 3393 3394 *) Reduced header interdepencies by declaring more opaque objects in 3395 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 3396 give fewer recursive includes, which could break lazy source code - so 3397 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 3398 developers should define this symbol when building and using openssl to 3399 ensure they track the recommended behaviour, interfaces, [etc], but 3400 backwards-compatible behaviour prevails when this isn't defined. 3401 [Geoff Thorpe] 3402 3403 *) New function X509_POLICY_NODE_print() which prints out policy nodes. 3404 [Steve Henson] 3405 3406 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 3407 This will generate a random key of the appropriate length based on the 3408 cipher context. The EVP_CIPHER can provide its own random key generation 3409 routine to support keys of a specific form. This is used in the des and 3410 3des routines to generate a key of the correct parity. Update S/MIME 3411 code to use new functions and hence generate correct parity DES keys. 3412 Add EVP_CHECK_DES_KEY #define to return an error if the key is not 3413 valid (weak or incorrect parity). 3414 [Steve Henson] 3415 3416 *) Add a local set of CRLs that can be used by X509_verify_cert() as well 3417 as looking them up. This is useful when the verified structure may contain 3418 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 3419 present unless the new PKCS7_NO_CRL flag is asserted. 3420 [Steve Henson] 3421 3422 *) Extend ASN1 oid configuration module. It now additionally accepts the 3423 syntax: 3424 3425 shortName = some long name, 1.2.3.4 3426 [Steve Henson] 3427 3428 *) Reimplemented the BN_CTX implementation. There is now no more static 3429 limitation on the number of variables it can handle nor the depth of the 3430 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 3431 information can now expand as required, and rather than having a single 3432 static array of bignums, BN_CTX now uses a linked-list of such arrays 3433 allowing it to expand on demand whilst maintaining the usefulness of 3434 BN_CTX's "bundling". 3435 [Geoff Thorpe] 3436 3437 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 3438 to allow all RSA operations to function using a single BN_CTX. 3439 [Geoff Thorpe] 3440 3441 *) Preliminary support for certificate policy evaluation and checking. This 3442 is initially intended to pass the tests outlined in "Conformance Testing 3443 of Relying Party Client Certificate Path Processing Logic" v1.07. 3444 [Steve Henson] 3445 3446 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 3447 remained unused and not that useful. A variety of other little bignum 3448 tweaks and fixes have also been made continuing on from the audit (see 3449 below). 3450 [Geoff Thorpe] 3451 3452 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 3453 associated ASN1, EVP and SSL functions and old ASN1 macros. 3454 [Richard Levitte] 3455 3456 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 3457 and this should never fail. So the return value from the use of 3458 BN_set_word() (which can fail due to needless expansion) is now deprecated; 3459 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 3460 [Geoff Thorpe] 3461 3462 *) BN_CTX_get() should return zero-valued bignums, providing the same 3463 initialised value as BN_new(). 3464 [Geoff Thorpe, suggested by Ulf M��ller] 3465 3466 *) Support for inhibitAnyPolicy certificate extension. 3467 [Steve Henson] 3468 3469 *) An audit of the BIGNUM code is underway, for which debugging code is 3470 enabled when BN_DEBUG is defined. This makes stricter enforcements on what 3471 is considered valid when processing BIGNUMs, and causes execution to 3472 assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 3473 further steps are taken to deliberately pollute unused data in BIGNUM 3474 structures to try and expose faulty code further on. For now, openssl will 3475 (in its default mode of operation) continue to tolerate the inconsistent 3476 forms that it has tolerated in the past, but authors and packagers should 3477 consider trying openssl and their own applications when compiled with 3478 these debugging symbols defined. It will help highlight potential bugs in 3479 their own code, and will improve the test coverage for OpenSSL itself. At 3480 some point, these tighter rules will become openssl's default to improve 3481 maintainability, though the assert()s and other overheads will remain only 3482 in debugging configurations. See bn.h for more details. 3483 [Geoff Thorpe, Nils Larsch, Ulf M��ller] 3484 3485 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 3486 that can only be obtained through BN_CTX_new() (which implicitly 3487 initialises it). The presence of this function only made it possible 3488 to overwrite an existing structure (and cause memory leaks). 3489 [Geoff Thorpe] 3490 3491 *) Because of the callback-based approach for implementing LHASH as a 3492 template type, lh_insert() adds opaque objects to hash-tables and 3493 lh_doall() or lh_doall_arg() are typically used with a destructor callback 3494 to clean up those corresponding objects before destroying the hash table 3495 (and losing the object pointers). So some over-zealous constifications in 3496 LHASH have been relaxed so that lh_insert() does not take (nor store) the 3497 objects as "const" and the lh_doall[_arg] callback wrappers are not 3498 prototyped to have "const" restrictions on the object pointers they are 3499 given (and so aren't required to cast them away any more). 3500 [Geoff Thorpe] 3501 3502 *) The tmdiff.h API was so ugly and minimal that our own timing utility 3503 (speed) prefers to use its own implementation. The two implementations 3504 haven't been consolidated as yet (volunteers?) but the tmdiff API has had 3505 its object type properly exposed (MS_TM) instead of casting to/from "char 3506 *". This may still change yet if someone realises MS_TM and "ms_time_***" 3507 aren't necessarily the greatest nomenclatures - but this is what was used 3508 internally to the implementation so I've used that for now. 3509 [Geoff Thorpe] 3510 3511 *) Ensure that deprecated functions do not get compiled when 3512 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 3513 the self-tests were still using deprecated key-generation functions so 3514 these have been updated also. 3515 [Geoff Thorpe] 3516 3517 *) Reorganise PKCS#7 code to separate the digest location functionality 3518 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 3519 New function PKCS7_set_digest() to set the digest type for PKCS#7 3520 digestedData type. Add additional code to correctly generate the 3521 digestedData type and add support for this type in PKCS7 initialization 3522 functions. 3523 [Steve Henson] 3524 3525 *) New function PKCS7_set0_type_other() this initializes a PKCS7 3526 structure of type "other". 3527 [Steve Henson] 3528 3529 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 3530 sure the loop does correctly stop and breaking ("division by zero") 3531 modulus operations are not performed. The (pre-generated) prime 3532 table crypto/bn/bn_prime.h was already correct, but it could not be 3533 re-generated on some platforms because of the "division by zero" 3534 situation in the script. 3535 [Ralf S. Engelschall] 3536 3537 *) Update support for ECC-based TLS ciphersuites according to 3538 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 3539 SHA-1 now is only used for "small" curves (where the 3540 representation of a field element takes up to 24 bytes); for 3541 larger curves, the field element resulting from ECDH is directly 3542 used as premaster secret. 3543 [Douglas Stebila (Sun Microsystems Laboratories)] 3544 3545 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 3546 curve secp160r1 to the tests. 3547 [Douglas Stebila (Sun Microsystems Laboratories)] 3548 3549 *) Add the possibility to load symbols globally with DSO. 3550 [G��tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 3551 3552 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 3553 control of the error stack. 3554 [Richard Levitte] 3555 3556 *) Add support for STORE in ENGINE. 3557 [Richard Levitte] 3558 3559 *) Add the STORE type. The intention is to provide a common interface 3560 to certificate and key stores, be they simple file-based stores, or 3561 HSM-type store, or LDAP stores, or... 3562 NOTE: The code is currently UNTESTED and isn't really used anywhere. 3563 [Richard Levitte] 3564 3565 *) Add a generic structure called OPENSSL_ITEM. This can be used to 3566 pass a list of arguments to any function as well as provide a way 3567 for a function to pass data back to the caller. 3568 [Richard Levitte] 3569 3570 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 3571 works like BUF_strdup() but can be used to duplicate a portion of 3572 a string. The copy gets NUL-terminated. BUF_memdup() duplicates 3573 a memory area. 3574 [Richard Levitte] 3575 3576 *) Add the function sk_find_ex() which works like sk_find(), but will 3577 return an index to an element even if an exact match couldn't be 3578 found. The index is guaranteed to point at the element where the 3579 searched-for key would be inserted to preserve sorting order. 3580 [Richard Levitte] 3581 3582 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 3583 takes an extra flags argument for optional functionality. Currently, 3584 the following flags are defined: 3585 3586 OBJ_BSEARCH_VALUE_ON_NOMATCH 3587 This one gets OBJ_bsearch_ex() to return a pointer to the first 3588 element where the comparing function returns a negative or zero 3589 number. 3590 3591 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 3592 This one gets OBJ_bsearch_ex() to return a pointer to the first 3593 element where the comparing function returns zero. This is useful 3594 if there are more than one element where the comparing function 3595 returns zero. 3596 [Richard Levitte] 3597 3598 *) Make it possible to create self-signed certificates with 'openssl ca' 3599 in such a way that the self-signed certificate becomes part of the 3600 CA database and uses the same mechanisms for serial number generation 3601 as all other certificate signing. The new flag '-selfsign' enables 3602 this functionality. Adapt CA.sh and CA.pl.in. 3603 [Richard Levitte] 3604 3605 *) Add functionality to check the public key of a certificate request 3606 against a given private. This is useful to check that a certificate 3607 request can be signed by that key (self-signing). 3608 [Richard Levitte] 3609 3610 *) Make it possible to have multiple active certificates with the same 3611 subject in the CA index file. This is done only if the keyword 3612 'unique_subject' is set to 'no' in the main CA section (default 3613 if 'CA_default') of the configuration file. The value is saved 3614 with the database itself in a separate index attribute file, 3615 named like the index file with '.attr' appended to the name. 3616 [Richard Levitte] 3617 3618 *) Generate muti valued AVAs using '+' notation in config files for 3619 req and dirName. 3620 [Steve Henson] 3621 3622 *) Support for nameConstraints certificate extension. 3623 [Steve Henson] 3624 3625 *) Support for policyConstraints certificate extension. 3626 [Steve Henson] 3627 3628 *) Support for policyMappings certificate extension. 3629 [Steve Henson] 3630 3631 *) Make sure the default DSA_METHOD implementation only uses its 3632 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 3633 and change its own handlers to be NULL so as to remove unnecessary 3634 indirection. This lets alternative implementations fallback to the 3635 default implementation more easily. 3636 [Geoff Thorpe] 3637 3638 *) Support for directoryName in GeneralName related extensions 3639 in config files. 3640 [Steve Henson] 3641 3642 *) Make it possible to link applications using Makefile.shared. 3643 Make that possible even when linking against static libraries! 3644 [Richard Levitte] 3645 3646 *) Support for single pass processing for S/MIME signing. This now 3647 means that S/MIME signing can be done from a pipe, in addition 3648 cleartext signing (multipart/signed type) is effectively streaming 3649 and the signed data does not need to be all held in memory. 3650 3651 This is done with a new flag PKCS7_STREAM. When this flag is set 3652 PKCS7_sign() only initializes the PKCS7 structure and the actual signing 3653 is done after the data is output (and digests calculated) in 3654 SMIME_write_PKCS7(). 3655 [Steve Henson] 3656 3657 *) Add full support for -rpath/-R, both in shared libraries and 3658 applications, at least on the platforms where it's known how 3659 to do it. 3660 [Richard Levitte] 3661 3662 *) In crypto/ec/ec_mult.c, implement fast point multiplication with 3663 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 3664 will now compute a table of multiples of the generator that 3665 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 3666 faster (notably in the case of a single point multiplication, 3667 scalar * generator). 3668 [Nils Larsch, Bodo Moeller] 3669 3670 *) IPv6 support for certificate extensions. The various extensions 3671 which use the IP:a.b.c.d can now take IPv6 addresses using the 3672 formats of RFC1884 2.2 . IPv6 addresses are now also displayed 3673 correctly. 3674 [Steve Henson] 3675 3676 *) Added an ENGINE that implements RSA by performing private key 3677 exponentiations with the GMP library. The conversions to and from 3678 GMP's mpz_t format aren't optimised nor are any montgomery forms 3679 cached, and on x86 it appears OpenSSL's own performance has caught up. 3680 However there are likely to be other architectures where GMP could 3681 provide a boost. This ENGINE is not built in by default, but it can be 3682 specified at Configure time and should be accompanied by the necessary 3683 linker additions, eg; 3684 ./config -DOPENSSL_USE_GMP -lgmp 3685 [Geoff Thorpe] 3686 3687 *) "openssl engine" will not display ENGINE/DSO load failure errors when 3688 testing availability of engines with "-t" - the old behaviour is 3689 produced by increasing the feature's verbosity with "-tt". 3690 [Geoff Thorpe] 3691 3692 *) ECDSA routines: under certain error conditions uninitialized BN objects 3693 could be freed. Solution: make sure initialization is performed early 3694 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 3695 via PR#459) 3696 [Lutz Jaenicke] 3697 3698 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 3699 and DH_METHOD (eg. by ENGINE implementations) to override the normal 3700 software implementations. For DSA and DH, parameter generation can 3701 also be overriden by providing the appropriate method callbacks. 3702 [Geoff Thorpe] 3703 3704 *) Change the "progress" mechanism used in key-generation and 3705 primality testing to functions that take a new BN_GENCB pointer in 3706 place of callback/argument pairs. The new API functions have "_ex" 3707 postfixes and the older functions are reimplemented as wrappers for 3708 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 3709 declarations of the old functions to help (graceful) attempts to 3710 migrate to the new functions. Also, the new key-generation API 3711 functions operate on a caller-supplied key-structure and return 3712 success/failure rather than returning a key or NULL - this is to 3713 help make "keygen" another member function of RSA_METHOD etc. 3714 3715 Example for using the new callback interface: 3716 3717 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 3718 void *my_arg = ...; 3719 BN_GENCB my_cb; 3720 3721 BN_GENCB_set(&my_cb, my_callback, my_arg); 3722 3723 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 3724 /* For the meaning of a, b in calls to my_callback(), see the 3725 * documentation of the function that calls the callback. 3726 * cb will point to my_cb; my_arg can be retrieved as cb->arg. 3727 * my_callback should return 1 if it wants BN_is_prime_ex() 3728 * to continue, or 0 to stop. 3729 */ 3730 3731 [Geoff Thorpe] 3732 3733 *) Change the ZLIB compression method to be stateful, and make it 3734 available to TLS with the number defined in 3735 draft-ietf-tls-compression-04.txt. 3736 [Richard Levitte] 3737 3738 *) Add the ASN.1 structures and functions for CertificatePair, which 3739 is defined as follows (according to X.509_4thEditionDraftV6.pdf): 3740 3741 CertificatePair ::= SEQUENCE { 3742 forward [0] Certificate OPTIONAL, 3743 reverse [1] Certificate OPTIONAL, 3744 -- at least one of the pair shall be present -- } 3745 3746 Also implement the PEM functions to read and write certificate 3747 pairs, and defined the PEM tag as "CERTIFICATE PAIR". 3748 3749 This needed to be defined, mostly for the sake of the LDAP 3750 attribute crossCertificatePair, but may prove useful elsewhere as 3751 well. 3752 [Richard Levitte] 3753 3754 *) Make it possible to inhibit symlinking of shared libraries in 3755 Makefile.shared, for Cygwin's sake. 3756 [Richard Levitte] 3757 3758 *) Extend the BIGNUM API by creating a function 3759 void BN_set_negative(BIGNUM *a, int neg); 3760 and a macro that behave like 3761 int BN_is_negative(const BIGNUM *a); 3762 3763 to avoid the need to access 'a->neg' directly in applications. 3764 [Nils Larsch] 3765 3766 *) Implement fast modular reduction for pseudo-Mersenne primes 3767 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 3768 EC_GROUP_new_curve_GFp() will now automatically use this 3769 if applicable. 3770 [Nils Larsch <nla@trustcenter.de>] 3771 3772 *) Add new lock type (CRYPTO_LOCK_BN). 3773 [Bodo Moeller] 3774 3775 *) Change the ENGINE framework to automatically load engines 3776 dynamically from specific directories unless they could be 3777 found to already be built in or loaded. Move all the 3778 current engines except for the cryptodev one to a new 3779 directory engines/. 3780 The engines in engines/ are built as shared libraries if 3781 the "shared" options was given to ./Configure or ./config. 3782 Otherwise, they are inserted in libcrypto.a. 3783 /usr/local/ssl/engines is the default directory for dynamic 3784 engines, but that can be overriden at configure time through 3785 the usual use of --prefix and/or --openssldir, and at run 3786 time with the environment variable OPENSSL_ENGINES. 3787 [Geoff Thorpe and Richard Levitte] 3788 3789 *) Add Makefile.shared, a helper makefile to build shared 3790 libraries. Addapt Makefile.org. 3791 [Richard Levitte] 3792 3793 *) Add version info to Win32 DLLs. 3794 [Peter 'Luna' Runestig" <peter@runestig.com>] 3795 3796 *) Add new 'medium level' PKCS#12 API. Certificates and keys 3797 can be added using this API to created arbitrary PKCS#12 3798 files while avoiding the low level API. 3799 3800 New options to PKCS12_create(), key or cert can be NULL and 3801 will then be omitted from the output file. The encryption 3802 algorithm NIDs can be set to -1 for no encryption, the mac 3803 iteration count can be set to 0 to omit the mac. 3804 3805 Enhance pkcs12 utility by making the -nokeys and -nocerts 3806 options work when creating a PKCS#12 file. New option -nomac 3807 to omit the mac, NONE can be set for an encryption algorithm. 3808 New code is modified to use the enhanced PKCS12_create() 3809 instead of the low level API. 3810 [Steve Henson] 3811 3812 *) Extend ASN1 encoder to support indefinite length constructed 3813 encoding. This can output sequences tags and octet strings in 3814 this form. Modify pk7_asn1.c to support indefinite length 3815 encoding. This is experimental and needs additional code to 3816 be useful, such as an ASN1 bio and some enhanced streaming 3817 PKCS#7 code. 3818 3819 Extend template encode functionality so that tagging is passed 3820 down to the template encoder. 3821 [Steve Henson] 3822 3823 *) Let 'openssl req' fail if an argument to '-newkey' is not 3824 recognized instead of using RSA as a default. 3825 [Bodo Moeller] 3826 3827 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 3828 As these are not official, they are not included in "ALL"; 3829 the "ECCdraft" ciphersuite group alias can be used to select them. 3830 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 3831 3832 *) Add ECDH engine support. 3833 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 3834 3835 *) Add ECDH in new directory crypto/ecdh/. 3836 [Douglas Stebila (Sun Microsystems Laboratories)] 3837 3838 *) Let BN_rand_range() abort with an error after 100 iterations 3839 without success (which indicates a broken PRNG). 3840 [Bodo Moeller] 3841 3842 *) Change BN_mod_sqrt() so that it verifies that the input value 3843 is really the square of the return value. (Previously, 3844 BN_mod_sqrt would show GIGO behaviour.) 3845 [Bodo Moeller] 3846 3847 *) Add named elliptic curves over binary fields from X9.62, SECG, 3848 and WAP/WTLS; add OIDs that were still missing. 3849 3850 [Sheueling Chang Shantz and Douglas Stebila 3851 (Sun Microsystems Laboratories)] 3852 3853 *) Extend the EC library for elliptic curves over binary fields 3854 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 3855 New EC_METHOD: 3856 3857 EC_GF2m_simple_method 3858 3859 New API functions: 3860 3861 EC_GROUP_new_curve_GF2m 3862 EC_GROUP_set_curve_GF2m 3863 EC_GROUP_get_curve_GF2m 3864 EC_POINT_set_affine_coordinates_GF2m 3865 EC_POINT_get_affine_coordinates_GF2m 3866 EC_POINT_set_compressed_coordinates_GF2m 3867 3868 Point compression for binary fields is disabled by default for 3869 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 3870 enable it). 3871 3872 As binary polynomials are represented as BIGNUMs, various members 3873 of the EC_GROUP and EC_POINT data structures can be shared 3874 between the implementations for prime fields and binary fields; 3875 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 3876 are essentially identical to their ..._GFp counterparts. 3877 (For simplicity, the '..._GFp' prefix has been dropped from 3878 various internal method names.) 3879 3880 An internal 'field_div' method (similar to 'field_mul' and 3881 'field_sqr') has been added; this is used only for binary fields. 3882 3883 [Sheueling Chang Shantz and Douglas Stebila 3884 (Sun Microsystems Laboratories)] 3885 3886 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 3887 through methods ('mul', 'precompute_mult'). 3888 3889 The generic implementations (now internally called 'ec_wNAF_mul' 3890 and 'ec_wNAF_precomputed_mult') remain the default if these 3891 methods are undefined. 3892 3893 [Sheueling Chang Shantz and Douglas Stebila 3894 (Sun Microsystems Laboratories)] 3895 3896 *) New function EC_GROUP_get_degree, which is defined through 3897 EC_METHOD. For curves over prime fields, this returns the bit 3898 length of the modulus. 3899 3900 [Sheueling Chang Shantz and Douglas Stebila 3901 (Sun Microsystems Laboratories)] 3902 3903 *) New functions EC_GROUP_dup, EC_POINT_dup. 3904 (These simply call ..._new and ..._copy). 3905 3906 [Sheueling Chang Shantz and Douglas Stebila 3907 (Sun Microsystems Laboratories)] 3908 3909 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 3910 Polynomials are represented as BIGNUMs (where the sign bit is not 3911 used) in the following functions [macros]: 3912 3913 BN_GF2m_add 3914 BN_GF2m_sub [= BN_GF2m_add] 3915 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 3916 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 3917 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 3918 BN_GF2m_mod_inv 3919 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 3920 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 3921 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 3922 BN_GF2m_cmp [= BN_ucmp] 3923 3924 (Note that only the 'mod' functions are actually for fields GF(2^m). 3925 BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 3926 3927 For some functions, an the irreducible polynomial defining a 3928 field can be given as an 'unsigned int[]' with strictly 3929 decreasing elements giving the indices of those bits that are set; 3930 i.e., p[] represents the polynomial 3931 f(t) = t^p[0] + t^p[1] + ... + t^p[k] 3932 where 3933 p[0] > p[1] > ... > p[k] = 0. 3934 This applies to the following functions: 3935 3936 BN_GF2m_mod_arr 3937 BN_GF2m_mod_mul_arr 3938 BN_GF2m_mod_sqr_arr 3939 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 3940 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 3941 BN_GF2m_mod_exp_arr 3942 BN_GF2m_mod_sqrt_arr 3943 BN_GF2m_mod_solve_quad_arr 3944 BN_GF2m_poly2arr 3945 BN_GF2m_arr2poly 3946 3947 Conversion can be performed by the following functions: 3948 3949 BN_GF2m_poly2arr 3950 BN_GF2m_arr2poly 3951 3952 bntest.c has additional tests for binary polynomial arithmetic. 3953 3954 Two implementations for BN_GF2m_mod_div() are available. 3955 The default algorithm simply uses BN_GF2m_mod_inv() and 3956 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 3957 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 3958 copyright notice in crypto/bn/bn_gf2m.c before enabling it). 3959 3960 [Sheueling Chang Shantz and Douglas Stebila 3961 (Sun Microsystems Laboratories)] 3962 3963 *) Add new error code 'ERR_R_DISABLED' that can be used when some 3964 functionality is disabled at compile-time. 3965 [Douglas Stebila <douglas.stebila@sun.com>] 3966 3967 *) Change default behaviour of 'openssl asn1parse' so that more 3968 information is visible when viewing, e.g., a certificate: 3969 3970 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 3971 mode the content of non-printable OCTET STRINGs is output in a 3972 style similar to INTEGERs, but with '[HEX DUMP]' prepended to 3973 avoid the appearance of a printable string. 3974 [Nils Larsch <nla@trustcenter.de>] 3975 3976 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 3977 functions 3978 EC_GROUP_set_asn1_flag() 3979 EC_GROUP_get_asn1_flag() 3980 EC_GROUP_set_point_conversion_form() 3981 EC_GROUP_get_point_conversion_form() 3982 These control ASN1 encoding details: 3983 - Curves (i.e., groups) are encoded explicitly unless asn1_flag 3984 has been set to OPENSSL_EC_NAMED_CURVE. 3985 - Points are encoded in uncompressed form by default; options for 3986 asn1_for are as for point2oct, namely 3987 POINT_CONVERSION_COMPRESSED 3988 POINT_CONVERSION_UNCOMPRESSED 3989 POINT_CONVERSION_HYBRID 3990 3991 Also add 'seed' and 'seed_len' members to EC_GROUP with access 3992 functions 3993 EC_GROUP_set_seed() 3994 EC_GROUP_get0_seed() 3995 EC_GROUP_get_seed_len() 3996 This is used only for ASN1 purposes (so far). 3997 [Nils Larsch <nla@trustcenter.de>] 3998 3999 *) Add 'field_type' member to EC_METHOD, which holds the NID 4000 of the appropriate field type OID. The new function 4001 EC_METHOD_get_field_type() returns this value. 4002 [Nils Larsch <nla@trustcenter.de>] 4003 4004 *) Add functions 4005 EC_POINT_point2bn() 4006 EC_POINT_bn2point() 4007 EC_POINT_point2hex() 4008 EC_POINT_hex2point() 4009 providing useful interfaces to EC_POINT_point2oct() and 4010 EC_POINT_oct2point(). 4011 [Nils Larsch <nla@trustcenter.de>] 4012 4013 *) Change internals of the EC library so that the functions 4014 EC_GROUP_set_generator() 4015 EC_GROUP_get_generator() 4016 EC_GROUP_get_order() 4017 EC_GROUP_get_cofactor() 4018 are implemented directly in crypto/ec/ec_lib.c and not dispatched 4019 to methods, which would lead to unnecessary code duplication when 4020 adding different types of curves. 4021 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 4022 4023 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 4024 arithmetic, and such that modified wNAFs are generated 4025 (which avoid length expansion in many cases). 4026 [Bodo Moeller] 4027 4028 *) Add a function EC_GROUP_check_discriminant() (defined via 4029 EC_METHOD) that verifies that the curve discriminant is non-zero. 4030 4031 Add a function EC_GROUP_check() that makes some sanity tests 4032 on a EC_GROUP, its generator and order. This includes 4033 EC_GROUP_check_discriminant(). 4034 [Nils Larsch <nla@trustcenter.de>] 4035 4036 *) Add ECDSA in new directory crypto/ecdsa/. 4037 4038 Add applications 'openssl ecparam' and 'openssl ecdsa' 4039 (these are based on 'openssl dsaparam' and 'openssl dsa'). 4040 4041 ECDSA support is also included in various other files across the 4042 library. Most notably, 4043 - 'openssl req' now has a '-newkey ecdsa:file' option; 4044 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 4045 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 4046 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 4047 them suitable for ECDSA where domain parameters must be 4048 extracted before the specific public key; 4049 - ECDSA engine support has been added. 4050 [Nils Larsch <nla@trustcenter.de>] 4051 4052 *) Include some named elliptic curves, and add OIDs from X9.62, 4053 SECG, and WAP/WTLS. Each curve can be obtained from the new 4054 function 4055 EC_GROUP_new_by_curve_name(), 4056 and the list of available named curves can be obtained with 4057 EC_get_builtin_curves(). 4058 Also add a 'curve_name' member to EC_GROUP objects, which can be 4059 accessed via 4060 EC_GROUP_set_curve_name() 4061 EC_GROUP_get_curve_name() 4062 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 4063 4064 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 4065 was actually never needed) and in BN_mul(). The removal in BN_mul() 4066 required a small change in bn_mul_part_recursive() and the addition 4067 of the functions bn_cmp_part_words(), bn_sub_part_words() and 4068 bn_add_part_words(), which do the same thing as bn_cmp_words(), 4069 bn_sub_words() and bn_add_words() except they take arrays with 4070 differing sizes. 4071 [Richard Levitte] 4072 4073 Changes between 0.9.7l and 0.9.7m [23 Feb 2007] 4074 4075 *) Cleanse PEM buffers before freeing them since they may contain 4076 sensitive data. 4077 [Benjamin Bennett <ben@psc.edu>] 4078 4079 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 4080 a ciphersuite string such as "DEFAULT:RSA" cannot enable 4081 authentication-only ciphersuites. 4082 [Bodo Moeller] 4083 4084 *) Since AES128 and AES256 share a single mask bit in the logic of 4085 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 4086 kludge to work properly if AES128 is available and AES256 isn't. 4087 [Victor Duchovni] 4088 4089 *) Expand security boundary to match 1.1.1 module. 4090 [Steve Henson] 4091 4092 *) Remove redundant features: hash file source, editing of test vectors 4093 modify fipsld to use external fips_premain.c signature. 4094 [Steve Henson] 4095 4096 *) New perl script mkfipsscr.pl to create shell scripts or batch files to 4097 run algorithm test programs. 4098 [Steve Henson] 4099 4100 *) Make algorithm test programs more tolerant of whitespace. 4101 [Steve Henson] 4102 4103 *) Have SSL/TLS server implementation tolerate "mismatched" record 4104 protocol version while receiving ClientHello even if the 4105 ClientHello is fragmented. (The server can't insist on the 4106 particular protocol version it has chosen before the ServerHello 4107 message has informed the client about his choice.) 4108 [Bodo Moeller] 4109 4110 *) Load error codes if they are not already present instead of using a 4111 static variable. This allows them to be cleanly unloaded and reloaded. 4112 [Steve Henson] 4113 4114 Changes between 0.9.7k and 0.9.7l [28 Sep 2006] 4115 4116 *) Introduce limits to prevent malicious keys being able to 4117 cause a denial of service. (CVE-2006-2940) 4118 [Steve Henson, Bodo Moeller] 4119 4120 *) Fix ASN.1 parsing of certain invalid structures that can result 4121 in a denial of service. (CVE-2006-2937) [Steve Henson] 4122 4123 *) Fix buffer overflow in SSL_get_shared_ciphers() function. 4124 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 4125 4126 *) Fix SSL client code which could crash if connecting to a 4127 malicious SSLv2 server. (CVE-2006-4343) 4128 [Tavis Ormandy and Will Drewry, Google Security Team] 4129 4130 *) Change ciphersuite string processing so that an explicit 4131 ciphersuite selects this one ciphersuite (so that "AES256-SHA" 4132 will no longer include "AES128-SHA"), and any other similar 4133 ciphersuite (same bitmap) from *other* protocol versions (so that 4134 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 4135 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 4136 changes from 0.9.8b and 0.9.8d. 4137 [Bodo Moeller] 4138 4139 Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 4140 4141 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 4142 (CVE-2006-4339) [Ben Laurie and Google Security Team] 4143 4144 *) Change the Unix randomness entropy gathering to use poll() when 4145 possible instead of select(), since the latter has some 4146 undesirable limitations. 4147 [Darryl Miles via Richard Levitte and Bodo Moeller] 4148 4149 *) Disable rogue ciphersuites: 4150 4151 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 4152 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 4153 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 4154 4155 The latter two were purportedly from 4156 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 4157 appear there. 4158 4159 Also deactive the remaining ciphersuites from 4160 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 4161 unofficial, and the ID has long expired. 4162 [Bodo Moeller] 4163 4164 *) Fix RSA blinding Heisenbug (problems sometimes occured on 4165 dual-core machines) and other potential thread-safety issues. 4166 [Bodo Moeller] 4167 4168 Changes between 0.9.7i and 0.9.7j [04 May 2006] 4169 4170 *) Adapt fipsld and the build system to link against the validated FIPS 4171 module in FIPS mode. 4172 [Steve Henson] 4173 4174 *) Fixes for VC++ 2005 build under Windows. 4175 [Steve Henson] 4176 4177 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 4178 from a Windows bash shell such as MSYS. It is autodetected from the 4179 "config" script when run from a VC++ environment. Modify standard VC++ 4180 build to use fipscanister.o from the GNU make build. 4181 [Steve Henson] 4182 4183 Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 4184 4185 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 4186 The value now differs depending on if you build for FIPS or not. 4187 BEWARE! A program linked with a shared FIPSed libcrypto can't be 4188 safely run with a non-FIPSed libcrypto, as it may crash because of 4189 the difference induced by this change. 4190 [Andy Polyakov] 4191 4192 Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 4193 4194 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 4195 (part of SSL_OP_ALL). This option used to disable the 4196 countermeasure against man-in-the-middle protocol-version 4197 rollback in the SSL 2.0 server implementation, which is a bad 4198 idea. (CVE-2005-2969) 4199 4200 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 4201 for Information Security, National Institute of Advanced Industrial 4202 Science and Technology [AIST], Japan)] 4203 4204 *) Minimal support for X9.31 signatures and PSS padding modes. This is 4205 mainly for FIPS compliance and not fully integrated at this stage. 4206 [Steve Henson] 4207 4208 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 4209 the exponentiation using a fixed-length exponent. (Otherwise, 4210 the information leaked through timing could expose the secret key 4211 after many signatures; cf. Bleichenbacher's attack on DSA with 4212 biased k.) 4213 [Bodo Moeller] 4214 4215 *) Make a new fixed-window mod_exp implementation the default for 4216 RSA, DSA, and DH private-key operations so that the sequence of 4217 squares and multiplies and the memory access pattern are 4218 independent of the particular secret key. This will mitigate 4219 cache-timing and potential related attacks. 4220 4221 BN_mod_exp_mont_consttime() is the new exponentiation implementation, 4222 and this is automatically used by BN_mod_exp_mont() if the new flag 4223 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 4224 will use this BN flag for private exponents unless the flag 4225 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 4226 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 4227 4228 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 4229 4230 *) Change the client implementation for SSLv23_method() and 4231 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 4232 Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 4233 (Previously, the SSL 2.0 backwards compatible Client Hello 4234 message format would be used even with SSL_OP_NO_SSLv2.) 4235 [Bodo Moeller] 4236 4237 *) Add support for smime-type MIME parameter in S/MIME messages which some 4238 clients need. 4239 [Steve Henson] 4240 4241 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 4242 a threadsafe manner. Modify rsa code to use new function and add calls 4243 to dsa and dh code (which had race conditions before). 4244 [Steve Henson] 4245 4246 *) Include the fixed error library code in the C error file definitions 4247 instead of fixing them up at runtime. This keeps the error code 4248 structures constant. 4249 [Steve Henson] 4250 4251 Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 4252 4253 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 4254 OpenSSL 0.9.8.] 4255 4256 *) Fixes for newer kerberos headers. NB: the casts are needed because 4257 the 'length' field is signed on one version and unsigned on another 4258 with no (?) obvious way to tell the difference, without these VC++ 4259 complains. Also the "definition" of FAR (blank) is no longer included 4260 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 4261 some needed definitions. 4262 [Steve Henson] 4263 4264 *) Undo Cygwin change. 4265 [Ulf M��ller] 4266 4267 *) Added support for proxy certificates according to RFC 3820. 4268 Because they may be a security thread to unaware applications, 4269 they must be explicitely allowed in run-time. See 4270 docs/HOWTO/proxy_certificates.txt for further information. 4271 [Richard Levitte] 4272 4273 Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 4274 4275 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 4276 server and client random values. Previously 4277 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 4278 less random data when sizeof(time_t) > 4 (some 64 bit platforms). 4279 4280 This change has negligible security impact because: 4281 4282 1. Server and client random values still have 24 bytes of pseudo random 4283 data. 4284 4285 2. Server and client random values are sent in the clear in the initial 4286 handshake. 4287 4288 3. The master secret is derived using the premaster secret (48 bytes in 4289 size for static RSA ciphersuites) as well as client server and random 4290 values. 4291 4292 The OpenSSL team would like to thank the UK NISCC for bringing this issue 4293 to our attention. 4294 4295 [Stephen Henson, reported by UK NISCC] 4296 4297 *) Use Windows randomness collection on Cygwin. 4298 [Ulf M��ller] 4299 4300 *) Fix hang in EGD/PRNGD query when communication socket is closed 4301 prematurely by EGD/PRNGD. 4302 [Darren Tucker <dtucker@zip.com.au> via Lutz J��nicke, resolves #1014] 4303 4304 *) Prompt for pass phrases when appropriate for PKCS12 input format. 4305 [Steve Henson] 4306 4307 *) Back-port of selected performance improvements from development 4308 branch, as well as improved support for PowerPC platforms. 4309 [Andy Polyakov] 4310 4311 *) Add lots of checks for memory allocation failure, error codes to indicate 4312 failure and freeing up memory if a failure occurs. 4313 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 4314 4315 *) Add new -passin argument to dgst. 4316 [Steve Henson] 4317 4318 *) Perform some character comparisons of different types in X509_NAME_cmp: 4319 this is needed for some certificates that reencode DNs into UTF8Strings 4320 (in violation of RFC3280) and can't or wont issue name rollover 4321 certificates. 4322 [Steve Henson] 4323 4324 *) Make an explicit check during certificate validation to see that 4325 the CA setting in each certificate on the chain is correct. As a 4326 side effect always do the following basic checks on extensions, 4327 not just when there's an associated purpose to the check: 4328 4329 - if there is an unhandled critical extension (unless the user 4330 has chosen to ignore this fault) 4331 - if the path length has been exceeded (if one is set at all) 4332 - that certain extensions fit the associated purpose (if one has 4333 been given) 4334 [Richard Levitte] 4335 4336 Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 4337 4338 *) Avoid a race condition when CRLs are checked in a multi threaded 4339 environment. This would happen due to the reordering of the revoked 4340 entries during signature checking and serial number lookup. Now the 4341 encoding is cached and the serial number sort performed under a lock. 4342 Add new STACK function sk_is_sorted(). 4343 [Steve Henson] 4344 4345 *) Add Delta CRL to the extension code. 4346 [Steve Henson] 4347 4348 *) Various fixes to s3_pkt.c so alerts are sent properly. 4349 [David Holmes <d.holmes@f5.com>] 4350 4351 *) Reduce the chances of duplicate issuer name and serial numbers (in 4352 violation of RFC3280) using the OpenSSL certificate creation utilities. 4353 This is done by creating a random 64 bit value for the initial serial 4354 number when a serial number file is created or when a self signed 4355 certificate is created using 'openssl req -x509'. The initial serial 4356 number file is created using 'openssl x509 -next_serial' in CA.pl 4357 rather than being initialized to 1. 4358 [Steve Henson] 4359 4360 Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 4361 4362 *) Fix null-pointer assignment in do_change_cipher_spec() revealed 4363 by using the Codenomicon TLS Test Tool (CVE-2004-0079) 4364 [Joe Orton, Steve Henson] 4365 4366 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 4367 (CVE-2004-0112) 4368 [Joe Orton, Steve Henson] 4369 4370 *) Make it possible to have multiple active certificates with the same 4371 subject in the CA index file. This is done only if the keyword 4372 'unique_subject' is set to 'no' in the main CA section (default 4373 if 'CA_default') of the configuration file. The value is saved 4374 with the database itself in a separate index attribute file, 4375 named like the index file with '.attr' appended to the name. 4376 [Richard Levitte] 4377 4378 *) X509 verify fixes. Disable broken certificate workarounds when 4379 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 4380 keyUsage extension present. Don't accept CRLs with unhandled critical 4381 extensions: since verify currently doesn't process CRL extensions this 4382 rejects a CRL with *any* critical extensions. Add new verify error codes 4383 for these cases. 4384 [Steve Henson] 4385 4386 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 4387 A clarification of RFC2560 will require the use of OCTET STRINGs and 4388 some implementations cannot handle the current raw format. Since OpenSSL 4389 copies and compares OCSP nonces as opaque blobs without any attempt at 4390 parsing them this should not create any compatibility issues. 4391 [Steve Henson] 4392 4393 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 4394 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 4395 this HMAC (and other) operations are several times slower than OpenSSL 4396 < 0.9.7. 4397 [Steve Henson] 4398 4399 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 4400 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 4401 4402 *) Use the correct content when signing type "other". 4403 [Steve Henson] 4404 4405 Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 4406 4407 *) Fix various bugs revealed by running the NISCC test suite: 4408 4409 Stop out of bounds reads in the ASN1 code when presented with 4410 invalid tags (CVE-2003-0543 and CVE-2003-0544). 4411 4412 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 4413 4414 If verify callback ignores invalid public key errors don't try to check 4415 certificate signature with the NULL public key. 4416 4417 [Steve Henson] 4418 4419 *) New -ignore_err option in ocsp application to stop the server 4420 exiting on the first error in a request. 4421 [Steve Henson] 4422 4423 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 4424 if the server requested one: as stated in TLS 1.0 and SSL 3.0 4425 specifications. 4426 [Steve Henson] 4427 4428 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 4429 extra data after the compression methods not only for TLS 1.0 4430 but also for SSL 3.0 (as required by the specification). 4431 [Bodo Moeller; problem pointed out by Matthias Loepfe] 4432 4433 *) Change X509_certificate_type() to mark the key as exported/exportable 4434 when it's 512 *bits* long, not 512 bytes. 4435 [Richard Levitte] 4436 4437 *) Change AES_cbc_encrypt() so it outputs exact multiple of 4438 blocks during encryption. 4439 [Richard Levitte] 4440 4441 *) Various fixes to base64 BIO and non blocking I/O. On write 4442 flushes were not handled properly if the BIO retried. On read 4443 data was not being buffered properly and had various logic bugs. 4444 This also affects blocking I/O when the data being decoded is a 4445 certain size. 4446 [Steve Henson] 4447 4448 *) Various S/MIME bugfixes and compatibility changes: 4449 output correct application/pkcs7 MIME type if 4450 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 4451 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 4452 of files as .eml work). Correctly handle very long lines in MIME 4453 parser. 4454 [Steve Henson] 4455 4456 Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 4457 4458 *) Countermeasure against the Klima-Pokorny-Rosa extension of 4459 Bleichbacher's attack on PKCS #1 v1.5 padding: treat 4460 a protocol version number mismatch like a decryption error 4461 in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 4462 [Bodo Moeller] 4463 4464 *) Turn on RSA blinding by default in the default implementation 4465 to avoid a timing attack. Applications that don't want it can call 4466 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 4467 They would be ill-advised to do so in most cases. 4468 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 4469 4470 *) Change RSA blinding code so that it works when the PRNG is not 4471 seeded (in this case, the secret RSA exponent is abused as 4472 an unpredictable seed -- if it is not unpredictable, there 4473 is no point in blinding anyway). Make RSA blinding thread-safe 4474 by remembering the creator's thread ID in rsa->blinding and 4475 having all other threads use local one-time blinding factors 4476 (this requires more computation than sharing rsa->blinding, but 4477 avoids excessive locking; and if an RSA object is not shared 4478 between threads, blinding will still be very fast). 4479 [Bodo Moeller] 4480 4481 *) Fixed a typo bug that would cause ENGINE_set_default() to set an 4482 ENGINE as defaults for all supported algorithms irrespective of 4483 the 'flags' parameter. 'flags' is now honoured, so applications 4484 should make sure they are passing it correctly. 4485 [Geoff Thorpe] 4486 4487 *) Target "mingw" now allows native Windows code to be generated in 4488 the Cygwin environment as well as with the MinGW compiler. 4489 [Ulf Moeller] 4490 4491 Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 4492 4493 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 4494 via timing by performing a MAC computation even if incorrrect 4495 block cipher padding has been found. This is a countermeasure 4496 against active attacks where the attacker has to distinguish 4497 between bad padding and a MAC verification error. (CVE-2003-0078) 4498 4499 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 4500 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 4501 Martin Vuagnoux (EPFL, Ilion)] 4502 4503 *) Make the no-err option work as intended. The intention with no-err 4504 is not to have the whole error stack handling routines removed from 4505 libcrypto, it's only intended to remove all the function name and 4506 reason texts, thereby removing some of the footprint that may not 4507 be interesting if those errors aren't displayed anyway. 4508 4509 NOTE: it's still possible for any application or module to have it's 4510 own set of error texts inserted. The routines are there, just not 4511 used by default when no-err is given. 4512 [Richard Levitte] 4513 4514 *) Add support for FreeBSD on IA64. 4515 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 4516 4517 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 4518 Kerberos function mit_des_cbc_cksum(). Before this change, 4519 the value returned by DES_cbc_cksum() was like the one from 4520 mit_des_cbc_cksum(), except the bytes were swapped. 4521 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 4522 4523 *) Allow an application to disable the automatic SSL chain building. 4524 Before this a rather primitive chain build was always performed in 4525 ssl3_output_cert_chain(): an application had no way to send the 4526 correct chain if the automatic operation produced an incorrect result. 4527 4528 Now the chain builder is disabled if either: 4529 4530 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 4531 4532 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 4533 4534 The reasoning behind this is that an application would not want the 4535 auto chain building to take place if extra chain certificates are 4536 present and it might also want a means of sending no additional 4537 certificates (for example the chain has two certificates and the 4538 root is omitted). 4539 [Steve Henson] 4540 4541 *) Add the possibility to build without the ENGINE framework. 4542 [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 4543 4544 *) Under Win32 gmtime() can return NULL: check return value in 4545 OPENSSL_gmtime(). Add error code for case where gmtime() fails. 4546 [Steve Henson] 4547 4548 *) DSA routines: under certain error conditions uninitialized BN objects 4549 could be freed. Solution: make sure initialization is performed early 4550 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 4551 Nils Larsch <nla@trustcenter.de> via PR#459) 4552 [Lutz Jaenicke] 4553 4554 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 4555 checked on reconnect on the client side, therefore session resumption 4556 could still fail with a "ssl session id is different" error. This 4557 behaviour is masked when SSL_OP_ALL is used due to 4558 SSL_OP_MICROSOFT_SESS_ID_BUG being set. 4559 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4560 followup to PR #377. 4561 [Lutz Jaenicke] 4562 4563 *) IA-32 assembler support enhancements: unified ELF targets, support 4564 for SCO/Caldera platforms, fix for Cygwin shared build. 4565 [Andy Polyakov] 4566 4567 *) Add support for FreeBSD on sparc64. As a consequence, support for 4568 FreeBSD on non-x86 processors is separate from x86 processors on 4569 the config script, much like the NetBSD support. 4570 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 4571 4572 Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 4573 4574 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 4575 OpenSSL 0.9.7.] 4576 4577 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 4578 code (06) was taken as the first octet of the session ID and the last 4579 octet was ignored consequently. As a result SSLv2 client side session 4580 caching could not have worked due to the session ID mismatch between 4581 client and server. 4582 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4583 PR #377. 4584 [Lutz Jaenicke] 4585 4586 *) Change the declaration of needed Kerberos libraries to use EX_LIBS 4587 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 4588 removed entirely. 4589 [Richard Levitte] 4590 4591 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 4592 seems that in spite of existing for more than a year, many application 4593 author have done nothing to provide the necessary callbacks, which 4594 means that this particular engine will not work properly anywhere. 4595 This is a very unfortunate situation which forces us, in the name 4596 of usability, to give the hw_ncipher.c a static lock, which is part 4597 of libcrypto. 4598 NOTE: This is for the 0.9.7 series ONLY. This hack will never 4599 appear in 0.9.8 or later. We EXPECT application authors to have 4600 dealt properly with this when 0.9.8 is released (unless we actually 4601 make such changes in the libcrypto locking code that changes will 4602 have to be made anyway). 4603 [Richard Levitte] 4604 4605 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 4606 octets have been read, EOF or an error occurs. Without this change 4607 some truncated ASN1 structures will not produce an error. 4608 [Steve Henson] 4609 4610 *) Disable Heimdal support, since it hasn't been fully implemented. 4611 Still give the possibility to force the use of Heimdal, but with 4612 warnings and a request that patches get sent to openssl-dev. 4613 [Richard Levitte] 4614 4615 *) Add the VC-CE target, introduce the WINCE sysname, and add 4616 INSTALL.WCE and appropriate conditionals to make it build. 4617 [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 4618 4619 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 4620 cygssl-x.y.z.dll, where x, y and z are the major, minor and 4621 edit numbers of the version. 4622 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 4623 4624 *) Introduce safe string copy and catenation functions 4625 (BUF_strlcpy() and BUF_strlcat()). 4626 [Ben Laurie (CHATS) and Richard Levitte] 4627 4628 *) Avoid using fixed-size buffers for one-line DNs. 4629 [Ben Laurie (CHATS)] 4630 4631 *) Add BUF_MEM_grow_clean() to avoid information leakage when 4632 resizing buffers containing secrets, and use where appropriate. 4633 [Ben Laurie (CHATS)] 4634 4635 *) Avoid using fixed size buffers for configuration file location. 4636 [Ben Laurie (CHATS)] 4637 4638 *) Avoid filename truncation for various CA files. 4639 [Ben Laurie (CHATS)] 4640 4641 *) Use sizeof in preference to magic numbers. 4642 [Ben Laurie (CHATS)] 4643 4644 *) Avoid filename truncation in cert requests. 4645 [Ben Laurie (CHATS)] 4646 4647 *) Add assertions to check for (supposedly impossible) buffer 4648 overflows. 4649 [Ben Laurie (CHATS)] 4650 4651 *) Don't cache truncated DNS entries in the local cache (this could 4652 potentially lead to a spoofing attack). 4653 [Ben Laurie (CHATS)] 4654 4655 *) Fix various buffers to be large enough for hex/decimal 4656 representations in a platform independent manner. 4657 [Ben Laurie (CHATS)] 4658 4659 *) Add CRYPTO_realloc_clean() to avoid information leakage when 4660 resizing buffers containing secrets, and use where appropriate. 4661 [Ben Laurie (CHATS)] 4662 4663 *) Add BIO_indent() to avoid much slightly worrying code to do 4664 indents. 4665 [Ben Laurie (CHATS)] 4666 4667 *) Convert sprintf()/BIO_puts() to BIO_printf(). 4668 [Ben Laurie (CHATS)] 4669 4670 *) buffer_gets() could terminate with the buffer only half 4671 full. Fixed. 4672 [Ben Laurie (CHATS)] 4673 4674 *) Add assertions to prevent user-supplied crypto functions from 4675 overflowing internal buffers by having large block sizes, etc. 4676 [Ben Laurie (CHATS)] 4677 4678 *) New OPENSSL_assert() macro (similar to assert(), but enabled 4679 unconditionally). 4680 [Ben Laurie (CHATS)] 4681 4682 *) Eliminate unused copy of key in RC4. 4683 [Ben Laurie (CHATS)] 4684 4685 *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 4686 [Ben Laurie (CHATS)] 4687 4688 *) Fix off-by-one error in EGD path. 4689 [Ben Laurie (CHATS)] 4690 4691 *) If RANDFILE path is too long, ignore instead of truncating. 4692 [Ben Laurie (CHATS)] 4693 4694 *) Eliminate unused and incorrectly sized X.509 structure 4695 CBCParameter. 4696 [Ben Laurie (CHATS)] 4697 4698 *) Eliminate unused and dangerous function knumber(). 4699 [Ben Laurie (CHATS)] 4700 4701 *) Eliminate unused and dangerous structure, KSSL_ERR. 4702 [Ben Laurie (CHATS)] 4703 4704 *) Protect against overlong session ID context length in an encoded 4705 session object. Since these are local, this does not appear to be 4706 exploitable. 4707 [Ben Laurie (CHATS)] 4708 4709 *) Change from security patch (see 0.9.6e below) that did not affect 4710 the 0.9.6 release series: 4711 4712 Remote buffer overflow in SSL3 protocol - an attacker could 4713 supply an oversized master key in Kerberos-enabled versions. 4714 (CVE-2002-0657) 4715 [Ben Laurie (CHATS)] 4716 4717 *) Change the SSL kerb5 codes to match RFC 2712. 4718 [Richard Levitte] 4719 4720 *) Make -nameopt work fully for req and add -reqopt switch. 4721 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 4722 4723 *) The "block size" for block ciphers in CFB and OFB mode should be 1. 4724 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 4725 4726 *) Make sure tests can be performed even if the corresponding algorithms 4727 have been removed entirely. This was also the last step to make 4728 OpenSSL compilable with DJGPP under all reasonable conditions. 4729 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 4730 4731 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 4732 to allow version independent disabling of normally unselected ciphers, 4733 which may be activated as a side-effect of selecting a single cipher. 4734 4735 (E.g., cipher list string "RSA" enables ciphersuites that are left 4736 out of "ALL" because they do not provide symmetric encryption. 4737 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 4738 [Lutz Jaenicke, Bodo Moeller] 4739 4740 *) Add appropriate support for separate platform-dependent build 4741 directories. The recommended way to make a platform-dependent 4742 build directory is the following (tested on Linux), maybe with 4743 some local tweaks: 4744 4745 # Place yourself outside of the OpenSSL source tree. In 4746 # this example, the environment variable OPENSSL_SOURCE 4747 # is assumed to contain the absolute OpenSSL source directory. 4748 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 4749 cd objtree/"`uname -s`-`uname -r`-`uname -m`" 4750 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 4751 mkdir -p `dirname $F` 4752 ln -s $OPENSSL_SOURCE/$F $F 4753 done 4754 4755 To be absolutely sure not to disturb the source tree, a "make clean" 4756 is a good thing. If it isn't successfull, don't worry about it, 4757 it probably means the source directory is very clean. 4758 [Richard Levitte] 4759 4760 *) Make sure any ENGINE control commands make local copies of string 4761 pointers passed to them whenever necessary. Otherwise it is possible 4762 the caller may have overwritten (or deallocated) the original string 4763 data when a later ENGINE operation tries to use the stored values. 4764 [G��tz Babin-Ebell <babinebell@trustcenter.de>] 4765 4766 *) Improve diagnostics in file reading and command-line digests. 4767 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 4768 4769 *) Add AES modes CFB and OFB to the object database. Correct an 4770 error in AES-CFB decryption. 4771 [Richard Levitte] 4772 4773 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 4774 allows existing EVP_CIPHER_CTX structures to be reused after 4775 calling EVP_*Final(). This behaviour is used by encryption 4776 BIOs and some applications. This has the side effect that 4777 applications must explicitly clean up cipher contexts with 4778 EVP_CIPHER_CTX_cleanup() or they will leak memory. 4779 [Steve Henson] 4780 4781 *) Check the values of dna and dnb in bn_mul_recursive before calling 4782 bn_mul_comba (a non zero value means the a or b arrays do not contain 4783 n2 elements) and fallback to bn_mul_normal if either is not zero. 4784 [Steve Henson] 4785 4786 *) Fix escaping of non-ASCII characters when using the -subj option 4787 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 4788 [Lutz Jaenicke] 4789 4790 *) Make object definitions compliant to LDAP (RFC2256): SN is the short 4791 form for "surname", serialNumber has no short form. 4792 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 4793 therefore remove "mail" short name for "internet 7". 4794 The OID for unique identifiers in X509 certificates is 4795 x500UniqueIdentifier, not uniqueIdentifier. 4796 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 4797 [Lutz Jaenicke] 4798 4799 *) Add an "init" command to the ENGINE config module and auto initialize 4800 ENGINEs. Without any "init" command the ENGINE will be initialized 4801 after all ctrl commands have been executed on it. If init=1 the 4802 ENGINE is initailized at that point (ctrls before that point are run 4803 on the uninitialized ENGINE and after on the initialized one). If 4804 init=0 then the ENGINE will not be iniatialized at all. 4805 [Steve Henson] 4806 4807 *) Fix the 'app_verify_callback' interface so that the user-defined 4808 argument is actually passed to the callback: In the 4809 SSL_CTX_set_cert_verify_callback() prototype, the callback 4810 declaration has been changed from 4811 int (*cb)() 4812 into 4813 int (*cb)(X509_STORE_CTX *,void *); 4814 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 4815 i=s->ctx->app_verify_callback(&ctx) 4816 has been changed into 4817 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 4818 4819 To update applications using SSL_CTX_set_cert_verify_callback(), 4820 a dummy argument can be added to their callback functions. 4821 [D. K. Smetters <smetters@parc.xerox.com>] 4822 4823 *) Added the '4758cca' ENGINE to support IBM 4758 cards. 4824 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 4825 4826 *) Add and OPENSSL_LOAD_CONF define which will cause 4827 OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 4828 This allows older applications to transparently support certain 4829 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 4830 Two new functions OPENSSL_add_all_algorithms_noconf() which will never 4831 load the config file and OPENSSL_add_all_algorithms_conf() which will 4832 always load it have also been added. 4833 [Steve Henson] 4834 4835 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 4836 Adjust NIDs and EVP layer. 4837 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4838 4839 *) Config modules support in openssl utility. 4840 4841 Most commands now load modules from the config file, 4842 though in a few (such as version) this isn't done 4843 because it couldn't be used for anything. 4844 4845 In the case of ca and req the config file used is 4846 the same as the utility itself: that is the -config 4847 command line option can be used to specify an 4848 alternative file. 4849 [Steve Henson] 4850 4851 *) Move default behaviour from OPENSSL_config(). If appname is NULL 4852 use "openssl_conf" if filename is NULL use default openssl config file. 4853 [Steve Henson] 4854 4855 *) Add an argument to OPENSSL_config() to allow the use of an alternative 4856 config section name. Add a new flag to tolerate a missing config file 4857 and move code to CONF_modules_load_file(). 4858 [Steve Henson] 4859 4860 *) Support for crypto accelerator cards from Accelerated Encryption 4861 Processing, www.aep.ie. (Use engine 'aep') 4862 The support was copied from 0.9.6c [engine] and adapted/corrected 4863 to work with the new engine framework. 4864 [AEP Inc. and Richard Levitte] 4865 4866 *) Support for SureWare crypto accelerator cards from Baltimore 4867 Technologies. (Use engine 'sureware') 4868 The support was copied from 0.9.6c [engine] and adapted 4869 to work with the new engine framework. 4870 [Richard Levitte] 4871 4872 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 4873 make the newer ENGINE framework commands for the CHIL engine work. 4874 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 4875 4876 *) Make it possible to produce shared libraries on ReliantUNIX. 4877 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 4878 4879 *) Add the configuration target debug-linux-ppro. 4880 Make 'openssl rsa' use the general key loading routines 4881 implemented in apps.c, and make those routines able to 4882 handle the key format FORMAT_NETSCAPE and the variant 4883 FORMAT_IISSGC. 4884 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4885 4886 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 4887 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4888 4889 *) Add -keyform to rsautl, and document -engine. 4890 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 4891 4892 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 4893 BIO_R_NO_SUCH_FILE error code rather than the generic 4894 ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 4895 [Ben Laurie] 4896 4897 *) Add new functions 4898 ERR_peek_last_error 4899 ERR_peek_last_error_line 4900 ERR_peek_last_error_line_data. 4901 These are similar to 4902 ERR_peek_error 4903 ERR_peek_error_line 4904 ERR_peek_error_line_data, 4905 but report on the latest error recorded rather than the first one 4906 still in the error queue. 4907 [Ben Laurie, Bodo Moeller] 4908 4909 *) default_algorithms option in ENGINE config module. This allows things 4910 like: 4911 default_algorithms = ALL 4912 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 4913 [Steve Henson] 4914 4915 *) Prelminary ENGINE config module. 4916 [Steve Henson] 4917 4918 *) New experimental application configuration code. 4919 [Steve Henson] 4920 4921 *) Change the AES code to follow the same name structure as all other 4922 symmetric ciphers, and behave the same way. Move everything to 4923 the directory crypto/aes, thereby obsoleting crypto/rijndael. 4924 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4925 4926 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 4927 [Ben Laurie and Theo de Raadt] 4928 4929 *) Add option to output public keys in req command. 4930 [Massimiliano Pala madwolf@openca.org] 4931 4932 *) Use wNAFs in EC_POINTs_mul() for improved efficiency 4933 (up to about 10% better than before for P-192 and P-224). 4934 [Bodo Moeller] 4935 4936 *) New functions/macros 4937 4938 SSL_CTX_set_msg_callback(ctx, cb) 4939 SSL_CTX_set_msg_callback_arg(ctx, arg) 4940 SSL_set_msg_callback(ssl, cb) 4941 SSL_set_msg_callback_arg(ssl, arg) 4942 4943 to request calling a callback function 4944 4945 void cb(int write_p, int version, int content_type, 4946 const void *buf, size_t len, SSL *ssl, void *arg) 4947 4948 whenever a protocol message has been completely received 4949 (write_p == 0) or sent (write_p == 1). Here 'version' is the 4950 protocol version according to which the SSL library interprets 4951 the current protocol message (SSL2_VERSION, SSL3_VERSION, or 4952 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 4953 the content type as defined in the SSL 3.0/TLS 1.0 protocol 4954 specification (change_cipher_spec(20), alert(21), handshake(22)). 4955 'buf' and 'len' point to the actual message, 'ssl' to the 4956 SSL object, and 'arg' is the application-defined value set by 4957 SSL[_CTX]_set_msg_callback_arg(). 4958 4959 'openssl s_client' and 'openssl s_server' have new '-msg' options 4960 to enable a callback that displays all protocol messages. 4961 [Bodo Moeller] 4962 4963 *) Change the shared library support so shared libraries are built as 4964 soon as the corresponding static library is finished, and thereby get 4965 openssl and the test programs linked against the shared library. 4966 This still only happens when the keyword "shard" has been given to 4967 the configuration scripts. 4968 4969 NOTE: shared library support is still an experimental thing, and 4970 backward binary compatibility is still not guaranteed. 4971 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 4972 4973 *) Add support for Subject Information Access extension. 4974 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 4975 4976 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 4977 additional bytes when new memory had to be allocated, not just 4978 when reusing an existing buffer. 4979 [Bodo Moeller] 4980 4981 *) New command line and configuration option 'utf8' for the req command. 4982 This allows field values to be specified as UTF8 strings. 4983 [Steve Henson] 4984 4985 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 4986 runs for the former and machine-readable output for the latter. 4987 [Ben Laurie] 4988 4989 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 4990 of the e-mail address in the DN (i.e., it will go into a certificate 4991 extension only). The new configuration file option 'email_in_dn = no' 4992 has the same effect. 4993 [Massimiliano Pala madwolf@openca.org] 4994 4995 *) Change all functions with names starting with des_ to be starting 4996 with DES_ instead. Add wrappers that are compatible with libdes, 4997 but are named _ossl_old_des_*. Finally, add macros that map the 4998 des_* symbols to the corresponding _ossl_old_des_* if libdes 4999 compatibility is desired. If OpenSSL 0.9.6c compatibility is 5000 desired, the des_* symbols will be mapped to DES_*, with one 5001 exception. 5002 5003 Since we provide two compatibility mappings, the user needs to 5004 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 5005 compatibility is desired. The default (i.e., when that macro 5006 isn't defined) is OpenSSL 0.9.6c compatibility. 5007 5008 There are also macros that enable and disable the support of old 5009 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 5010 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 5011 are defined, the default will apply: to support the old des routines. 5012 5013 In either case, one must include openssl/des.h to get the correct 5014 definitions. Do not try to just include openssl/des_old.h, that 5015 won't work. 5016 5017 NOTE: This is a major break of an old API into a new one. Software 5018 authors are encouraged to switch to the DES_ style functions. Some 5019 time in the future, des_old.h and the libdes compatibility functions 5020 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 5021 default), and then completely removed. 5022 [Richard Levitte] 5023 5024 *) Test for certificates which contain unsupported critical extensions. 5025 If such a certificate is found during a verify operation it is 5026 rejected by default: this behaviour can be overridden by either 5027 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 5028 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 5029 X509_supported_extension() has also been added which returns 1 if a 5030 particular extension is supported. 5031 [Steve Henson] 5032 5033 *) Modify the behaviour of EVP cipher functions in similar way to digests 5034 to retain compatibility with existing code. 5035 [Steve Henson] 5036 5037 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 5038 compatibility with existing code. In particular the 'ctx' parameter does 5039 not have to be to be initialized before the call to EVP_DigestInit() and 5040 it is tidied up after a call to EVP_DigestFinal(). New function 5041 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 5042 EVP_MD_CTX_copy() changed to not require the destination to be 5043 initialized valid and new function EVP_MD_CTX_copy_ex() added which 5044 requires the destination to be valid. 5045 5046 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 5047 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 5048 [Steve Henson] 5049 5050 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 5051 so that complete 'Handshake' protocol structures are kept in memory 5052 instead of overwriting 'msg_type' and 'length' with 'body' data. 5053 [Bodo Moeller] 5054 5055 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 5056 [Massimo Santin via Richard Levitte] 5057 5058 *) Major restructuring to the underlying ENGINE code. This includes 5059 reduction of linker bloat, separation of pure "ENGINE" manipulation 5060 (initialisation, etc) from functionality dealing with implementations 5061 of specific crypto iterfaces. This change also introduces integrated 5062 support for symmetric ciphers and digest implementations - so ENGINEs 5063 can now accelerate these by providing EVP_CIPHER and EVP_MD 5064 implementations of their own. This is detailed in crypto/engine/README 5065 as it couldn't be adequately described here. However, there are a few 5066 API changes worth noting - some RSA, DSA, DH, and RAND functions that 5067 were changed in the original introduction of ENGINE code have now 5068 reverted back - the hooking from this code to ENGINE is now a good 5069 deal more passive and at run-time, operations deal directly with 5070 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 5071 dereferencing through an ENGINE pointer any more. Also, the ENGINE 5072 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 5073 they were not being used by the framework as there is no concept of a 5074 BIGNUM_METHOD and they could not be generalised to the new 5075 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 5076 ENGINE_cpy() has been removed as it cannot be consistently defined in 5077 the new code. 5078 [Geoff Thorpe] 5079 5080 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 5081 [Steve Henson] 5082 5083 *) Change mkdef.pl to sort symbols that get the same entry number, 5084 and make sure the automatically generated functions ERR_load_* 5085 become part of libeay.num as well. 5086 [Richard Levitte] 5087 5088 *) New function SSL_renegotiate_pending(). This returns true once 5089 renegotiation has been requested (either SSL_renegotiate() call 5090 or HelloRequest/ClientHello receveived from the peer) and becomes 5091 false once a handshake has been completed. 5092 (For servers, SSL_renegotiate() followed by SSL_do_handshake() 5093 sends a HelloRequest, but does not ensure that a handshake takes 5094 place. SSL_renegotiate_pending() is useful for checking if the 5095 client has followed the request.) 5096 [Bodo Moeller] 5097 5098 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 5099 By default, clients may request session resumption even during 5100 renegotiation (if session ID contexts permit); with this option, 5101 session resumption is possible only in the first handshake. 5102 5103 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 5104 more bits available for options that should not be part of 5105 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 5106 [Bodo Moeller] 5107 5108 *) Add some demos for certificate and certificate request creation. 5109 [Steve Henson] 5110 5111 *) Make maximum certificate chain size accepted from the peer application 5112 settable (SSL*_get/set_max_cert_list()), as proposed by 5113 "Douglas E. Engert" <deengert@anl.gov>. 5114 [Lutz Jaenicke] 5115 5116 *) Add support for shared libraries for Unixware-7 5117 (Boyd Lynn Gerber <gerberb@zenez.com>). 5118 [Lutz Jaenicke] 5119 5120 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 5121 be done prior to destruction. Use this to unload error strings from 5122 ENGINEs that load their own error strings. NB: This adds two new API 5123 functions to "get" and "set" this destroy handler in an ENGINE. 5124 [Geoff Thorpe] 5125 5126 *) Alter all existing ENGINE implementations (except "openssl" and 5127 "openbsd") to dynamically instantiate their own error strings. This 5128 makes them more flexible to be built both as statically-linked ENGINEs 5129 and self-contained shared-libraries loadable via the "dynamic" ENGINE. 5130 Also, add stub code to each that makes building them as self-contained 5131 shared-libraries easier (see README.ENGINE). 5132 [Geoff Thorpe] 5133 5134 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 5135 implementations into applications that are completely implemented in 5136 self-contained shared-libraries. The "dynamic" ENGINE exposes control 5137 commands that can be used to configure what shared-library to load and 5138 to control aspects of the way it is handled. Also, made an update to 5139 the README.ENGINE file that brings its information up-to-date and 5140 provides some information and instructions on the "dynamic" ENGINE 5141 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 5142 [Geoff Thorpe] 5143 5144 *) Make it possible to unload ranges of ERR strings with a new 5145 "ERR_unload_strings" function. 5146 [Geoff Thorpe] 5147 5148 *) Add a copy() function to EVP_MD. 5149 [Ben Laurie] 5150 5151 *) Make EVP_MD routines take a context pointer instead of just the 5152 md_data void pointer. 5153 [Ben Laurie] 5154 5155 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 5156 that the digest can only process a single chunk of data 5157 (typically because it is provided by a piece of 5158 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 5159 is only going to provide a single chunk of data, and hence the 5160 framework needn't accumulate the data for oneshot drivers. 5161 [Ben Laurie] 5162 5163 *) As with "ERR", make it possible to replace the underlying "ex_data" 5164 functions. This change also alters the storage and management of global 5165 ex_data state - it's now all inside ex_data.c and all "class" code (eg. 5166 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 5167 index counters. The API functions that use this state have been changed 5168 to take a "class_index" rather than pointers to the class's local STACK 5169 and counter, and there is now an API function to dynamically create new 5170 classes. This centralisation allows us to (a) plug a lot of the 5171 thread-safety problems that existed, and (b) makes it possible to clean 5172 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 5173 such data would previously have always leaked in application code and 5174 workarounds were in place to make the memory debugging turn a blind eye 5175 to it. Application code that doesn't use this new function will still 5176 leak as before, but their memory debugging output will announce it now 5177 rather than letting it slide. 5178 5179 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 5180 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 5181 has a return value to indicate success or failure. 5182 [Geoff Thorpe] 5183 5184 *) Make it possible to replace the underlying "ERR" functions such that the 5185 global state (2 LHASH tables and 2 locks) is only used by the "default" 5186 implementation. This change also adds two functions to "get" and "set" 5187 the implementation prior to it being automatically set the first time 5188 any other ERR function takes place. Ie. an application can call "get", 5189 pass the return value to a module it has just loaded, and that module 5190 can call its own "set" function using that value. This means the 5191 module's "ERR" operations will use (and modify) the error state in the 5192 application and not in its own statically linked copy of OpenSSL code. 5193 [Geoff Thorpe] 5194 5195 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 5196 reference counts. This performs normal REF_PRINT/REF_CHECK macros on 5197 the operation, and provides a more encapsulated way for external code 5198 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 5199 to use these functions rather than manually incrementing the counts. 5200 5201 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 5202 [Geoff Thorpe] 5203 5204 *) Add EVP test program. 5205 [Ben Laurie] 5206 5207 *) Add symmetric cipher support to ENGINE. Expect the API to change! 5208 [Ben Laurie] 5209 5210 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 5211 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 5212 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 5213 These allow a CRL to be built without having to access X509_CRL fields 5214 directly. Modify 'ca' application to use new functions. 5215 [Steve Henson] 5216 5217 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 5218 bug workarounds. Rollback attack detection is a security feature. 5219 The problem will only arise on OpenSSL servers when TLSv1 is not 5220 available (sslv3_server_method() or SSL_OP_NO_TLSv1). 5221 Software authors not wanting to support TLSv1 will have special reasons 5222 for their choice and can explicitly enable this option. 5223 [Bodo Moeller, Lutz Jaenicke] 5224 5225 *) Rationalise EVP so it can be extended: don't include a union of 5226 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 5227 (similar to those existing for EVP_CIPHER_CTX). 5228 Usage example: 5229 5230 EVP_MD_CTX md; 5231 5232 EVP_MD_CTX_init(&md); /* new function call */ 5233 EVP_DigestInit(&md, EVP_sha1()); 5234 EVP_DigestUpdate(&md, in, len); 5235 EVP_DigestFinal(&md, out, NULL); 5236 EVP_MD_CTX_cleanup(&md); /* new function call */ 5237 5238 [Ben Laurie] 5239 5240 *) Make DES key schedule conform to the usual scheme, as well as 5241 correcting its structure. This means that calls to DES functions 5242 now have to pass a pointer to a des_key_schedule instead of a 5243 plain des_key_schedule (which was actually always a pointer 5244 anyway): E.g., 5245 5246 des_key_schedule ks; 5247 5248 des_set_key_checked(..., &ks); 5249 des_ncbc_encrypt(..., &ks, ...); 5250 5251 (Note that a later change renames 'des_...' into 'DES_...'.) 5252 [Ben Laurie] 5253 5254 *) Initial reduction of linker bloat: the use of some functions, such as 5255 PEM causes large amounts of unused functions to be linked in due to 5256 poor organisation. For example pem_all.c contains every PEM function 5257 which has a knock on effect of linking in large amounts of (unused) 5258 ASN1 code. Grouping together similar functions and splitting unrelated 5259 functions prevents this. 5260 [Steve Henson] 5261 5262 *) Cleanup of EVP macros. 5263 [Ben Laurie] 5264 5265 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 5266 correct _ecb suffix. 5267 [Ben Laurie] 5268 5269 *) Add initial OCSP responder support to ocsp application. The 5270 revocation information is handled using the text based index 5271 use by the ca application. The responder can either handle 5272 requests generated internally, supplied in files (for example 5273 via a CGI script) or using an internal minimal server. 5274 [Steve Henson] 5275 5276 *) Add configuration choices to get zlib compression for TLS. 5277 [Richard Levitte] 5278 5279 *) Changes to Kerberos SSL for RFC 2712 compliance: 5280 1. Implemented real KerberosWrapper, instead of just using 5281 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 5282 2. Implemented optional authenticator field of KerberosWrapper. 5283 5284 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 5285 and authenticator structs; see crypto/krb5/. 5286 5287 Generalized Kerberos calls to support multiple Kerberos libraries. 5288 [Vern Staats <staatsvr@asc.hpc.mil>, 5289 Jeffrey Altman <jaltman@columbia.edu> 5290 via Richard Levitte] 5291 5292 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 5293 already does with RSA. testdsa.h now has 'priv_key/pub_key' 5294 values for each of the key sizes rather than having just 5295 parameters (and 'speed' generating keys each time). 5296 [Geoff Thorpe] 5297 5298 *) Speed up EVP routines. 5299 Before: 5300encrypt 5301type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 5302des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 5303des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 5304des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 5305decrypt 5306des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 5307des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 5308des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 5309 After: 5310encrypt 5311des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 5312decrypt 5313des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 5314 [Ben Laurie] 5315 5316 *) Added the OS2-EMX target. 5317 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 5318 5319 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 5320 to support NCONF routines in extension code. New function CONF_set_nconf() 5321 to allow functions which take an NCONF to also handle the old LHASH 5322 structure: this means that the old CONF compatible routines can be 5323 retained (in particular wrt extensions) without having to duplicate the 5324 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 5325 [Steve Henson] 5326 5327 *) Enhance the general user interface with mechanisms for inner control 5328 and with possibilities to have yes/no kind of prompts. 5329 [Richard Levitte] 5330 5331 *) Change all calls to low level digest routines in the library and 5332 applications to use EVP. Add missing calls to HMAC_cleanup() and 5333 don't assume HMAC_CTX can be copied using memcpy(). 5334 [Verdon Walker <VWalker@novell.com>, Steve Henson] 5335 5336 *) Add the possibility to control engines through control names but with 5337 arbitrary arguments instead of just a string. 5338 Change the key loaders to take a UI_METHOD instead of a callback 5339 function pointer. NOTE: this breaks binary compatibility with earlier 5340 versions of OpenSSL [engine]. 5341 Adapt the nCipher code for these new conditions and add a card insertion 5342 callback. 5343 [Richard Levitte] 5344 5345 *) Enhance the general user interface with mechanisms to better support 5346 dialog box interfaces, application-defined prompts, the possibility 5347 to use defaults (for example default passwords from somewhere else) 5348 and interrupts/cancellations. 5349 [Richard Levitte] 5350 5351 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 5352 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 5353 [Steve Henson] 5354 5355 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 5356 tidy up some unnecessarily weird code in 'sk_new()'). 5357 [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 5358 5359 *) Change the key loading routines for ENGINEs to use the same kind 5360 callback (pem_password_cb) as all other routines that need this 5361 kind of callback. 5362 [Richard Levitte] 5363 5364 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 5365 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 5366 than this minimum value is recommended. 5367 [Lutz Jaenicke] 5368 5369 *) New random seeder for OpenVMS, using the system process statistics 5370 that are easily reachable. 5371 [Richard Levitte] 5372 5373 *) Windows apparently can't transparently handle global 5374 variables defined in DLLs. Initialisations such as: 5375 5376 const ASN1_ITEM *it = &ASN1_INTEGER_it; 5377 5378 wont compile. This is used by the any applications that need to 5379 declare their own ASN1 modules. This was fixed by adding the option 5380 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 5381 needed for static libraries under Win32. 5382 [Steve Henson] 5383 5384 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 5385 setting of purpose and trust fields. New X509_STORE trust and 5386 purpose functions and tidy up setting in other SSL functions. 5387 [Steve Henson] 5388 5389 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 5390 structure. These are inherited by X509_STORE_CTX when it is 5391 initialised. This allows various defaults to be set in the 5392 X509_STORE structure (such as flags for CRL checking and custom 5393 purpose or trust settings) for functions which only use X509_STORE_CTX 5394 internally such as S/MIME. 5395 5396 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 5397 trust settings if they are not set in X509_STORE. This allows X509_STORE 5398 purposes and trust (in S/MIME for example) to override any set by default. 5399 5400 Add command line options for CRL checking to smime, s_client and s_server 5401 applications. 5402 [Steve Henson] 5403 5404 *) Initial CRL based revocation checking. If the CRL checking flag(s) 5405 are set then the CRL is looked up in the X509_STORE structure and 5406 its validity and signature checked, then if the certificate is found 5407 in the CRL the verify fails with a revoked error. 5408 5409 Various new CRL related callbacks added to X509_STORE_CTX structure. 5410 5411 Command line options added to 'verify' application to support this. 5412 5413 This needs some additional work, such as being able to handle multiple 5414 CRLs with different times, extension based lookup (rather than just 5415 by subject name) and ultimately more complete V2 CRL extension 5416 handling. 5417 [Steve Henson] 5418 5419 *) Add a general user interface API (crypto/ui/). This is designed 5420 to replace things like des_read_password and friends (backward 5421 compatibility functions using this new API are provided). 5422 The purpose is to remove prompting functions from the DES code 5423 section as well as provide for prompting through dialog boxes in 5424 a window system and the like. 5425 [Richard Levitte] 5426 5427 *) Add "ex_data" support to ENGINE so implementations can add state at a 5428 per-structure level rather than having to store it globally. 5429 [Geoff] 5430 5431 *) Make it possible for ENGINE structures to be copied when retrieved by 5432 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 5433 This causes the "original" ENGINE structure to act like a template, 5434 analogous to the RSA vs. RSA_METHOD type of separation. Because of this 5435 operational state can be localised to each ENGINE structure, despite the 5436 fact they all share the same "methods". New ENGINE structures returned in 5437 this case have no functional references and the return value is the single 5438 structural reference. This matches the single structural reference returned 5439 by ENGINE_by_id() normally, when it is incremented on the pre-existing 5440 ENGINE structure. 5441 [Geoff] 5442 5443 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 5444 needs to match any other type at all we need to manually clear the 5445 tag cache. 5446 [Steve Henson] 5447 5448 *) Changes to the "openssl engine" utility to include; 5449 - verbosity levels ('-v', '-vv', and '-vvv') that provide information 5450 about an ENGINE's available control commands. 5451 - executing control commands from command line arguments using the 5452 '-pre' and '-post' switches. '-post' is only used if '-t' is 5453 specified and the ENGINE is successfully initialised. The syntax for 5454 the individual commands are colon-separated, for example; 5455 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 5456 [Geoff] 5457 5458 *) New dynamic control command support for ENGINEs. ENGINEs can now 5459 declare their own commands (numbers), names (strings), descriptions, 5460 and input types for run-time discovery by calling applications. A 5461 subset of these commands are implicitly classed as "executable" 5462 depending on their input type, and only these can be invoked through 5463 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 5464 can be based on user input, config files, etc). The distinction is 5465 that "executable" commands cannot return anything other than a boolean 5466 result and can only support numeric or string input, whereas some 5467 discoverable commands may only be for direct use through 5468 ENGINE_ctrl(), eg. supporting the exchange of binary data, function 5469 pointers, or other custom uses. The "executable" commands are to 5470 support parameterisations of ENGINE behaviour that can be 5471 unambiguously defined by ENGINEs and used consistently across any 5472 OpenSSL-based application. Commands have been added to all the 5473 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 5474 control over shared-library paths without source code alterations. 5475 [Geoff] 5476 5477 *) Changed all ENGINE implementations to dynamically allocate their 5478 ENGINEs rather than declaring them statically. Apart from this being 5479 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 5480 this also allows the implementations to compile without using the 5481 internal engine_int.h header. 5482 [Geoff] 5483 5484 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 5485 'const' value. Any code that should be able to modify a RAND_METHOD 5486 should already have non-const pointers to it (ie. they should only 5487 modify their own ones). 5488 [Geoff] 5489 5490 *) Made a variety of little tweaks to the ENGINE code. 5491 - "atalla" and "ubsec" string definitions were moved from header files 5492 to C code. "nuron" string definitions were placed in variables 5493 rather than hard-coded - allowing parameterisation of these values 5494 later on via ctrl() commands. 5495 - Removed unused "#if 0"'d code. 5496 - Fixed engine list iteration code so it uses ENGINE_free() to release 5497 structural references. 5498 - Constified the RAND_METHOD element of ENGINE structures. 5499 - Constified various get/set functions as appropriate and added 5500 missing functions (including a catch-all ENGINE_cpy that duplicates 5501 all ENGINE values onto a new ENGINE except reference counts/state). 5502 - Removed NULL parameter checks in get/set functions. Setting a method 5503 or function to NULL is a way of cancelling out a previously set 5504 value. Passing a NULL ENGINE parameter is just plain stupid anyway 5505 and doesn't justify the extra error symbols and code. 5506 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 5507 flags from engine_int.h to engine.h. 5508 - Changed prototypes for ENGINE handler functions (init(), finish(), 5509 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 5510 [Geoff] 5511 5512 *) Implement binary inversion algorithm for BN_mod_inverse in addition 5513 to the algorithm using long division. The binary algorithm can be 5514 used only if the modulus is odd. On 32-bit systems, it is faster 5515 only for relatively small moduli (roughly 20-30% for 128-bit moduli, 5516 roughly 5-15% for 256-bit moduli), so we use it only for moduli 5517 up to 450 bits. In 64-bit environments, the binary algorithm 5518 appears to be advantageous for much longer moduli; here we use it 5519 for moduli up to 2048 bits. 5520 [Bodo Moeller] 5521 5522 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 5523 could not support the combine flag in choice fields. 5524 [Steve Henson] 5525 5526 *) Add a 'copy_extensions' option to the 'ca' utility. This copies 5527 extensions from a certificate request to the certificate. 5528 [Steve Henson] 5529 5530 *) Allow multiple 'certopt' and 'nameopt' options to be separated 5531 by commas. Add 'namopt' and 'certopt' options to the 'ca' config 5532 file: this allows the display of the certificate about to be 5533 signed to be customised, to allow certain fields to be included 5534 or excluded and extension details. The old system didn't display 5535 multicharacter strings properly, omitted fields not in the policy 5536 and couldn't display additional details such as extensions. 5537 [Steve Henson] 5538 5539 *) Function EC_POINTs_mul for multiple scalar multiplication 5540 of an arbitrary number of elliptic curve points 5541 \sum scalars[i]*points[i], 5542 optionally including the generator defined for the EC_GROUP: 5543 scalar*generator + \sum scalars[i]*points[i]. 5544 5545 EC_POINT_mul is a simple wrapper function for the typical case 5546 that the point list has just one item (besides the optional 5547 generator). 5548 [Bodo Moeller] 5549 5550 *) First EC_METHODs for curves over GF(p): 5551 5552 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 5553 operations and provides various method functions that can also 5554 operate with faster implementations of modular arithmetic. 5555 5556 EC_GFp_mont_method() reuses most functions that are part of 5557 EC_GFp_simple_method, but uses Montgomery arithmetic. 5558 5559 [Bodo Moeller; point addition and point doubling 5560 implementation directly derived from source code provided by 5561 Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 5562 5563 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 5564 crypto/ec/ec_lib.c): 5565 5566 Curves are EC_GROUP objects (with an optional group generator) 5567 based on EC_METHODs that are built into the library. 5568 5569 Points are EC_POINT objects based on EC_GROUP objects. 5570 5571 Most of the framework would be able to handle curves over arbitrary 5572 finite fields, but as there are no obvious types for fields other 5573 than GF(p), some functions are limited to that for now. 5574 [Bodo Moeller] 5575 5576 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 5577 that the file contains a complete HTTP response. 5578 [Richard Levitte] 5579 5580 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 5581 change the def and num file printf format specifier from "%-40sXXX" 5582 to "%-39s XXX". The latter will always guarantee a space after the 5583 field while the former will cause them to run together if the field 5584 is 40 of more characters long. 5585 [Steve Henson] 5586 5587 *) Constify the cipher and digest 'method' functions and structures 5588 and modify related functions to take constant EVP_MD and EVP_CIPHER 5589 pointers. 5590 [Steve Henson] 5591 5592 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 5593 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 5594 [Bodo Moeller] 5595 5596 *) Modify EVP_Digest*() routines so they now return values. Although the 5597 internal software routines can never fail additional hardware versions 5598 might. 5599 [Steve Henson] 5600 5601 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 5602 5603 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 5604 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 5605 5606 ASN1 error codes 5607 ERR_R_NESTED_ASN1_ERROR 5608 ... 5609 ERR_R_MISSING_ASN1_EOS 5610 were 4 .. 9, conflicting with 5611 ERR_LIB_RSA (= ERR_R_RSA_LIB) 5612 ... 5613 ERR_LIB_PEM (= ERR_R_PEM_LIB). 5614 They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 5615 5616 Add new error code 'ERR_R_INTERNAL_ERROR'. 5617 [Bodo Moeller] 5618 5619 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 5620 suffices. 5621 [Bodo Moeller] 5622 5623 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 5624 sets the subject name for a new request or supersedes the 5625 subject name in a given request. Formats that can be parsed are 5626 'CN=Some Name, OU=myOU, C=IT' 5627 and 5628 'CN=Some Name/OU=myOU/C=IT'. 5629 5630 Add options '-batch' and '-verbose' to 'openssl req'. 5631 [Massimiliano Pala <madwolf@hackmasters.net>] 5632 5633 *) Introduce the possibility to access global variables through 5634 functions on platform were that's the best way to handle exporting 5635 global variables in shared libraries. To enable this functionality, 5636 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 5637 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 5638 is normally done by Configure or something similar). 5639 5640 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 5641 in the source file (foo.c) like this: 5642 5643 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 5644 OPENSSL_IMPLEMENT_GLOBAL(double,bar); 5645 5646 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 5647 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 5648 5649 OPENSSL_DECLARE_GLOBAL(int,foo); 5650 #define foo OPENSSL_GLOBAL_REF(foo) 5651 OPENSSL_DECLARE_GLOBAL(double,bar); 5652 #define bar OPENSSL_GLOBAL_REF(bar) 5653 5654 The #defines are very important, and therefore so is including the 5655 header file everywhere where the defined globals are used. 5656 5657 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 5658 of ASN.1 items, but that structure is a bit different. 5659 5660 The largest change is in util/mkdef.pl which has been enhanced with 5661 better and easier to understand logic to choose which symbols should 5662 go into the Windows .def files as well as a number of fixes and code 5663 cleanup (among others, algorithm keywords are now sorted 5664 lexicographically to avoid constant rewrites). 5665 [Richard Levitte] 5666 5667 *) In BN_div() keep a copy of the sign of 'num' before writing the 5668 result to 'rm' because if rm==num the value will be overwritten 5669 and produce the wrong result if 'num' is negative: this caused 5670 problems with BN_mod() and BN_nnmod(). 5671 [Steve Henson] 5672 5673 *) Function OCSP_request_verify(). This checks the signature on an 5674 OCSP request and verifies the signer certificate. The signer 5675 certificate is just checked for a generic purpose and OCSP request 5676 trust settings. 5677 [Steve Henson] 5678 5679 *) Add OCSP_check_validity() function to check the validity of OCSP 5680 responses. OCSP responses are prepared in real time and may only 5681 be a few seconds old. Simply checking that the current time lies 5682 between thisUpdate and nextUpdate max reject otherwise valid responses 5683 caused by either OCSP responder or client clock inaccuracy. Instead 5684 we allow thisUpdate and nextUpdate to fall within a certain period of 5685 the current time. The age of the response can also optionally be 5686 checked. Two new options -validity_period and -status_age added to 5687 ocsp utility. 5688 [Steve Henson] 5689 5690 *) If signature or public key algorithm is unrecognized print out its 5691 OID rather that just UNKNOWN. 5692 [Steve Henson] 5693 5694 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 5695 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 5696 ID to be generated from the issuer certificate alone which can then be 5697 passed to OCSP_id_issuer_cmp(). 5698 [Steve Henson] 5699 5700 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 5701 ASN1 modules to export functions returning ASN1_ITEM pointers 5702 instead of the ASN1_ITEM structures themselves. This adds several 5703 new macros which allow the underlying ASN1 function/structure to 5704 be accessed transparently. As a result code should not use ASN1_ITEM 5705 references directly (such as &X509_it) but instead use the relevant 5706 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 5707 use of the new ASN1 code on platforms where exporting structures 5708 is problematical (for example in shared libraries) but exporting 5709 functions returning pointers to structures is not. 5710 [Steve Henson] 5711 5712 *) Add support for overriding the generation of SSL/TLS session IDs. 5713 These callbacks can be registered either in an SSL_CTX or per SSL. 5714 The purpose of this is to allow applications to control, if they wish, 5715 the arbitrary values chosen for use as session IDs, particularly as it 5716 can be useful for session caching in multiple-server environments. A 5717 command-line switch for testing this (and any client code that wishes 5718 to use such a feature) has been added to "s_server". 5719 [Geoff Thorpe, Lutz Jaenicke] 5720 5721 *) Modify mkdef.pl to recognise and parse preprocessor conditionals 5722 of the form '#if defined(...) || defined(...) || ...' and 5723 '#if !defined(...) && !defined(...) && ...'. This also avoids 5724 the growing number of special cases it was previously handling. 5725 [Richard Levitte] 5726 5727 *) Make all configuration macros available for application by making 5728 sure they are available in opensslconf.h, by giving them names starting 5729 with "OPENSSL_" to avoid conflicts with other packages and by making 5730 sure e_os2.h will cover all platform-specific cases together with 5731 opensslconf.h. 5732 Additionally, it is now possible to define configuration/platform- 5733 specific names (called "system identities"). In the C code, these 5734 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 5735 macro with the name beginning with "OPENSSL_SYS_", which is determined 5736 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 5737 what is available. 5738 [Richard Levitte] 5739 5740 *) New option -set_serial to 'req' and 'x509' this allows the serial 5741 number to use to be specified on the command line. Previously self 5742 signed certificates were hard coded with serial number 0 and the 5743 CA options of 'x509' had to use a serial number in a file which was 5744 auto incremented. 5745 [Steve Henson] 5746 5747 *) New options to 'ca' utility to support V2 CRL entry extensions. 5748 Currently CRL reason, invalidity date and hold instruction are 5749 supported. Add new CRL extensions to V3 code and some new objects. 5750 [Steve Henson] 5751 5752 *) New function EVP_CIPHER_CTX_set_padding() this is used to 5753 disable standard block padding (aka PKCS#5 padding) in the EVP 5754 API, which was previously mandatory. This means that the data is 5755 not padded in any way and so the total length much be a multiple 5756 of the block size, otherwise an error occurs. 5757 [Steve Henson] 5758 5759 *) Initial (incomplete) OCSP SSL support. 5760 [Steve Henson] 5761 5762 *) New function OCSP_parse_url(). This splits up a URL into its host, 5763 port and path components: primarily to parse OCSP URLs. New -url 5764 option to ocsp utility. 5765 [Steve Henson] 5766 5767 *) New nonce behavior. The return value of OCSP_check_nonce() now 5768 reflects the various checks performed. Applications can decide 5769 whether to tolerate certain situations such as an absent nonce 5770 in a response when one was present in a request: the ocsp application 5771 just prints out a warning. New function OCSP_add1_basic_nonce() 5772 this is to allow responders to include a nonce in a response even if 5773 the request is nonce-less. 5774 [Steve Henson] 5775 5776 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 5777 skipped when using openssl x509 multiple times on a single input file, 5778 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 5779 [Bodo Moeller] 5780 5781 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 5782 set string type: to handle setting ASN1_TIME structures. Fix ca 5783 utility to correctly initialize revocation date of CRLs. 5784 [Steve Henson] 5785 5786 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 5787 the clients preferred ciphersuites and rather use its own preferences. 5788 Should help to work around M$ SGC (Server Gated Cryptography) bug in 5789 Internet Explorer by ensuring unchanged hash method during stepup. 5790 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 5791 [Lutz Jaenicke] 5792 5793 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 5794 to aes and add a new 'exist' option to print out symbols that don't 5795 appear to exist. 5796 [Steve Henson] 5797 5798 *) Additional options to ocsp utility to allow flags to be set and 5799 additional certificates supplied. 5800 [Steve Henson] 5801 5802 *) Add the option -VAfile to 'openssl ocsp', so the user can give the 5803 OCSP client a number of certificate to only verify the response 5804 signature against. 5805 [Richard Levitte] 5806 5807 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 5808 handle the new API. Currently only ECB, CBC modes supported. Add new 5809 AES OIDs. 5810 5811 Add TLS AES ciphersuites as described in RFC3268, "Advanced 5812 Encryption Standard (AES) Ciphersuites for Transport Layer 5813 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 5814 not enabled by default and were not part of the "ALL" ciphersuite 5815 alias because they were not yet official; they could be 5816 explicitly requested by specifying the "AESdraft" ciphersuite 5817 group alias. In the final release of OpenSSL 0.9.7, the group 5818 alias is called "AES" and is part of "ALL".) 5819 [Ben Laurie, Steve Henson, Bodo Moeller] 5820 5821 *) New function OCSP_copy_nonce() to copy nonce value (if present) from 5822 request to response. 5823 [Steve Henson] 5824 5825 *) Functions for OCSP responders. OCSP_request_onereq_count(), 5826 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 5827 extract information from a certificate request. OCSP_response_create() 5828 creates a response and optionally adds a basic response structure. 5829 OCSP_basic_add1_status() adds a complete single response to a basic 5830 response and returns the OCSP_SINGLERESP structure just added (to allow 5831 extensions to be included for example). OCSP_basic_add1_cert() adds a 5832 certificate to a basic response and OCSP_basic_sign() signs a basic 5833 response with various flags. New helper functions ASN1_TIME_check() 5834 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 5835 (converts ASN1_TIME to GeneralizedTime). 5836 [Steve Henson] 5837 5838 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 5839 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 5840 structure from a certificate. X509_pubkey_digest() digests the public_key 5841 contents: this is used in various key identifiers. 5842 [Steve Henson] 5843 5844 *) Make sk_sort() tolerate a NULL argument. 5845 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 5846 5847 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 5848 passed by the function are trusted implicitly. If any of them signed the 5849 response then it is assumed to be valid and is not verified. 5850 [Steve Henson] 5851 5852 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 5853 to data. This was previously part of the PKCS7 ASN1 code. This 5854 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 5855 [Steve Henson, reported by Kenneth R. Robinette 5856 <support@securenetterm.com>] 5857 5858 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 5859 routines: without these tracing memory leaks is very painful. 5860 Fix leaks in PKCS12 and PKCS7 routines. 5861 [Steve Henson] 5862 5863 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 5864 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 5865 effectively meant GeneralizedTime would never be used. Now it 5866 is initialised to -1 but X509_time_adj() now has to check the value 5867 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 5868 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 5869 [Steve Henson, reported by Kenneth R. Robinette 5870 <support@securenetterm.com>] 5871 5872 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 5873 result in a zero length in the ASN1_INTEGER structure which was 5874 not consistent with the structure when d2i_ASN1_INTEGER() was used 5875 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 5876 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 5877 where it did not print out a minus for negative ASN1_INTEGER. 5878 [Steve Henson] 5879 5880 *) Add summary printout to ocsp utility. The various functions which 5881 convert status values to strings have been renamed to: 5882 OCSP_response_status_str(), OCSP_cert_status_str() and 5883 OCSP_crl_reason_str() and are no longer static. New options 5884 to verify nonce values and to disable verification. OCSP response 5885 printout format cleaned up. 5886 [Steve Henson] 5887 5888 *) Add additional OCSP certificate checks. These are those specified 5889 in RFC2560. This consists of two separate checks: the CA of the 5890 certificate being checked must either be the OCSP signer certificate 5891 or the issuer of the OCSP signer certificate. In the latter case the 5892 OCSP signer certificate must contain the OCSP signing extended key 5893 usage. This check is performed by attempting to match the OCSP 5894 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 5895 in the OCSP_CERTID structures of the response. 5896 [Steve Henson] 5897 5898 *) Initial OCSP certificate verification added to OCSP_basic_verify() 5899 and related routines. This uses the standard OpenSSL certificate 5900 verify routines to perform initial checks (just CA validity) and 5901 to obtain the certificate chain. Then additional checks will be 5902 performed on the chain. Currently the root CA is checked to see 5903 if it is explicitly trusted for OCSP signing. This is used to set 5904 a root CA as a global signing root: that is any certificate that 5905 chains to that CA is an acceptable OCSP signing certificate. 5906 [Steve Henson] 5907 5908 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 5909 extensions from a separate configuration file. 5910 As when reading extensions from the main configuration file, 5911 the '-extensions ...' option may be used for specifying the 5912 section to use. 5913 [Massimiliano Pala <madwolf@comune.modena.it>] 5914 5915 *) New OCSP utility. Allows OCSP requests to be generated or 5916 read. The request can be sent to a responder and the output 5917 parsed, outputed or printed in text form. Not complete yet: 5918 still needs to check the OCSP response validity. 5919 [Steve Henson] 5920 5921 *) New subcommands for 'openssl ca': 5922 'openssl ca -status <serial>' prints the status of the cert with 5923 the given serial number (according to the index file). 5924 'openssl ca -updatedb' updates the expiry status of certificates 5925 in the index file. 5926 [Massimiliano Pala <madwolf@comune.modena.it>] 5927 5928 *) New '-newreq-nodes' command option to CA.pl. This is like 5929 '-newreq', but calls 'openssl req' with the '-nodes' option 5930 so that the resulting key is not encrypted. 5931 [Damien Miller <djm@mindrot.org>] 5932 5933 *) New configuration for the GNU Hurd. 5934 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 5935 5936 *) Initial code to implement OCSP basic response verify. This 5937 is currently incomplete. Currently just finds the signer's 5938 certificate and verifies the signature on the response. 5939 [Steve Henson] 5940 5941 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 5942 value of OPENSSLDIR. This is available via the new '-d' option 5943 to 'openssl version', and is also included in 'openssl version -a'. 5944 [Bodo Moeller] 5945 5946 *) Allowing defining memory allocation callbacks that will be given 5947 file name and line number information in additional arguments 5948 (a const char* and an int). The basic functionality remains, as 5949 well as the original possibility to just replace malloc(), 5950 realloc() and free() by functions that do not know about these 5951 additional arguments. To register and find out the current 5952 settings for extended allocation functions, the following 5953 functions are provided: 5954 5955 CRYPTO_set_mem_ex_functions 5956 CRYPTO_set_locked_mem_ex_functions 5957 CRYPTO_get_mem_ex_functions 5958 CRYPTO_get_locked_mem_ex_functions 5959 5960 These work the same way as CRYPTO_set_mem_functions and friends. 5961 CRYPTO_get_[locked_]mem_functions now writes 0 where such an 5962 extended allocation function is enabled. 5963 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 5964 a conventional allocation function is enabled. 5965 [Richard Levitte, Bodo Moeller] 5966 5967 *) Finish off removing the remaining LHASH function pointer casts. 5968 There should no longer be any prototype-casting required when using 5969 the LHASH abstraction, and any casts that remain are "bugs". See 5970 the callback types and macros at the head of lhash.h for details 5971 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 5972 [Geoff Thorpe] 5973 5974 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 5975 If /dev/[u]random devices are not available or do not return enough 5976 entropy, EGD style sockets (served by EGD or PRNGD) will automatically 5977 be queried. 5978 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 5979 /etc/entropy will be queried once each in this sequence, quering stops 5980 when enough entropy was collected without querying more sockets. 5981 [Lutz Jaenicke] 5982 5983 *) Change the Unix RAND_poll() variant to be able to poll several 5984 random devices, as specified by DEVRANDOM, until a sufficient amount 5985 of data has been collected. We spend at most 10 ms on each file 5986 (select timeout) and read in non-blocking mode. DEVRANDOM now 5987 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 5988 (previously it was just the string "/dev/urandom"), so on typical 5989 platforms the 10 ms delay will never occur. 5990 Also separate out the Unix variant to its own file, rand_unix.c. 5991 For VMS, there's a currently-empty rand_vms.c. 5992 [Richard Levitte] 5993 5994 *) Move OCSP client related routines to ocsp_cl.c. These 5995 provide utility functions which an application needing 5996 to issue a request to an OCSP responder and analyse the 5997 response will typically need: as opposed to those which an 5998 OCSP responder itself would need which will be added later. 5999 6000 OCSP_request_sign() signs an OCSP request with an API similar 6001 to PKCS7_sign(). OCSP_response_status() returns status of OCSP 6002 response. OCSP_response_get1_basic() extracts basic response 6003 from response. OCSP_resp_find_status(): finds and extracts status 6004 information from an OCSP_CERTID structure (which will be created 6005 when the request structure is built). These are built from lower 6006 level functions which work on OCSP_SINGLERESP structures but 6007 wont normally be used unless the application wishes to examine 6008 extensions in the OCSP response for example. 6009 6010 Replace nonce routines with a pair of functions. 6011 OCSP_request_add1_nonce() adds a nonce value and optionally 6012 generates a random value. OCSP_check_nonce() checks the 6013 validity of the nonce in an OCSP response. 6014 [Steve Henson] 6015 6016 *) Change function OCSP_request_add() to OCSP_request_add0_id(). 6017 This doesn't copy the supplied OCSP_CERTID and avoids the 6018 need to free up the newly created id. Change return type 6019 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 6020 This can then be used to add extensions to the request. 6021 Deleted OCSP_request_new(), since most of its functionality 6022 is now in OCSP_REQUEST_new() (and the case insensitive name 6023 clash) apart from the ability to set the request name which 6024 will be added elsewhere. 6025 [Steve Henson] 6026 6027 *) Update OCSP API. Remove obsolete extensions argument from 6028 various functions. Extensions are now handled using the new 6029 OCSP extension code. New simple OCSP HTTP function which 6030 can be used to send requests and parse the response. 6031 [Steve Henson] 6032 6033 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 6034 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 6035 uses the special reorder version of SET OF to sort the attributes 6036 and reorder them to match the encoded order. This resolves a long 6037 standing problem: a verify on a PKCS7 structure just after signing 6038 it used to fail because the attribute order did not match the 6039 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 6040 it uses the received order. This is necessary to tolerate some broken 6041 software that does not order SET OF. This is handled by encoding 6042 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 6043 to produce the required SET OF. 6044 [Steve Henson] 6045 6046 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 6047 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 6048 files to get correct declarations of the ASN.1 item variables. 6049 [Richard Levitte] 6050 6051 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 6052 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 6053 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 6054 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 6055 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 6056 ASN1_ITEM and no wrapper functions. 6057 [Steve Henson] 6058 6059 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 6060 replace the old function pointer based I/O routines. Change most of 6061 the *_d2i_bio() and *_d2i_fp() functions to use these. 6062 [Steve Henson] 6063 6064 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 6065 lines, recognice more "algorithms" that can be deselected, and make 6066 it complain about algorithm deselection that isn't recognised. 6067 [Richard Levitte] 6068 6069 *) New ASN1 functions to handle dup, sign, verify, digest, pack and 6070 unpack operations in terms of ASN1_ITEM. Modify existing wrappers 6071 to use new functions. Add NO_ASN1_OLD which can be set to remove 6072 some old style ASN1 functions: this can be used to determine if old 6073 code will still work when these eventually go away. 6074 [Steve Henson] 6075 6076 *) New extension functions for OCSP structures, these follow the 6077 same conventions as certificates and CRLs. 6078 [Steve Henson] 6079 6080 *) New function X509V3_add1_i2d(). This automatically encodes and 6081 adds an extension. Its behaviour can be customised with various 6082 flags to append, replace or delete. Various wrappers added for 6083 certifcates and CRLs. 6084 [Steve Henson] 6085 6086 *) Fix to avoid calling the underlying ASN1 print routine when 6087 an extension cannot be parsed. Correct a typo in the 6088 OCSP_SERVICELOC extension. Tidy up print OCSP format. 6089 [Steve Henson] 6090 6091 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 6092 entries for variables. 6093 [Steve Henson] 6094 6095 *) Add functionality to apps/openssl.c for detecting locking 6096 problems: As the program is single-threaded, all we have 6097 to do is register a locking callback using an array for 6098 storing which locks are currently held by the program. 6099 [Bodo Moeller] 6100 6101 *) Use a lock around the call to CRYPTO_get_ex_new_index() in 6102 SSL_get_ex_data_X509_STORE_idx(), which is used in 6103 ssl_verify_cert_chain() and thus can be called at any time 6104 during TLS/SSL handshakes so that thread-safety is essential. 6105 Unfortunately, the ex_data design is not at all suited 6106 for multi-threaded use, so it probably should be abolished. 6107 [Bodo Moeller] 6108 6109 *) Added Broadcom "ubsec" ENGINE to OpenSSL. 6110 [Broadcom, tweaked and integrated by Geoff Thorpe] 6111 6112 *) Move common extension printing code to new function 6113 X509V3_print_extensions(). Reorganise OCSP print routines and 6114 implement some needed OCSP ASN1 functions. Add OCSP extensions. 6115 [Steve Henson] 6116 6117 *) New function X509_signature_print() to remove duplication in some 6118 print routines. 6119 [Steve Henson] 6120 6121 *) Add a special meaning when SET OF and SEQUENCE OF flags are both 6122 set (this was treated exactly the same as SET OF previously). This 6123 is used to reorder the STACK representing the structure to match the 6124 encoding. This will be used to get round a problem where a PKCS7 6125 structure which was signed could not be verified because the STACK 6126 order did not reflect the encoded order. 6127 [Steve Henson] 6128 6129 *) Reimplement the OCSP ASN1 module using the new code. 6130 [Steve Henson] 6131 6132 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 6133 for its ASN1 operations. The old style function pointers still exist 6134 for now but they will eventually go away. 6135 [Steve Henson] 6136 6137 *) Merge in replacement ASN1 code from the ASN1 branch. This almost 6138 completely replaces the old ASN1 functionality with a table driven 6139 encoder and decoder which interprets an ASN1_ITEM structure describing 6140 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 6141 largely maintained. Almost all of the old asn1_mac.h macro based ASN1 6142 has also been converted to the new form. 6143 [Steve Henson] 6144 6145 *) Change BN_mod_exp_recp so that negative moduli are tolerated 6146 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 6147 so that BN_mod_exp_mont and BN_mod_exp_mont_word work 6148 for negative moduli. 6149 [Bodo Moeller] 6150 6151 *) Fix BN_uadd and BN_usub: Always return non-negative results instead 6152 of not touching the result's sign bit. 6153 [Bodo Moeller] 6154 6155 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 6156 set. 6157 [Bodo Moeller] 6158 6159 *) Changed the LHASH code to use prototypes for callbacks, and created 6160 macros to declare and implement thin (optionally static) functions 6161 that provide type-safety and avoid function pointer casting for the 6162 type-specific callbacks. 6163 [Geoff Thorpe] 6164 6165 *) Added Kerberos Cipher Suites to be used with TLS, as written in 6166 RFC 2712. 6167 [Veers Staats <staatsvr@asc.hpc.mil>, 6168 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 6169 6170 *) Reformat the FAQ so the different questions and answers can be divided 6171 in sections depending on the subject. 6172 [Richard Levitte] 6173 6174 *) Have the zlib compression code load ZLIB.DLL dynamically under 6175 Windows. 6176 [Richard Levitte] 6177 6178 *) New function BN_mod_sqrt for computing square roots modulo a prime 6179 (using the probabilistic Tonelli-Shanks algorithm unless 6180 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 6181 be handled deterministically). 6182 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 6183 6184 *) Make BN_mod_inverse faster by explicitly handling small quotients 6185 in the Euclid loop. (Speed gain about 20% for small moduli [256 or 6186 512 bits], about 30% for larger ones [1024 or 2048 bits].) 6187 [Bodo Moeller] 6188 6189 *) New function BN_kronecker. 6190 [Bodo Moeller] 6191 6192 *) Fix BN_gcd so that it works on negative inputs; the result is 6193 positive unless both parameters are zero. 6194 Previously something reasonably close to an infinite loop was 6195 possible because numbers could be growing instead of shrinking 6196 in the implementation of Euclid's algorithm. 6197 [Bodo Moeller] 6198 6199 *) Fix BN_is_word() and BN_is_one() macros to take into account the 6200 sign of the number in question. 6201 6202 Fix BN_is_word(a,w) to work correctly for w == 0. 6203 6204 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 6205 because its test if the absolute value of 'a' equals 'w'. 6206 Note that BN_abs_is_word does *not* handle w == 0 reliably; 6207 it exists mostly for use in the implementations of BN_is_zero(), 6208 BN_is_one(), and BN_is_word(). 6209 [Bodo Moeller] 6210 6211 *) New function BN_swap. 6212 [Bodo Moeller] 6213 6214 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 6215 the exponentiation functions are more likely to produce reasonable 6216 results on negative inputs. 6217 [Bodo Moeller] 6218 6219 *) Change BN_mod_mul so that the result is always non-negative. 6220 Previously, it could be negative if one of the factors was negative; 6221 I don't think anyone really wanted that behaviour. 6222 [Bodo Moeller] 6223 6224 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 6225 (except for exponentiation, which stays in crypto/bn/bn_exp.c, 6226 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 6227 and add new functions: 6228 6229 BN_nnmod 6230 BN_mod_sqr 6231 BN_mod_add 6232 BN_mod_add_quick 6233 BN_mod_sub 6234 BN_mod_sub_quick 6235 BN_mod_lshift1 6236 BN_mod_lshift1_quick 6237 BN_mod_lshift 6238 BN_mod_lshift_quick 6239 6240 These functions always generate non-negative results. 6241 6242 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 6243 such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 6244 6245 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 6246 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 6247 be reduced modulo m. 6248 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 6249 6250#if 0 6251 The following entry accidentily appeared in the CHANGES file 6252 distributed with OpenSSL 0.9.7. The modifications described in 6253 it do *not* apply to OpenSSL 0.9.7. 6254 6255 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 6256 was actually never needed) and in BN_mul(). The removal in BN_mul() 6257 required a small change in bn_mul_part_recursive() and the addition 6258 of the functions bn_cmp_part_words(), bn_sub_part_words() and 6259 bn_add_part_words(), which do the same thing as bn_cmp_words(), 6260 bn_sub_words() and bn_add_words() except they take arrays with 6261 differing sizes. 6262 [Richard Levitte] 6263#endif 6264 6265 *) In 'openssl passwd', verify passwords read from the terminal 6266 unless the '-salt' option is used (which usually means that 6267 verification would just waste user's time since the resulting 6268 hash is going to be compared with some given password hash) 6269 or the new '-noverify' option is used. 6270 6271 This is an incompatible change, but it does not affect 6272 non-interactive use of 'openssl passwd' (passwords on the command 6273 line, '-stdin' option, '-in ...' option) and thus should not 6274 cause any problems. 6275 [Bodo Moeller] 6276 6277 *) Remove all references to RSAref, since there's no more need for it. 6278 [Richard Levitte] 6279 6280 *) Make DSO load along a path given through an environment variable 6281 (SHLIB_PATH) with shl_load(). 6282 [Richard Levitte] 6283 6284 *) Constify the ENGINE code as a result of BIGNUM constification. 6285 Also constify the RSA code and most things related to it. In a 6286 few places, most notable in the depth of the ASN.1 code, ugly 6287 casts back to non-const were required (to be solved at a later 6288 time) 6289 [Richard Levitte] 6290 6291 *) Make it so the openssl application has all engines loaded by default. 6292 [Richard Levitte] 6293 6294 *) Constify the BIGNUM routines a little more. 6295 [Richard Levitte] 6296 6297 *) Add the following functions: 6298 6299 ENGINE_load_cswift() 6300 ENGINE_load_chil() 6301 ENGINE_load_atalla() 6302 ENGINE_load_nuron() 6303 ENGINE_load_builtin_engines() 6304 6305 That way, an application can itself choose if external engines that 6306 are built-in in OpenSSL shall ever be used or not. The benefit is 6307 that applications won't have to be linked with libdl or other dso 6308 libraries unless it's really needed. 6309 6310 Changed 'openssl engine' to load all engines on demand. 6311 Changed the engine header files to avoid the duplication of some 6312 declarations (they differed!). 6313 [Richard Levitte] 6314 6315 *) 'openssl engine' can now list capabilities. 6316 [Richard Levitte] 6317 6318 *) Better error reporting in 'openssl engine'. 6319 [Richard Levitte] 6320 6321 *) Never call load_dh_param(NULL) in s_server. 6322 [Bodo Moeller] 6323 6324 *) Add engine application. It can currently list engines by name and 6325 identity, and test if they are actually available. 6326 [Richard Levitte] 6327 6328 *) Improve RPM specification file by forcing symbolic linking and making 6329 sure the installed documentation is also owned by root.root. 6330 [Damien Miller <djm@mindrot.org>] 6331 6332 *) Give the OpenSSL applications more possibilities to make use of 6333 keys (public as well as private) handled by engines. 6334 [Richard Levitte] 6335 6336 *) Add OCSP code that comes from CertCo. 6337 [Richard Levitte] 6338 6339 *) Add VMS support for the Rijndael code. 6340 [Richard Levitte] 6341 6342 *) Added untested support for Nuron crypto accelerator. 6343 [Ben Laurie] 6344 6345 *) Add support for external cryptographic devices. This code was 6346 previously distributed separately as the "engine" branch. 6347 [Geoff Thorpe, Richard Levitte] 6348 6349 *) Rework the filename-translation in the DSO code. It is now possible to 6350 have far greater control over how a "name" is turned into a filename 6351 depending on the operating environment and any oddities about the 6352 different shared library filenames on each system. 6353 [Geoff Thorpe] 6354 6355 *) Support threads on FreeBSD-elf in Configure. 6356 [Richard Levitte] 6357 6358 *) Fix for SHA1 assembly problem with MASM: it produces 6359 warnings about corrupt line number information when assembling 6360 with debugging information. This is caused by the overlapping 6361 of two sections. 6362 [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 6363 6364 *) NCONF changes. 6365 NCONF_get_number() has no error checking at all. As a replacement, 6366 NCONF_get_number_e() is defined (_e for "error checking") and is 6367 promoted strongly. The old NCONF_get_number is kept around for 6368 binary backward compatibility. 6369 Make it possible for methods to load from something other than a BIO, 6370 by providing a function pointer that is given a name instead of a BIO. 6371 For example, this could be used to load configuration data from an 6372 LDAP server. 6373 [Richard Levitte] 6374 6375 *) Fix for non blocking accept BIOs. Added new I/O special reason 6376 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 6377 with non blocking I/O was not possible because no retry code was 6378 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 6379 this case. 6380 [Steve Henson] 6381 6382 *) Added the beginnings of Rijndael support. 6383 [Ben Laurie] 6384 6385 *) Fix for bug in DirectoryString mask setting. Add support for 6386 X509_NAME_print_ex() in 'req' and X509_print_ex() function 6387 to allow certificate printing to more controllable, additional 6388 'certopt' option to 'x509' to allow new printing options to be 6389 set. 6390 [Steve Henson] 6391 6392 *) Clean old EAY MD5 hack from e_os.h. 6393 [Richard Levitte] 6394 6395 Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 6396 6397 *) Fix null-pointer assignment in do_change_cipher_spec() revealed 6398 by using the Codenomicon TLS Test Tool (CVE-2004-0079) 6399 [Joe Orton, Steve Henson] 6400 6401 Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 6402 6403 *) Fix additional bug revealed by the NISCC test suite: 6404 6405 Stop bug triggering large recursion when presented with 6406 certain ASN.1 tags (CVE-2003-0851) 6407 [Steve Henson] 6408 6409 Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 6410 6411 *) Fix various bugs revealed by running the NISCC test suite: 6412 6413 Stop out of bounds reads in the ASN1 code when presented with 6414 invalid tags (CVE-2003-0543 and CVE-2003-0544). 6415 6416 If verify callback ignores invalid public key errors don't try to check 6417 certificate signature with the NULL public key. 6418 6419 [Steve Henson] 6420 6421 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 6422 if the server requested one: as stated in TLS 1.0 and SSL 3.0 6423 specifications. 6424 [Steve Henson] 6425 6426 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 6427 extra data after the compression methods not only for TLS 1.0 6428 but also for SSL 3.0 (as required by the specification). 6429 [Bodo Moeller; problem pointed out by Matthias Loepfe] 6430 6431 *) Change X509_certificate_type() to mark the key as exported/exportable 6432 when it's 512 *bits* long, not 512 bytes. 6433 [Richard Levitte] 6434 6435 Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 6436 6437 *) Countermeasure against the Klima-Pokorny-Rosa extension of 6438 Bleichbacher's attack on PKCS #1 v1.5 padding: treat 6439 a protocol version number mismatch like a decryption error 6440 in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 6441 [Bodo Moeller] 6442 6443 *) Turn on RSA blinding by default in the default implementation 6444 to avoid a timing attack. Applications that don't want it can call 6445 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 6446 They would be ill-advised to do so in most cases. 6447 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 6448 6449 *) Change RSA blinding code so that it works when the PRNG is not 6450 seeded (in this case, the secret RSA exponent is abused as 6451 an unpredictable seed -- if it is not unpredictable, there 6452 is no point in blinding anyway). Make RSA blinding thread-safe 6453 by remembering the creator's thread ID in rsa->blinding and 6454 having all other threads use local one-time blinding factors 6455 (this requires more computation than sharing rsa->blinding, but 6456 avoids excessive locking; and if an RSA object is not shared 6457 between threads, blinding will still be very fast). 6458 [Bodo Moeller] 6459 6460 Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 6461 6462 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 6463 via timing by performing a MAC computation even if incorrrect 6464 block cipher padding has been found. This is a countermeasure 6465 against active attacks where the attacker has to distinguish 6466 between bad padding and a MAC verification error. (CVE-2003-0078) 6467 6468 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 6469 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 6470 Martin Vuagnoux (EPFL, Ilion)] 6471 6472 Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 6473 6474 *) New function OPENSSL_cleanse(), which is used to cleanse a section of 6475 memory from it's contents. This is done with a counter that will 6476 place alternating values in each byte. This can be used to solve 6477 two issues: 1) the removal of calls to memset() by highly optimizing 6478 compilers, and 2) cleansing with other values than 0, since those can 6479 be read through on certain media, for example a swap space on disk. 6480 [Geoff Thorpe] 6481 6482 *) Bugfix: client side session caching did not work with external caching, 6483 because the session->cipher setting was not restored when reloading 6484 from the external cache. This problem was masked, when 6485 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 6486 (Found by Steve Haslam <steve@araqnid.ddts.net>.) 6487 [Lutz Jaenicke] 6488 6489 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 6490 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 6491 [Zeev Lieber <zeev-l@yahoo.com>] 6492 6493 *) Undo an undocumented change introduced in 0.9.6e which caused 6494 repeated calls to OpenSSL_add_all_ciphers() and 6495 OpenSSL_add_all_digests() to be ignored, even after calling 6496 EVP_cleanup(). 6497 [Richard Levitte] 6498 6499 *) Change the default configuration reader to deal with last line not 6500 being properly terminated. 6501 [Richard Levitte] 6502 6503 *) Change X509_NAME_cmp() so it applies the special rules on handling 6504 DN values that are of type PrintableString, as well as RDNs of type 6505 emailAddress where the value has the type ia5String. 6506 [stefank@valicert.com via Richard Levitte] 6507 6508 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 6509 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 6510 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 6511 the bitwise-OR of the two for use by the majority of applications 6512 wanting this behaviour, and update the docs. The documented 6513 behaviour and actual behaviour were inconsistent and had been 6514 changing anyway, so this is more a bug-fix than a behavioural 6515 change. 6516 [Geoff Thorpe, diagnosed by Nadav Har'El] 6517 6518 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 6519 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 6520 [Bodo Moeller] 6521 6522 *) Fix initialization code race conditions in 6523 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 6524 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 6525 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 6526 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 6527 ssl2_get_cipher_by_char(), 6528 ssl3_get_cipher_by_char(). 6529 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 6530 6531 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 6532 the cached sessions are flushed, as the remove_cb() might use ex_data 6533 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 6534 (see [openssl.org #212]). 6535 [Geoff Thorpe, Lutz Jaenicke] 6536 6537 *) Fix typo in OBJ_txt2obj which incorrectly passed the content 6538 length, instead of the encoding length to d2i_ASN1_OBJECT. 6539 [Steve Henson] 6540 6541 Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 6542 6543 *) [In 0.9.6g-engine release:] 6544 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 6545 [Lynn Gazis <lgazis@rainbow.com>] 6546 6547 Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 6548 6549 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 6550 and get fix the header length calculation. 6551 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 6552 Alon Kantor <alonk@checkpoint.com> (and others), 6553 Steve Henson] 6554 6555 *) Use proper error handling instead of 'assertions' in buffer 6556 overflow checks added in 0.9.6e. This prevents DoS (the 6557 assertions could call abort()). 6558 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 6559 6560 Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 6561 6562 *) Add various sanity checks to asn1_get_length() to reject 6563 the ASN1 length bytes if they exceed sizeof(long), will appear 6564 negative or the content length exceeds the length of the 6565 supplied buffer. 6566 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6567 6568 *) Fix cipher selection routines: ciphers without encryption had no flags 6569 for the cipher strength set and where therefore not handled correctly 6570 by the selection routines (PR #130). 6571 [Lutz Jaenicke] 6572 6573 *) Fix EVP_dsa_sha macro. 6574 [Nils Larsch] 6575 6576 *) New option 6577 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 6578 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 6579 that was added in OpenSSL 0.9.6d. 6580 6581 As the countermeasure turned out to be incompatible with some 6582 broken SSL implementations, the new option is part of SSL_OP_ALL. 6583 SSL_OP_ALL is usually employed when compatibility with weird SSL 6584 implementations is desired (e.g. '-bugs' option to 's_client' and 6585 's_server'), so the new option is automatically set in many 6586 applications. 6587 [Bodo Moeller] 6588 6589 *) Changes in security patch: 6590 6591 Changes marked "(CHATS)" were sponsored by the Defense Advanced 6592 Research Projects Agency (DARPA) and Air Force Research Laboratory, 6593 Air Force Materiel Command, USAF, under agreement number 6594 F30602-01-2-0537. 6595 6596 *) Add various sanity checks to asn1_get_length() to reject 6597 the ASN1 length bytes if they exceed sizeof(long), will appear 6598 negative or the content length exceeds the length of the 6599 supplied buffer. (CVE-2002-0659) 6600 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6601 6602 *) Assertions for various potential buffer overflows, not known to 6603 happen in practice. 6604 [Ben Laurie (CHATS)] 6605 6606 *) Various temporary buffers to hold ASCII versions of integers were 6607 too small for 64 bit platforms. (CVE-2002-0655) 6608 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 6609 6610 *) Remote buffer overflow in SSL3 protocol - an attacker could 6611 supply an oversized session ID to a client. (CVE-2002-0656) 6612 [Ben Laurie (CHATS)] 6613 6614 *) Remote buffer overflow in SSL2 protocol - an attacker could 6615 supply an oversized client master key. (CVE-2002-0656) 6616 [Ben Laurie (CHATS)] 6617 6618 Changes between 0.9.6c and 0.9.6d [9 May 2002] 6619 6620 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 6621 encoded as NULL) with id-dsa-with-sha1. 6622 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 6623 6624 *) Check various X509_...() return values in apps/req.c. 6625 [Nils Larsch <nla@trustcenter.de>] 6626 6627 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 6628 an end-of-file condition would erronously be flagged, when the CRLF 6629 was just at the end of a processed block. The bug was discovered when 6630 processing data through a buffering memory BIO handing the data to a 6631 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 6632 <ptsekov@syntrex.com> and Nedelcho Stanev. 6633 [Lutz Jaenicke] 6634 6635 *) Implement a countermeasure against a vulnerability recently found 6636 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 6637 before application data chunks to avoid the use of known IVs 6638 with data potentially chosen by the attacker. 6639 [Bodo Moeller] 6640 6641 *) Fix length checks in ssl3_get_client_hello(). 6642 [Bodo Moeller] 6643 6644 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 6645 to prevent ssl3_read_internal() from incorrectly assuming that 6646 ssl3_read_bytes() found application data while handshake 6647 processing was enabled when in fact s->s3->in_read_app_data was 6648 merely automatically cleared during the initial handshake. 6649 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 6650 6651 *) Fix object definitions for Private and Enterprise: they were not 6652 recognized in their shortname (=lowercase) representation. Extend 6653 obj_dat.pl to issue an error when using undefined keywords instead 6654 of silently ignoring the problem (Svenning Sorensen 6655 <sss@sss.dnsalias.net>). 6656 [Lutz Jaenicke] 6657 6658 *) Fix DH_generate_parameters() so that it works for 'non-standard' 6659 generators, i.e. generators other than 2 and 5. (Previously, the 6660 code did not properly initialise the 'add' and 'rem' values to 6661 BN_generate_prime().) 6662 6663 In the new general case, we do not insist that 'generator' is 6664 actually a primitive root: This requirement is rather pointless; 6665 a generator of the order-q subgroup is just as good, if not 6666 better. 6667 [Bodo Moeller] 6668 6669 *) Map new X509 verification errors to alerts. Discovered and submitted by 6670 Tom Wu <tom@arcot.com>. 6671 [Lutz Jaenicke] 6672 6673 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 6674 returning non-zero before the data has been completely received 6675 when using non-blocking I/O. 6676 [Bodo Moeller; problem pointed out by John Hughes] 6677 6678 *) Some of the ciphers missed the strength entry (SSL_LOW etc). 6679 [Ben Laurie, Lutz Jaenicke] 6680 6681 *) Fix bug in SSL_clear(): bad sessions were not removed (found by 6682 Yoram Zahavi <YoramZ@gilian.com>). 6683 [Lutz Jaenicke] 6684 6685 *) Add information about CygWin 1.3 and on, and preserve proper 6686 configuration for the versions before that. 6687 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 6688 6689 *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 6690 check whether we deal with a copy of a session and do not delete from 6691 the cache in this case. Problem reported by "Izhar Shoshani Levi" 6692 <izhar@checkpoint.com>. 6693 [Lutz Jaenicke] 6694 6695 *) Do not store session data into the internal session cache, if it 6696 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 6697 flag is set). Proposed by Aslam <aslam@funk.com>. 6698 [Lutz Jaenicke] 6699 6700 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 6701 value is 0. 6702 [Richard Levitte] 6703 6704 *) [In 0.9.6d-engine release:] 6705 Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 6706 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 6707 6708 *) Add the configuration target linux-s390x. 6709 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 6710 6711 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 6712 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 6713 variable as an indication that a ClientHello message has been 6714 received. As the flag value will be lost between multiple 6715 invocations of ssl3_accept when using non-blocking I/O, the 6716 function may not be aware that a handshake has actually taken 6717 place, thus preventing a new session from being added to the 6718 session cache. 6719 6720 To avoid this problem, we now set s->new_session to 2 instead of 6721 using a local variable. 6722 [Lutz Jaenicke, Bodo Moeller] 6723 6724 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 6725 if the SSL_R_LENGTH_MISMATCH error is detected. 6726 [Geoff Thorpe, Bodo Moeller] 6727 6728 *) New 'shared_ldflag' column in Configure platform table. 6729 [Richard Levitte] 6730 6731 *) Fix EVP_CIPHER_mode macro. 6732 ["Dan S. Camper" <dan@bti.net>] 6733 6734 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 6735 type, we must throw them away by setting rr->length to 0. 6736 [D P Chang <dpc@qualys.com>] 6737 6738 Changes between 0.9.6b and 0.9.6c [21 dec 2001] 6739 6740 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 6741 <Dominikus.Scherkl@biodata.com>. (The previous implementation 6742 worked incorrectly for those cases where range = 10..._2 and 6743 3*range is two bits longer than range.) 6744 [Bodo Moeller] 6745 6746 *) Only add signing time to PKCS7 structures if it is not already 6747 present. 6748 [Steve Henson] 6749 6750 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 6751 OBJ_ld_ce should be OBJ_id_ce. 6752 Also some ip-pda OIDs in crypto/objects/objects.txt were 6753 incorrect (cf. RFC 3039). 6754 [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 6755 6756 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 6757 returns early because it has nothing to do. 6758 [Andy Schneider <andy.schneider@bjss.co.uk>] 6759 6760 *) [In 0.9.6c-engine release:] 6761 Fix mutex callback return values in crypto/engine/hw_ncipher.c. 6762 [Andy Schneider <andy.schneider@bjss.co.uk>] 6763 6764 *) [In 0.9.6c-engine release:] 6765 Add support for Cryptographic Appliance's keyserver technology. 6766 (Use engine 'keyclient') 6767 [Cryptographic Appliances and Geoff Thorpe] 6768 6769 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 6770 is called via tools/c89.sh because arguments have to be 6771 rearranged (all '-L' options must appear before the first object 6772 modules). 6773 [Richard Shapiro <rshapiro@abinitio.com>] 6774 6775 *) [In 0.9.6c-engine release:] 6776 Add support for Broadcom crypto accelerator cards, backported 6777 from 0.9.7. 6778 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 6779 6780 *) [In 0.9.6c-engine release:] 6781 Add support for SureWare crypto accelerator cards from 6782 Baltimore Technologies. (Use engine 'sureware') 6783 [Baltimore Technologies and Mark Cox] 6784 6785 *) [In 0.9.6c-engine release:] 6786 Add support for crypto accelerator cards from Accelerated 6787 Encryption Processing, www.aep.ie. (Use engine 'aep') 6788 [AEP Inc. and Mark Cox] 6789 6790 *) Add a configuration entry for gcc on UnixWare. 6791 [Gary Benson <gbenson@redhat.com>] 6792 6793 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 6794 messages are stored in a single piece (fixed-length part and 6795 variable-length part combined) and fix various bugs found on the way. 6796 [Bodo Moeller] 6797 6798 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 6799 instead. BIO_gethostbyname() does not know what timeouts are 6800 appropriate, so entries would stay in cache even when they have 6801 become invalid. 6802 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 6803 6804 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 6805 faced with a pathologically small ClientHello fragment that does 6806 not contain client_version: Instead of aborting with an error, 6807 simply choose the highest available protocol version (i.e., 6808 TLS 1.0 unless it is disabled). In practice, ClientHello 6809 messages are never sent like this, but this change gives us 6810 strictly correct behaviour at least for TLS. 6811 [Bodo Moeller] 6812 6813 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 6814 never resets s->method to s->ctx->method when called from within 6815 one of the SSL handshake functions. 6816 [Bodo Moeller; problem pointed out by Niko Baric] 6817 6818 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 6819 (sent using the client's version number) if client_version is 6820 smaller than the protocol version in use. Also change 6821 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 6822 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 6823 the client will at least see that alert. 6824 [Bodo Moeller] 6825 6826 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 6827 correctly. 6828 [Bodo Moeller] 6829 6830 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 6831 client receives HelloRequest while in a handshake. 6832 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 6833 6834 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 6835 should end in 'break', not 'goto end' which circuments various 6836 cleanups done in state SSL_ST_OK. But session related stuff 6837 must be disabled for SSL_ST_OK in the case that we just sent a 6838 HelloRequest. 6839 6840 Also avoid some overhead by not calling ssl_init_wbio_buffer() 6841 before just sending a HelloRequest. 6842 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 6843 6844 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 6845 reveal whether illegal block cipher padding was found or a MAC 6846 verification error occured. (Neither SSLerr() codes nor alerts 6847 are directly visible to potential attackers, but the information 6848 may leak via logfiles.) 6849 6850 Similar changes are not required for the SSL 2.0 implementation 6851 because the number of padding bytes is sent in clear for SSL 2.0, 6852 and the extra bytes are just ignored. However ssl/s2_pkt.c 6853 failed to verify that the purported number of padding bytes is in 6854 the legal range. 6855 [Bodo Moeller] 6856 6857 *) Add OpenUNIX-8 support including shared libraries 6858 (Boyd Lynn Gerber <gerberb@zenez.com>). 6859 [Lutz Jaenicke] 6860 6861 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 6862 'wristwatch attack' using huge encoding parameters (cf. 6863 James H. Manger's CRYPTO 2001 paper). Note that the 6864 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 6865 encoding parameters and hence was not vulnerable. 6866 [Bodo Moeller] 6867 6868 *) BN_sqr() bug fix. 6869 [Ulf M��ller, reported by Jim Ellis <jim.ellis@cavium.com>] 6870 6871 *) Rabin-Miller test analyses assume uniformly distributed witnesses, 6872 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 6873 followed by modular reduction. 6874 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 6875 6876 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 6877 equivalent based on BN_pseudo_rand() instead of BN_rand(). 6878 [Bodo Moeller] 6879 6880 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 6881 This function was broken, as the check for a new client hello message 6882 to handle SGC did not allow these large messages. 6883 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 6884 [Lutz Jaenicke] 6885 6886 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 6887 [Lutz Jaenicke] 6888 6889 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 6890 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 6891 [Lutz Jaenicke] 6892 6893 *) Rework the configuration and shared library support for Tru64 Unix. 6894 The configuration part makes use of modern compiler features and 6895 still retains old compiler behavior for those that run older versions 6896 of the OS. The shared library support part includes a variant that 6897 uses the RPATH feature, and is available through the special 6898 configuration target "alpha-cc-rpath", which will never be selected 6899 automatically. 6900 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 6901 6902 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 6903 with the same message size as in ssl3_get_certificate_request(). 6904 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 6905 messages might inadvertently be reject as too long. 6906 [Petr Lampa <lampa@fee.vutbr.cz>] 6907 6908 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 6909 [Andy Polyakov] 6910 6911 *) Modified SSL library such that the verify_callback that has been set 6912 specificly for an SSL object with SSL_set_verify() is actually being 6913 used. Before the change, a verify_callback set with this function was 6914 ignored and the verify_callback() set in the SSL_CTX at the time of 6915 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 6916 to allow the necessary settings. 6917 [Lutz Jaenicke] 6918 6919 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 6920 explicitly to NULL, as at least on Solaris 8 this seems not always to be 6921 done automatically (in contradiction to the requirements of the C 6922 standard). This made problems when used from OpenSSH. 6923 [Lutz Jaenicke] 6924 6925 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 6926 dh->length and always used 6927 6928 BN_rand_range(priv_key, dh->p). 6929 6930 BN_rand_range() is not necessary for Diffie-Hellman, and this 6931 specific range makes Diffie-Hellman unnecessarily inefficient if 6932 dh->length (recommended exponent length) is much smaller than the 6933 length of dh->p. We could use BN_rand_range() if the order of 6934 the subgroup was stored in the DH structure, but we only have 6935 dh->length. 6936 6937 So switch back to 6938 6939 BN_rand(priv_key, l, ...) 6940 6941 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 6942 otherwise. 6943 [Bodo Moeller] 6944 6945 *) In 6946 6947 RSA_eay_public_encrypt 6948 RSA_eay_private_decrypt 6949 RSA_eay_private_encrypt (signing) 6950 RSA_eay_public_decrypt (signature verification) 6951 6952 (default implementations for RSA_public_encrypt, 6953 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 6954 always reject numbers >= n. 6955 [Bodo Moeller] 6956 6957 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 6958 to synchronize access to 'locking_thread'. This is necessary on 6959 systems where access to 'locking_thread' (an 'unsigned long' 6960 variable) is not atomic. 6961 [Bodo Moeller] 6962 6963 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 6964 *before* setting the 'crypto_lock_rand' flag. The previous code had 6965 a race condition if 0 is a valid thread ID. 6966 [Travis Vitek <vitek@roguewave.com>] 6967 6968 *) Add support for shared libraries under Irix. 6969 [Albert Chin-A-Young <china@thewrittenword.com>] 6970 6971 *) Add configuration option to build on Linux on both big-endian and 6972 little-endian MIPS. 6973 [Ralf Baechle <ralf@uni-koblenz.de>] 6974 6975 *) Add the possibility to create shared libraries on HP-UX. 6976 [Richard Levitte] 6977 6978 Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 6979 6980 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 6981 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 6982 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 6983 PRNG state recovery was possible based on the output of 6984 one PRNG request appropriately sized to gain knowledge on 6985 'md' followed by enough consecutive 1-byte PRNG requests 6986 to traverse all of 'state'. 6987 6988 1. When updating 'md_local' (the current thread's copy of 'md') 6989 during PRNG output generation, hash all of the previous 6990 'md_local' value, not just the half used for PRNG output. 6991 6992 2. Make the number of bytes from 'state' included into the hash 6993 independent from the number of PRNG bytes requested. 6994 6995 The first measure alone would be sufficient to avoid 6996 Markku-Juhani's attack. (Actually it had never occurred 6997 to me that the half of 'md_local' used for chaining was the 6998 half from which PRNG output bytes were taken -- I had always 6999 assumed that the secret half would be used.) The second 7000 measure makes sure that additional data from 'state' is never 7001 mixed into 'md_local' in small portions; this heuristically 7002 further strengthens the PRNG. 7003 [Bodo Moeller] 7004 7005 *) Fix crypto/bn/asm/mips3.s. 7006 [Andy Polyakov] 7007 7008 *) When only the key is given to "enc", the IV is undefined. Print out 7009 an error message in this case. 7010 [Lutz Jaenicke] 7011 7012 *) Handle special case when X509_NAME is empty in X509 printing routines. 7013 [Steve Henson] 7014 7015 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 7016 positive and less than q. 7017 [Bodo Moeller] 7018 7019 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 7020 used: it isn't thread safe and the add_lock_callback should handle 7021 that itself. 7022 [Paul Rose <Paul.Rose@bridge.com>] 7023 7024 *) Verify that incoming data obeys the block size in 7025 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 7026 [Bodo Moeller] 7027 7028 *) Fix OAEP check. 7029 [Ulf M��ller, Bodo M��ller] 7030 7031 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 7032 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 7033 when fixing the server behaviour for backwards-compatible 'client 7034 hello' messages. (Note that the attack is impractical against 7035 SSL 3.0 and TLS 1.0 anyway because length and version checking 7036 means that the probability of guessing a valid ciphertext is 7037 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 7038 paper.) 7039 7040 Before 0.9.5, the countermeasure (hide the error by generating a 7041 random 'decryption result') did not work properly because 7042 ERR_clear_error() was missing, meaning that SSL_get_error() would 7043 detect the supposedly ignored error. 7044 7045 Both problems are now fixed. 7046 [Bodo Moeller] 7047 7048 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 7049 (previously it was 1024). 7050 [Bodo Moeller] 7051 7052 *) Fix for compatibility mode trust settings: ignore trust settings 7053 unless some valid trust or reject settings are present. 7054 [Steve Henson] 7055 7056 *) Fix for blowfish EVP: its a variable length cipher. 7057 [Steve Henson] 7058 7059 *) Fix various bugs related to DSA S/MIME verification. Handle missing 7060 parameters in DSA public key structures and return an error in the 7061 DSA routines if parameters are absent. 7062 [Steve Henson] 7063 7064 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 7065 in the current directory if neither $RANDFILE nor $HOME was set. 7066 RAND_file_name() in 0.9.6a returned NULL in this case. This has 7067 caused some confusion to Windows users who haven't defined $HOME. 7068 Thus RAND_file_name() is changed again: e_os.h can define a 7069 DEFAULT_HOME, which will be used if $HOME is not set. 7070 For Windows, we use "C:"; on other platforms, we still require 7071 environment variables. 7072 7073 *) Move 'if (!initialized) RAND_poll()' into regions protected by 7074 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 7075 having multiple threads call RAND_poll() concurrently. 7076 [Bodo Moeller] 7077 7078 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 7079 combination of a flag and a thread ID variable. 7080 Otherwise while one thread is in ssleay_rand_bytes (which sets the 7081 flag), *other* threads can enter ssleay_add_bytes without obeying 7082 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 7083 that they do not hold after the first thread unsets add_do_not_lock). 7084 [Bodo Moeller] 7085 7086 *) Change bctest again: '-x' expressions are not available in all 7087 versions of 'test'. 7088 [Bodo Moeller] 7089 7090 Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 7091 7092 *) Fix a couple of memory leaks in PKCS7_dataDecode() 7093 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 7094 7095 *) Change Configure and Makefiles to provide EXE_EXT, which will contain 7096 the default extension for executables, if any. Also, make the perl 7097 scripts that use symlink() to test if it really exists and use "cp" 7098 if it doesn't. All this made OpenSSL compilable and installable in 7099 CygWin. 7100 [Richard Levitte] 7101 7102 *) Fix for asn1_GetSequence() for indefinite length constructed data. 7103 If SEQUENCE is length is indefinite just set c->slen to the total 7104 amount of data available. 7105 [Steve Henson, reported by shige@FreeBSD.org] 7106 [This change does not apply to 0.9.7.] 7107 7108 *) Change bctest to avoid here-documents inside command substitution 7109 (workaround for FreeBSD /bin/sh bug). 7110 For compatibility with Ultrix, avoid shell functions (introduced 7111 in the bctest version that searches along $PATH). 7112 [Bodo Moeller] 7113 7114 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 7115 with des_encrypt() defined on some operating systems, like Solaris 7116 and UnixWare. 7117 [Richard Levitte] 7118 7119 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 7120 On the Importance of Eliminating Errors in Cryptographic 7121 Computations, J. Cryptology 14 (2001) 2, 101-119, 7122 http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 7123 [Ulf Moeller] 7124 7125 *) MIPS assembler BIGNUM division bug fix. 7126 [Andy Polyakov] 7127 7128 *) Disabled incorrect Alpha assembler code. 7129 [Richard Levitte] 7130 7131 *) Fix PKCS#7 decode routines so they correctly update the length 7132 after reading an EOC for the EXPLICIT tag. 7133 [Steve Henson] 7134 [This change does not apply to 0.9.7.] 7135 7136 *) Fix bug in PKCS#12 key generation routines. This was triggered 7137 if a 3DES key was generated with a 0 initial byte. Include 7138 PKCS12_BROKEN_KEYGEN compilation option to retain the old 7139 (but broken) behaviour. 7140 [Steve Henson] 7141 7142 *) Enhance bctest to search for a working bc along $PATH and print 7143 it when found. 7144 [Tim Rice <tim@multitalents.net> via Richard Levitte] 7145 7146 *) Fix memory leaks in err.c: free err_data string if necessary; 7147 don't write to the wrong index in ERR_set_error_data. 7148 [Bodo Moeller] 7149 7150 *) Implement ssl23_peek (analogous to ssl23_read), which previously 7151 did not exist. 7152 [Bodo Moeller] 7153 7154 *) Replace rdtsc with _emit statements for VC++ version 5. 7155 [Jeremy Cooper <jeremy@baymoo.org>] 7156 7157 *) Make it possible to reuse SSLv2 sessions. 7158 [Richard Levitte] 7159 7160 *) In copy_email() check for >= 0 as a return value for 7161 X509_NAME_get_index_by_NID() since 0 is a valid index. 7162 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 7163 7164 *) Avoid coredump with unsupported or invalid public keys by checking if 7165 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 7166 PKCS7_verify() fails with non detached data. 7167 [Steve Henson] 7168 7169 *) Don't use getenv in library functions when run as setuid/setgid. 7170 New function OPENSSL_issetugid(). 7171 [Ulf Moeller] 7172 7173 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 7174 due to incorrect handling of multi-threading: 7175 7176 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 7177 7178 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 7179 7180 3. Count how many times MemCheck_off() has been called so that 7181 nested use can be treated correctly. This also avoids 7182 inband-signalling in the previous code (which relied on the 7183 assumption that thread ID 0 is impossible). 7184 [Bodo Moeller] 7185 7186 *) Add "-rand" option also to s_client and s_server. 7187 [Lutz Jaenicke] 7188 7189 *) Fix CPU detection on Irix 6.x. 7190 [Kurt Hockenbury <khockenb@stevens-tech.edu> and 7191 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 7192 7193 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 7194 was empty. 7195 [Steve Henson] 7196 [This change does not apply to 0.9.7.] 7197 7198 *) Use the cached encoding of an X509_NAME structure rather than 7199 copying it. This is apparently the reason for the libsafe "errors" 7200 but the code is actually correct. 7201 [Steve Henson] 7202 7203 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 7204 Bleichenbacher's DSA attack. 7205 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 7206 to be set and top=0 forces the highest bit to be set; top=-1 is new 7207 and leaves the highest bit random. 7208 [Ulf Moeller, Bodo Moeller] 7209 7210 *) In the NCONF_...-based implementations for CONF_... queries 7211 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 7212 a temporary CONF structure with the data component set to NULL 7213 (which gives segmentation faults in lh_retrieve). 7214 Instead, use NULL for the CONF pointer in CONF_get_string and 7215 CONF_get_number (which may use environment variables) and directly 7216 return NULL from CONF_get_section. 7217 [Bodo Moeller] 7218 7219 *) Fix potential buffer overrun for EBCDIC. 7220 [Ulf Moeller] 7221 7222 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 7223 keyUsage if basicConstraints absent for a CA. 7224 [Steve Henson] 7225 7226 *) Make SMIME_write_PKCS7() write mail header values with a format that 7227 is more generally accepted (no spaces before the semicolon), since 7228 some programs can't parse those values properly otherwise. Also make 7229 sure BIO's that break lines after each write do not create invalid 7230 headers. 7231 [Richard Levitte] 7232 7233 *) Make the CRL encoding routines work with empty SEQUENCE OF. The 7234 macros previously used would not encode an empty SEQUENCE OF 7235 and break the signature. 7236 [Steve Henson] 7237 [This change does not apply to 0.9.7.] 7238 7239 *) Zero the premaster secret after deriving the master secret in 7240 DH ciphersuites. 7241 [Steve Henson] 7242 7243 *) Add some EVP_add_digest_alias registrations (as found in 7244 OpenSSL_add_all_digests()) to SSL_library_init() 7245 aka OpenSSL_add_ssl_algorithms(). This provides improved 7246 compatibility with peers using X.509 certificates 7247 with unconventional AlgorithmIdentifier OIDs. 7248 [Bodo Moeller] 7249 7250 *) Fix for Irix with NO_ASM. 7251 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 7252 7253 *) ./config script fixes. 7254 [Ulf Moeller, Richard Levitte] 7255 7256 *) Fix 'openssl passwd -1'. 7257 [Bodo Moeller] 7258 7259 *) Change PKCS12_key_gen_asc() so it can cope with non null 7260 terminated strings whose length is passed in the passlen 7261 parameter, for example from PEM callbacks. This was done 7262 by adding an extra length parameter to asc2uni(). 7263 [Steve Henson, reported by <oddissey@samsung.co.kr>] 7264 7265 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 7266 call failed, free the DSA structure. 7267 [Bodo Moeller] 7268 7269 *) Fix to uni2asc() to cope with zero length Unicode strings. 7270 These are present in some PKCS#12 files. 7271 [Steve Henson] 7272 7273 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 7274 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 7275 when writing a 32767 byte record. 7276 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 7277 7278 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 7279 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 7280 7281 (RSA objects have a reference count access to which is protected 7282 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 7283 so they are meant to be shared between threads.) 7284 [Bodo Moeller, Geoff Thorpe; original patch submitted by 7285 "Reddie, Steven" <Steven.Reddie@ca.com>] 7286 7287 *) Fix a deadlock in CRYPTO_mem_leaks(). 7288 [Bodo Moeller] 7289 7290 *) Use better test patterns in bntest. 7291 [Ulf M��ller] 7292 7293 *) rand_win.c fix for Borland C. 7294 [Ulf M��ller] 7295 7296 *) BN_rshift bugfix for n == 0. 7297 [Bodo Moeller] 7298 7299 *) Add a 'bctest' script that checks for some known 'bc' bugs 7300 so that 'make test' does not abort just because 'bc' is broken. 7301 [Bodo Moeller] 7302 7303 *) Store verify_result within SSL_SESSION also for client side to 7304 avoid potential security hole. (Re-used sessions on the client side 7305 always resulted in verify_result==X509_V_OK, not using the original 7306 result of the server certificate verification.) 7307 [Lutz Jaenicke] 7308 7309 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 7310 SSL3_RT_APPLICATION_DATA, return 0. 7311 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 7312 [Bodo Moeller] 7313 7314 *) Fix SSL_peek: 7315 Both ssl2_peek and ssl3_peek, which were totally broken in earlier 7316 releases, have been re-implemented by renaming the previous 7317 implementations of ssl2_read and ssl3_read to ssl2_read_internal 7318 and ssl3_read_internal, respectively, and adding 'peek' parameters 7319 to them. The new ssl[23]_{read,peek} functions are calls to 7320 ssl[23]_read_internal with the 'peek' flag set appropriately. 7321 A 'peek' parameter has also been added to ssl3_read_bytes, which 7322 does the actual work for ssl3_read_internal. 7323 [Bodo Moeller] 7324 7325 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 7326 the method-specific "init()" handler. Also clean up ex_data after 7327 calling the method-specific "finish()" handler. Previously, this was 7328 happening the other way round. 7329 [Geoff Thorpe] 7330 7331 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 7332 The previous value, 12, was not always sufficient for BN_mod_exp(). 7333 [Bodo Moeller] 7334 7335 *) Make sure that shared libraries get the internal name engine with 7336 the full version number and not just 0. This should mark the 7337 shared libraries as not backward compatible. Of course, this should 7338 be changed again when we can guarantee backward binary compatibility. 7339 [Richard Levitte] 7340 7341 *) Fix typo in get_cert_by_subject() in by_dir.c 7342 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 7343 7344 *) Rework the system to generate shared libraries: 7345 7346 - Make note of the expected extension for the shared libraries and 7347 if there is a need for symbolic links from for example libcrypto.so.0 7348 to libcrypto.so.0.9.7. There is extended info in Configure for 7349 that. 7350 7351 - Make as few rebuilds of the shared libraries as possible. 7352 7353 - Still avoid linking the OpenSSL programs with the shared libraries. 7354 7355 - When installing, install the shared libraries separately from the 7356 static ones. 7357 [Richard Levitte] 7358 7359 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 7360 7361 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 7362 and not in SSL_clear because the latter is also used by the 7363 accept/connect functions; previously, the settings made by 7364 SSL_set_read_ahead would be lost during the handshake. 7365 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 7366 7367 *) Correct util/mkdef.pl to be selective about disabled algorithms. 7368 Previously, it would create entries for disableed algorithms no 7369 matter what. 7370 [Richard Levitte] 7371 7372 *) Added several new manual pages for SSL_* function. 7373 [Lutz Jaenicke] 7374 7375 Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 7376 7377 *) In ssl23_get_client_hello, generate an error message when faced 7378 with an initial SSL 3.0/TLS record that is too small to contain the 7379 first two bytes of the ClientHello message, i.e. client_version. 7380 (Note that this is a pathologic case that probably has never happened 7381 in real life.) The previous approach was to use the version number 7382 from the record header as a substitute; but our protocol choice 7383 should not depend on that one because it is not authenticated 7384 by the Finished messages. 7385 [Bodo Moeller] 7386 7387 *) More robust randomness gathering functions for Windows. 7388 [Jeffrey Altman <jaltman@columbia.edu>] 7389 7390 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 7391 not set then we don't setup the error code for issuer check errors 7392 to avoid possibly overwriting other errors which the callback does 7393 handle. If an application does set the flag then we assume it knows 7394 what it is doing and can handle the new informational codes 7395 appropriately. 7396 [Steve Henson] 7397 7398 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 7399 a general "ANY" type, as such it should be able to decode anything 7400 including tagged types. However it didn't check the class so it would 7401 wrongly interpret tagged types in the same way as their universal 7402 counterpart and unknown types were just rejected. Changed so that the 7403 tagged and unknown types are handled in the same way as a SEQUENCE: 7404 that is the encoding is stored intact. There is also a new type 7405 "V_ASN1_OTHER" which is used when the class is not universal, in this 7406 case we have no idea what the actual type is so we just lump them all 7407 together. 7408 [Steve Henson] 7409 7410 *) On VMS, stdout may very well lead to a file that is written to 7411 in a record-oriented fashion. That means that every write() will 7412 write a separate record, which will be read separately by the 7413 programs trying to read from it. This can be very confusing. 7414 7415 The solution is to put a BIO filter in the way that will buffer 7416 text until a linefeed is reached, and then write everything a 7417 line at a time, so every record written will be an actual line, 7418 not chunks of lines and not (usually doesn't happen, but I've 7419 seen it once) several lines in one record. BIO_f_linebuffer() is 7420 the answer. 7421 7422 Currently, it's a VMS-only method, because that's where it has 7423 been tested well enough. 7424 [Richard Levitte] 7425 7426 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 7427 it can return incorrect results. 7428 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 7429 but it was in 0.9.6-beta[12].) 7430 [Bodo Moeller] 7431 7432 *) Disable the check for content being present when verifying detached 7433 signatures in pk7_smime.c. Some versions of Netscape (wrongly) 7434 include zero length content when signing messages. 7435 [Steve Henson] 7436 7437 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 7438 BIO_ctrl (for BIO pairs). 7439 [Bodo M��ller] 7440 7441 *) Add DSO method for VMS. 7442 [Richard Levitte] 7443 7444 *) Bug fix: Montgomery multiplication could produce results with the 7445 wrong sign. 7446 [Ulf M��ller] 7447 7448 *) Add RPM specification openssl.spec and modify it to build three 7449 packages. The default package contains applications, application 7450 documentation and run-time libraries. The devel package contains 7451 include files, static libraries and function documentation. The 7452 doc package contains the contents of the doc directory. The original 7453 openssl.spec was provided by Damien Miller <djm@mindrot.org>. 7454 [Richard Levitte] 7455 7456 *) Add a large number of documentation files for many SSL routines. 7457 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 7458 7459 *) Add a configuration entry for Sony News 4. 7460 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 7461 7462 *) Don't set the two most significant bits to one when generating a 7463 random number < q in the DSA library. 7464 [Ulf M��ller] 7465 7466 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 7467 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 7468 the underlying transport is blocking) if a handshake took place. 7469 (The default behaviour is needed by applications such as s_client 7470 and s_server that use select() to determine when to use SSL_read; 7471 but for applications that know in advance when to expect data, it 7472 just makes things more complicated.) 7473 [Bodo Moeller] 7474 7475 *) Add RAND_egd_bytes(), which gives control over the number of bytes read 7476 from EGD. 7477 [Ben Laurie] 7478 7479 *) Add a few more EBCDIC conditionals that make `req' and `x509' 7480 work better on such systems. 7481 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 7482 7483 *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 7484 Update PKCS12_parse() so it copies the friendlyName and the 7485 keyid to the certificates aux info. 7486 [Steve Henson] 7487 7488 *) Fix bug in PKCS7_verify() which caused an infinite loop 7489 if there was more than one signature. 7490 [Sven Uszpelkat <su@celocom.de>] 7491 7492 *) Major change in util/mkdef.pl to include extra information 7493 about each symbol, as well as presentig variables as well 7494 as functions. This change means that there's n more need 7495 to rebuild the .num files when some algorithms are excluded. 7496 [Richard Levitte] 7497 7498 *) Allow the verify time to be set by an application, 7499 rather than always using the current time. 7500 [Steve Henson] 7501 7502 *) Phase 2 verify code reorganisation. The certificate 7503 verify code now looks up an issuer certificate by a 7504 number of criteria: subject name, authority key id 7505 and key usage. It also verifies self signed certificates 7506 by the same criteria. The main comparison function is 7507 X509_check_issued() which performs these checks. 7508 7509 Lot of changes were necessary in order to support this 7510 without completely rewriting the lookup code. 7511 7512 Authority and subject key identifier are now cached. 7513 7514 The LHASH 'certs' is X509_STORE has now been replaced 7515 by a STACK_OF(X509_OBJECT). This is mainly because an 7516 LHASH can't store or retrieve multiple objects with 7517 the same hash value. 7518 7519 As a result various functions (which were all internal 7520 use only) have changed to handle the new X509_STORE 7521 structure. This will break anything that messed round 7522 with X509_STORE internally. 7523 7524 The functions X509_STORE_add_cert() now checks for an 7525 exact match, rather than just subject name. 7526 7527 The X509_STORE API doesn't directly support the retrieval 7528 of multiple certificates matching a given criteria, however 7529 this can be worked round by performing a lookup first 7530 (which will fill the cache with candidate certificates) 7531 and then examining the cache for matches. This is probably 7532 the best we can do without throwing out X509_LOOKUP 7533 entirely (maybe later...). 7534 7535 The X509_VERIFY_CTX structure has been enhanced considerably. 7536 7537 All certificate lookup operations now go via a get_issuer() 7538 callback. Although this currently uses an X509_STORE it 7539 can be replaced by custom lookups. This is a simple way 7540 to bypass the X509_STORE hackery necessary to make this 7541 work and makes it possible to use more efficient techniques 7542 in future. A very simple version which uses a simple 7543 STACK for its trusted certificate store is also provided 7544 using X509_STORE_CTX_trusted_stack(). 7545 7546 The verify_cb() and verify() callbacks now have equivalents 7547 in the X509_STORE_CTX structure. 7548 7549 X509_STORE_CTX also has a 'flags' field which can be used 7550 to customise the verify behaviour. 7551 [Steve Henson] 7552 7553 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 7554 excludes S/MIME capabilities. 7555 [Steve Henson] 7556 7557 *) When a certificate request is read in keep a copy of the 7558 original encoding of the signed data and use it when outputing 7559 again. Signatures then use the original encoding rather than 7560 a decoded, encoded version which may cause problems if the 7561 request is improperly encoded. 7562 [Steve Henson] 7563 7564 *) For consistency with other BIO_puts implementations, call 7565 buffer_write(b, ...) directly in buffer_puts instead of calling 7566 BIO_write(b, ...). 7567 7568 In BIO_puts, increment b->num_write as in BIO_write. 7569 [Peter.Sylvester@EdelWeb.fr] 7570 7571 *) Fix BN_mul_word for the case where the word is 0. (We have to use 7572 BN_zero, we may not return a BIGNUM with an array consisting of 7573 words set to zero.) 7574 [Bodo Moeller] 7575 7576 *) Avoid calling abort() from within the library when problems are 7577 detected, except if preprocessor symbols have been defined 7578 (such as REF_CHECK, BN_DEBUG etc.). 7579 [Bodo Moeller] 7580 7581 *) New openssl application 'rsautl'. This utility can be 7582 used for low level RSA operations. DER public key 7583 BIO/fp routines also added. 7584 [Steve Henson] 7585 7586 *) New Configure entry and patches for compiling on QNX 4. 7587 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 7588 7589 *) A demo state-machine implementation was sponsored by 7590 Nuron (http://www.nuron.com/) and is now available in 7591 demos/state_machine. 7592 [Ben Laurie] 7593 7594 *) New options added to the 'dgst' utility for signature 7595 generation and verification. 7596 [Steve Henson] 7597 7598 *) Unrecognized PKCS#7 content types are now handled via a 7599 catch all ASN1_TYPE structure. This allows unsupported 7600 types to be stored as a "blob" and an application can 7601 encode and decode it manually. 7602 [Steve Henson] 7603 7604 *) Fix various signed/unsigned issues to make a_strex.c 7605 compile under VC++. 7606 [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 7607 7608 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 7609 length if passed a buffer. ASN1_INTEGER_to_BN failed 7610 if passed a NULL BN and its argument was negative. 7611 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 7612 7613 *) Modification to PKCS#7 encoding routines to output definite 7614 length encoding. Since currently the whole structures are in 7615 memory there's not real point in using indefinite length 7616 constructed encoding. However if OpenSSL is compiled with 7617 the flag PKCS7_INDEFINITE_ENCODING the old form is used. 7618 [Steve Henson] 7619 7620 *) Added BIO_vprintf() and BIO_vsnprintf(). 7621 [Richard Levitte] 7622 7623 *) Added more prefixes to parse for in the the strings written 7624 through a logging bio, to cover all the levels that are available 7625 through syslog. The prefixes are now: 7626 7627 PANIC, EMERG, EMR => LOG_EMERG 7628 ALERT, ALR => LOG_ALERT 7629 CRIT, CRI => LOG_CRIT 7630 ERROR, ERR => LOG_ERR 7631 WARNING, WARN, WAR => LOG_WARNING 7632 NOTICE, NOTE, NOT => LOG_NOTICE 7633 INFO, INF => LOG_INFO 7634 DEBUG, DBG => LOG_DEBUG 7635 7636 and as before, if none of those prefixes are present at the 7637 beginning of the string, LOG_ERR is chosen. 7638 7639 On Win32, the LOG_* levels are mapped according to this: 7640 7641 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 7642 LOG_WARNING => EVENTLOG_WARNING_TYPE 7643 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 7644 7645 [Richard Levitte] 7646 7647 *) Made it possible to reconfigure with just the configuration 7648 argument "reconf" or "reconfigure". The command line arguments 7649 are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 7650 and are retrieved from there when reconfiguring. 7651 [Richard Levitte] 7652 7653 *) MD4 implemented. 7654 [Assar Westerlund <assar@sics.se>, Richard Levitte] 7655 7656 *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 7657 [Richard Levitte] 7658 7659 *) The obj_dat.pl script was messing up the sorting of object 7660 names. The reason was that it compared the quoted version 7661 of strings as a result "OCSP" > "OCSP Signing" because 7662 " > SPACE. Changed script to store unquoted versions of 7663 names and add quotes on output. It was also omitting some 7664 names from the lookup table if they were given a default 7665 value (that is if SN is missing it is given the same 7666 value as LN and vice versa), these are now added on the 7667 grounds that if an object has a name we should be able to 7668 look it up. Finally added warning output when duplicate 7669 short or long names are found. 7670 [Steve Henson] 7671 7672 *) Changes needed for Tandem NSK. 7673 [Scott Uroff <scott@xypro.com>] 7674 7675 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 7676 RSA_padding_check_SSLv23(), special padding was never detected 7677 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 7678 version rollback attacks was not effective. 7679 7680 In s23_clnt.c, don't use special rollback-attack detection padding 7681 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 7682 client; similarly, in s23_srvr.c, don't do the rollback check if 7683 SSL 2.0 is the only protocol enabled in the server. 7684 [Bodo Moeller] 7685 7686 *) Make it possible to get hexdumps of unprintable data with 'openssl 7687 asn1parse'. By implication, the functions ASN1_parse_dump() and 7688 BIO_dump_indent() are added. 7689 [Richard Levitte] 7690 7691 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 7692 these print out strings and name structures based on various 7693 flags including RFC2253 support and proper handling of 7694 multibyte characters. Added options to the 'x509' utility 7695 to allow the various flags to be set. 7696 [Steve Henson] 7697 7698 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 7699 Also change the functions X509_cmp_current_time() and 7700 X509_gmtime_adj() work with an ASN1_TIME structure, 7701 this will enable certificates using GeneralizedTime in validity 7702 dates to be checked. 7703 [Steve Henson] 7704 7705 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 7706 negative public key encodings) on by default, 7707 NO_NEG_PUBKEY_BUG can be set to disable it. 7708 [Steve Henson] 7709 7710 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 7711 content octets. An i2c_ASN1_OBJECT is unnecessary because 7712 the encoding can be trivially obtained from the structure. 7713 [Steve Henson] 7714 7715 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 7716 not read locks (CRYPTO_r_[un]lock). 7717 [Bodo Moeller] 7718 7719 *) A first attempt at creating official support for shared 7720 libraries through configuration. I've kept it so the 7721 default is static libraries only, and the OpenSSL programs 7722 are always statically linked for now, but there are 7723 preparations for dynamic linking in place. 7724 This has been tested on Linux and Tru64. 7725 [Richard Levitte] 7726 7727 *) Randomness polling function for Win9x, as described in: 7728 Peter Gutmann, Software Generation of Practically Strong 7729 Random Numbers. 7730 [Ulf M��ller] 7731 7732 *) Fix so PRNG is seeded in req if using an already existing 7733 DSA key. 7734 [Steve Henson] 7735 7736 *) New options to smime application. -inform and -outform 7737 allow alternative formats for the S/MIME message including 7738 PEM and DER. The -content option allows the content to be 7739 specified separately. This should allow things like Netscape 7740 form signing output easier to verify. 7741 [Steve Henson] 7742 7743 *) Fix the ASN1 encoding of tags using the 'long form'. 7744 [Steve Henson] 7745 7746 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 7747 STRING types. These convert content octets to and from the 7748 underlying type. The actual tag and length octets are 7749 already assumed to have been read in and checked. These 7750 are needed because all other string types have virtually 7751 identical handling apart from the tag. By having versions 7752 of the ASN1 functions that just operate on content octets 7753 IMPLICIT tagging can be handled properly. It also allows 7754 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 7755 and ASN1_INTEGER are identical apart from the tag. 7756 [Steve Henson] 7757 7758 *) Change the handling of OID objects as follows: 7759 7760 - New object identifiers are inserted in objects.txt, following 7761 the syntax given in objects.README. 7762 - objects.pl is used to process obj_mac.num and create a new 7763 obj_mac.h. 7764 - obj_dat.pl is used to create a new obj_dat.h, using the data in 7765 obj_mac.h. 7766 7767 This is currently kind of a hack, and the perl code in objects.pl 7768 isn't very elegant, but it works as I intended. The simplest way 7769 to check that it worked correctly is to look in obj_dat.h and 7770 check the array nid_objs and make sure the objects haven't moved 7771 around (this is important!). Additions are OK, as well as 7772 consistent name changes. 7773 [Richard Levitte] 7774 7775 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 7776 [Bodo Moeller] 7777 7778 *) Addition of the command line parameter '-rand file' to 'openssl req'. 7779 The given file adds to whatever has already been seeded into the 7780 random pool through the RANDFILE configuration file option or 7781 environment variable, or the default random state file. 7782 [Richard Levitte] 7783 7784 *) mkstack.pl now sorts each macro group into lexical order. 7785 Previously the output order depended on the order the files 7786 appeared in the directory, resulting in needless rewriting 7787 of safestack.h . 7788 [Steve Henson] 7789 7790 *) Patches to make OpenSSL compile under Win32 again. Mostly 7791 work arounds for the VC++ problem that it treats func() as 7792 func(void). Also stripped out the parts of mkdef.pl that 7793 added extra typesafe functions: these no longer exist. 7794 [Steve Henson] 7795 7796 *) Reorganisation of the stack code. The macros are now all 7797 collected in safestack.h . Each macro is defined in terms of 7798 a "stack macro" of the form SKM_<name>(type, a, b). The 7799 DEBUG_SAFESTACK is now handled in terms of function casts, 7800 this has the advantage of retaining type safety without the 7801 use of additional functions. If DEBUG_SAFESTACK is not defined 7802 then the non typesafe macros are used instead. Also modified the 7803 mkstack.pl script to handle the new form. Needs testing to see 7804 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 7805 the default if no major problems. Similar behaviour for ASN1_SET_OF 7806 and PKCS12_STACK_OF. 7807 [Steve Henson] 7808 7809 *) When some versions of IIS use the 'NET' form of private key the 7810 key derivation algorithm is different. Normally MD5(password) is 7811 used as a 128 bit RC4 key. In the modified case 7812 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 7813 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 7814 as the old Netscape_RSA functions except they have an additional 7815 'sgckey' parameter which uses the modified algorithm. Also added 7816 an -sgckey command line option to the rsa utility. Thanks to 7817 Adrian Peck <bertie@ncipher.com> for posting details of the modified 7818 algorithm to openssl-dev. 7819 [Steve Henson] 7820 7821 *) The evp_local.h macros were using 'c.##kname' which resulted in 7822 invalid expansion on some systems (SCO 5.0.5 for example). 7823 Corrected to 'c.kname'. 7824 [Phillip Porch <root@theporch.com>] 7825 7826 *) New X509_get1_email() and X509_REQ_get1_email() functions that return 7827 a STACK of email addresses from a certificate or request, these look 7828 in the subject name and the subject alternative name extensions and 7829 omit any duplicate addresses. 7830 [Steve Henson] 7831 7832 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 7833 This makes DSA verification about 2 % faster. 7834 [Bodo Moeller] 7835 7836 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 7837 (meaning that now 2^5 values will be precomputed, which is only 4 KB 7838 plus overhead for 1024 bit moduli). 7839 This makes exponentiations about 0.5 % faster for 1024 bit 7840 exponents (as measured by "openssl speed rsa2048"). 7841 [Bodo Moeller] 7842 7843 *) Rename memory handling macros to avoid conflicts with other 7844 software: 7845 Malloc => OPENSSL_malloc 7846 Malloc_locked => OPENSSL_malloc_locked 7847 Realloc => OPENSSL_realloc 7848 Free => OPENSSL_free 7849 [Richard Levitte] 7850 7851 *) New function BN_mod_exp_mont_word for small bases (roughly 15% 7852 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 7853 [Bodo Moeller] 7854 7855 *) CygWin32 support. 7856 [John Jarvie <jjarvie@newsguy.com>] 7857 7858 *) The type-safe stack code has been rejigged. It is now only compiled 7859 in when OpenSSL is configured with the DEBUG_SAFESTACK option and 7860 by default all type-specific stack functions are "#define"d back to 7861 standard stack functions. This results in more streamlined output 7862 but retains the type-safety checking possibilities of the original 7863 approach. 7864 [Geoff Thorpe] 7865 7866 *) The STACK code has been cleaned up, and certain type declarations 7867 that didn't make a lot of sense have been brought in line. This has 7868 also involved a cleanup of sorts in safestack.h to more correctly 7869 map type-safe stack functions onto their plain stack counterparts. 7870 This work has also resulted in a variety of "const"ifications of 7871 lots of the code, especially "_cmp" operations which should normally 7872 be prototyped with "const" parameters anyway. 7873 [Geoff Thorpe] 7874 7875 *) When generating bytes for the first time in md_rand.c, 'stir the pool' 7876 by seeding with STATE_SIZE dummy bytes (with zero entropy count). 7877 (The PRNG state consists of two parts, the large pool 'state' and 'md', 7878 where all of 'md' is used each time the PRNG is used, but 'state' 7879 is used only indexed by a cyclic counter. As entropy may not be 7880 well distributed from the beginning, 'md' is important as a 7881 chaining variable. However, the output function chains only half 7882 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 7883 all of 'md', and seeding with STATE_SIZE dummy bytes will result 7884 in all of 'state' being rewritten, with the new values depending 7885 on virtually all of 'md'. This overcomes the 80 bit limitation.) 7886 [Bodo Moeller] 7887 7888 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 7889 the handshake is continued after ssl_verify_cert_chain(); 7890 otherwise, if SSL_VERIFY_NONE is set, remaining error codes 7891 can lead to 'unexplainable' connection aborts later. 7892 [Bodo Moeller; problem tracked down by Lutz Jaenicke] 7893 7894 *) Major EVP API cipher revision. 7895 Add hooks for extra EVP features. This allows various cipher 7896 parameters to be set in the EVP interface. Support added for variable 7897 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 7898 setting of RC2 and RC5 parameters. 7899 7900 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 7901 ciphers. 7902 7903 Remove lots of duplicated code from the EVP library. For example *every* 7904 cipher init() function handles the 'iv' in the same way according to the 7905 cipher mode. They also all do nothing if the 'key' parameter is NULL and 7906 for CFB and OFB modes they zero ctx->num. 7907 7908 New functionality allows removal of S/MIME code RC2 hack. 7909 7910 Most of the routines have the same form and so can be declared in terms 7911 of macros. 7912 7913 By shifting this to the top level EVP_CipherInit() it can be removed from 7914 all individual ciphers. If the cipher wants to handle IVs or keys 7915 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 7916 flags. 7917 7918 Change lots of functions like EVP_EncryptUpdate() to now return a 7919 value: although software versions of the algorithms cannot fail 7920 any installed hardware versions can. 7921 [Steve Henson] 7922 7923 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 7924 this option is set, tolerate broken clients that send the negotiated 7925 protocol version number instead of the requested protocol version 7926 number. 7927 [Bodo Moeller] 7928 7929 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 7930 i.e. non-zero for export ciphersuites, zero otherwise. 7931 Previous versions had this flag inverted, inconsistent with 7932 rsa_tmp_cb (..._TMP_RSA_CB). 7933 [Bodo Moeller; problem reported by Amit Chopra] 7934 7935 *) Add missing DSA library text string. Work around for some IIS 7936 key files with invalid SEQUENCE encoding. 7937 [Steve Henson] 7938 7939 *) Add a document (doc/standards.txt) that list all kinds of standards 7940 and so on that are implemented in OpenSSL. 7941 [Richard Levitte] 7942 7943 *) Enhance c_rehash script. Old version would mishandle certificates 7944 with the same subject name hash and wouldn't handle CRLs at all. 7945 Added -fingerprint option to crl utility, to support new c_rehash 7946 features. 7947 [Steve Henson] 7948 7949 *) Eliminate non-ANSI declarations in crypto.h and stack.h. 7950 [Ulf M��ller] 7951 7952 *) Fix for SSL server purpose checking. Server checking was 7953 rejecting certificates which had extended key usage present 7954 but no ssl client purpose. 7955 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 7956 7957 *) Make PKCS#12 code work with no password. The PKCS#12 spec 7958 is a little unclear about how a blank password is handled. 7959 Since the password in encoded as a BMPString with terminating 7960 double NULL a zero length password would end up as just the 7961 double NULL. However no password at all is different and is 7962 handled differently in the PKCS#12 key generation code. NS 7963 treats a blank password as zero length. MSIE treats it as no 7964 password on export: but it will try both on import. We now do 7965 the same: PKCS12_parse() tries zero length and no password if 7966 the password is set to "" or NULL (NULL is now a valid password: 7967 it wasn't before) as does the pkcs12 application. 7968 [Steve Henson] 7969 7970 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 7971 perror when PEM_read_bio_X509_REQ fails, the error message must 7972 be obtained from the error queue. 7973 [Bodo Moeller] 7974 7975 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 7976 it in ERR_remove_state if appropriate, and change ERR_get_state 7977 accordingly to avoid race conditions (this is necessary because 7978 thread_hash is no longer constant once set). 7979 [Bodo Moeller] 7980 7981 *) Bugfix for linux-elf makefile.one. 7982 [Ulf M��ller] 7983 7984 *) RSA_get_default_method() will now cause a default 7985 RSA_METHOD to be chosen if one doesn't exist already. 7986 Previously this was only set during a call to RSA_new() 7987 or RSA_new_method(NULL) meaning it was possible for 7988 RSA_get_default_method() to return NULL. 7989 [Geoff Thorpe] 7990 7991 *) Added native name translation to the existing DSO code 7992 that will convert (if the flag to do so is set) filenames 7993 that are sufficiently small and have no path information 7994 into a canonical native form. Eg. "blah" converted to 7995 "libblah.so" or "blah.dll" etc. 7996 [Geoff Thorpe] 7997 7998 *) New function ERR_error_string_n(e, buf, len) which is like 7999 ERR_error_string(e, buf), but writes at most 'len' bytes 8000 including the 0 terminator. For ERR_error_string_n, 'buf' 8001 may not be NULL. 8002 [Damien Miller <djm@mindrot.org>, Bodo Moeller] 8003 8004 *) CONF library reworked to become more general. A new CONF 8005 configuration file reader "class" is implemented as well as a 8006 new functions (NCONF_*, for "New CONF") to handle it. The now 8007 old CONF_* functions are still there, but are reimplemented to 8008 work in terms of the new functions. Also, a set of functions 8009 to handle the internal storage of the configuration data is 8010 provided to make it easier to write new configuration file 8011 reader "classes" (I can definitely see something reading a 8012 configuration file in XML format, for example), called _CONF_*, 8013 or "the configuration storage API"... 8014 8015 The new configuration file reading functions are: 8016 8017 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 8018 NCONF_get_section, NCONF_get_string, NCONF_get_numbre 8019 8020 NCONF_default, NCONF_WIN32 8021 8022 NCONF_dump_fp, NCONF_dump_bio 8023 8024 NCONF_default and NCONF_WIN32 are method (or "class") choosers, 8025 NCONF_new creates a new CONF object. This works in the same way 8026 as other interfaces in OpenSSL, like the BIO interface. 8027 NCONF_dump_* dump the internal storage of the configuration file, 8028 which is useful for debugging. All other functions take the same 8029 arguments as the old CONF_* functions wth the exception of the 8030 first that must be a `CONF *' instead of a `LHASH *'. 8031 8032 To make it easer to use the new classes with the old CONF_* functions, 8033 the function CONF_set_default_method is provided. 8034 [Richard Levitte] 8035 8036 *) Add '-tls1' option to 'openssl ciphers', which was already 8037 mentioned in the documentation but had not been implemented. 8038 (This option is not yet really useful because even the additional 8039 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 8040 [Bodo Moeller] 8041 8042 *) Initial DSO code added into libcrypto for letting OpenSSL (and 8043 OpenSSL-based applications) load shared libraries and bind to 8044 them in a portable way. 8045 [Geoff Thorpe, with contributions from Richard Levitte] 8046 8047 Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 8048 8049 *) Make sure _lrotl and _lrotr are only used with MSVC. 8050 8051 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 8052 (the default implementation of RAND_status). 8053 8054 *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 8055 to '-clrext' (= clear extensions), as intended and documented. 8056 [Bodo Moeller; inconsistency pointed out by Michael Attili 8057 <attili@amaxo.com>] 8058 8059 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 8060 was larger than the MD block size. 8061 [Steve Henson, pointed out by Yost William <YostW@tce.com>] 8062 8063 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 8064 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 8065 using the passed key: if the passed key was a private key the result 8066 of X509_print(), for example, would be to print out all the private key 8067 components. 8068 [Steve Henson] 8069 8070 *) des_quad_cksum() byte order bug fix. 8071 [Ulf M��ller, using the problem description in krb4-0.9.7, where 8072 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 8073 8074 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 8075 discouraged. 8076 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 8077 8078 *) For easily testing in shell scripts whether some command 8079 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 8080 returns with exit code 0 iff no command of the given name is available. 8081 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 8082 the output goes to stdout and nothing is printed to stderr. 8083 Additional arguments are always ignored. 8084 8085 Since for each cipher there is a command of the same name, 8086 the 'no-cipher' compilation switches can be tested this way. 8087 8088 ('openssl no-XXX' is not able to detect pseudo-commands such 8089 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 8090 [Bodo Moeller] 8091 8092 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 8093 [Bodo Moeller] 8094 8095 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 8096 is set; it will be thrown away anyway because each handshake creates 8097 its own key. 8098 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 8099 to parameters -- in previous versions (since OpenSSL 0.9.3) the 8100 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 8101 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 8102 [Bodo Moeller] 8103 8104 *) New s_client option -ign_eof: EOF at stdin is ignored, and 8105 'Q' and 'R' lose their special meanings (quit/renegotiate). 8106 This is part of what -quiet does; unlike -quiet, -ign_eof 8107 does not suppress any output. 8108 [Richard Levitte] 8109 8110 *) Add compatibility options to the purpose and trust code. The 8111 purpose X509_PURPOSE_ANY is "any purpose" which automatically 8112 accepts a certificate or CA, this was the previous behaviour, 8113 with all the associated security issues. 8114 8115 X509_TRUST_COMPAT is the old trust behaviour: only and 8116 automatically trust self signed roots in certificate store. A 8117 new trust setting X509_TRUST_DEFAULT is used to specify that 8118 a purpose has no associated trust setting and it should instead 8119 use the value in the default purpose. 8120 [Steve Henson] 8121 8122 *) Fix the PKCS#8 DSA private key code so it decodes keys again 8123 and fix a memory leak. 8124 [Steve Henson] 8125 8126 *) In util/mkerr.pl (which implements 'make errors'), preserve 8127 reason strings from the previous version of the .c file, as 8128 the default to have only downcase letters (and digits) in 8129 automatically generated reasons codes is not always appropriate. 8130 [Bodo Moeller] 8131 8132 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 8133 using strerror. Previously, ERR_reason_error_string() returned 8134 library names as reason strings for SYSerr; but SYSerr is a special 8135 case where small numbers are errno values, not library numbers. 8136 [Bodo Moeller] 8137 8138 *) Add '-dsaparam' option to 'openssl dhparam' application. This 8139 converts DSA parameters into DH parameters. (When creating parameters, 8140 DSA_generate_parameters is used.) 8141 [Bodo Moeller] 8142 8143 *) Include 'length' (recommended exponent length) in C code generated 8144 by 'openssl dhparam -C'. 8145 [Bodo Moeller] 8146 8147 *) The second argument to set_label in perlasm was already being used 8148 so couldn't be used as a "file scope" flag. Moved to third argument 8149 which was free. 8150 [Steve Henson] 8151 8152 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 8153 instead of RAND_bytes for encryption IVs and salts. 8154 [Bodo Moeller] 8155 8156 *) Include RAND_status() into RAND_METHOD instead of implementing 8157 it only for md_rand.c Otherwise replacing the PRNG by calling 8158 RAND_set_rand_method would be impossible. 8159 [Bodo Moeller] 8160 8161 *) Don't let DSA_generate_key() enter an infinite loop if the random 8162 number generation fails. 8163 [Bodo Moeller] 8164 8165 *) New 'rand' application for creating pseudo-random output. 8166 [Bodo Moeller] 8167 8168 *) Added configuration support for Linux/IA64 8169 [Rolf Haberrecker <rolf@suse.de>] 8170 8171 *) Assembler module support for Mingw32. 8172 [Ulf M��ller] 8173 8174 *) Shared library support for HPUX (in shlib/). 8175 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 8176 8177 *) Shared library support for Solaris gcc. 8178 [Lutz Behnke <behnke@trustcenter.de>] 8179 8180 Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 8181 8182 *) PKCS7_encrypt() was adding text MIME headers twice because they 8183 were added manually and by SMIME_crlf_copy(). 8184 [Steve Henson] 8185 8186 *) In bntest.c don't call BN_rand with zero bits argument. 8187 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 8188 8189 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 8190 case was implemented. This caused BN_div_recp() to fail occasionally. 8191 [Ulf M��ller] 8192 8193 *) Add an optional second argument to the set_label() in the perl 8194 assembly language builder. If this argument exists and is set 8195 to 1 it signals that the assembler should use a symbol whose 8196 scope is the entire file, not just the current function. This 8197 is needed with MASM which uses the format label:: for this scope. 8198 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 8199 8200 *) Change the ASN1 types so they are typedefs by default. Before 8201 almost all types were #define'd to ASN1_STRING which was causing 8202 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 8203 for example. 8204 [Steve Henson] 8205 8206 *) Change names of new functions to the new get1/get0 naming 8207 convention: After 'get1', the caller owns a reference count 8208 and has to call ..._free; 'get0' returns a pointer to some 8209 data structure without incrementing reference counters. 8210 (Some of the existing 'get' functions increment a reference 8211 counter, some don't.) 8212 Similarly, 'set1' and 'add1' functions increase reference 8213 counters or duplicate objects. 8214 [Steve Henson] 8215 8216 *) Allow for the possibility of temp RSA key generation failure: 8217 the code used to assume it always worked and crashed on failure. 8218 [Steve Henson] 8219 8220 *) Fix potential buffer overrun problem in BIO_printf(). 8221 [Ulf M��ller, using public domain code by Patrick Powell; problem 8222 pointed out by David Sacerdote <das33@cornell.edu>] 8223 8224 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 8225 RAND_egd() and RAND_status(). In the command line application, 8226 the EGD socket can be specified like a seed file using RANDFILE 8227 or -rand. 8228 [Ulf M��ller] 8229 8230 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 8231 Some CAs (e.g. Verisign) distribute certificates in this form. 8232 [Steve Henson] 8233 8234 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 8235 list to exclude them. This means that no special compilation option 8236 is needed to use anonymous DH: it just needs to be included in the 8237 cipher list. 8238 [Steve Henson] 8239 8240 *) Change the EVP_MD_CTX_type macro so its meaning consistent with 8241 EVP_MD_type. The old functionality is available in a new macro called 8242 EVP_MD_md(). Change code that uses it and update docs. 8243 [Steve Henson] 8244 8245 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 8246 where the 'void *' argument is replaced by a function pointer argument. 8247 Previously 'void *' was abused to point to functions, which works on 8248 many platforms, but is not correct. As these functions are usually 8249 called by macros defined in OpenSSL header files, most source code 8250 should work without changes. 8251 [Richard Levitte] 8252 8253 *) <openssl/opensslconf.h> (which is created by Configure) now contains 8254 sections with information on -D... compiler switches used for 8255 compiling the library so that applications can see them. To enable 8256 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 8257 must be defined. E.g., 8258 #define OPENSSL_ALGORITHM_DEFINES 8259 #include <openssl/opensslconf.h> 8260 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 8261 [Richard Levitte, Ulf and Bodo M��ller] 8262 8263 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 8264 record layer. 8265 [Bodo Moeller] 8266 8267 *) Change the 'other' type in certificate aux info to a STACK_OF 8268 X509_ALGOR. Although not an AlgorithmIdentifier as such it has 8269 the required ASN1 format: arbitrary types determined by an OID. 8270 [Steve Henson] 8271 8272 *) Add some PEM_write_X509_REQ_NEW() functions and a command line 8273 argument to 'req'. This is not because the function is newer or 8274 better than others it just uses the work 'NEW' in the certificate 8275 request header lines. Some software needs this. 8276 [Steve Henson] 8277 8278 *) Reorganise password command line arguments: now passwords can be 8279 obtained from various sources. Delete the PEM_cb function and make 8280 it the default behaviour: i.e. if the callback is NULL and the 8281 usrdata argument is not NULL interpret it as a null terminated pass 8282 phrase. If usrdata and the callback are NULL then the pass phrase 8283 is prompted for as usual. 8284 [Steve Henson] 8285 8286 *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 8287 the support is automatically enabled. The resulting binaries will 8288 autodetect the card and use it if present. 8289 [Ben Laurie and Compaq Inc.] 8290 8291 *) Work around for Netscape hang bug. This sends certificate request 8292 and server done in one record. Since this is perfectly legal in the 8293 SSL/TLS protocol it isn't a "bug" option and is on by default. See 8294 the bugs/SSLv3 entry for more info. 8295 [Steve Henson] 8296 8297 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 8298 [Andy Polyakov] 8299 8300 *) Add -rand argument to smime and pkcs12 applications and read/write 8301 of seed file. 8302 [Steve Henson] 8303 8304 *) New 'passwd' tool for crypt(3) and apr1 password hashes. 8305 [Bodo Moeller] 8306 8307 *) Add command line password options to the remaining applications. 8308 [Steve Henson] 8309 8310 *) Bug fix for BN_div_recp() for numerators with an even number of 8311 bits. 8312 [Ulf M��ller] 8313 8314 *) More tests in bntest.c, and changed test_bn output. 8315 [Ulf M��ller] 8316 8317 *) ./config recognizes MacOS X now. 8318 [Andy Polyakov] 8319 8320 *) Bug fix for BN_div() when the first words of num and divsor are 8321 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 8322 [Ulf M��ller] 8323 8324 *) Add support for various broken PKCS#8 formats, and command line 8325 options to produce them. 8326 [Steve Henson] 8327 8328 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 8329 get temporary BIGNUMs from a BN_CTX. 8330 [Ulf M��ller] 8331 8332 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 8333 for p == 0. 8334 [Ulf M��ller] 8335 8336 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 8337 include a #define from the old name to the new. The original intent 8338 was that statically linked binaries could for example just call 8339 SSLeay_add_all_ciphers() to just add ciphers to the table and not 8340 link with digests. This never worked becayse SSLeay_add_all_digests() 8341 and SSLeay_add_all_ciphers() were in the same source file so calling 8342 one would link with the other. They are now in separate source files. 8343 [Steve Henson] 8344 8345 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 8346 [Steve Henson] 8347 8348 *) Use a less unusual form of the Miller-Rabin primality test (it used 8349 a binary algorithm for exponentiation integrated into the Miller-Rabin 8350 loop, our standard modexp algorithms are faster). 8351 [Bodo Moeller] 8352 8353 *) Support for the EBCDIC character set completed. 8354 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 8355 8356 *) Source code cleanups: use const where appropriate, eliminate casts, 8357 use void * instead of char * in lhash. 8358 [Ulf M��ller] 8359 8360 *) Bugfix: ssl3_send_server_key_exchange was not restartable 8361 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 8362 this the server could overwrite ephemeral keys that the client 8363 has already seen). 8364 [Bodo Moeller] 8365 8366 *) Turn DSA_is_prime into a macro that calls BN_is_prime, 8367 using 50 iterations of the Rabin-Miller test. 8368 8369 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 8370 iterations of the Rabin-Miller test as required by the appendix 8371 to FIPS PUB 186[-1]) instead of DSA_is_prime. 8372 As BN_is_prime_fasttest includes trial division, DSA parameter 8373 generation becomes much faster. 8374 8375 This implies a change for the callback functions in DSA_is_prime 8376 and DSA_generate_parameters: The callback function is called once 8377 for each positive witness in the Rabin-Miller test, not just 8378 occasionally in the inner loop; and the parameters to the 8379 callback function now provide an iteration count for the outer 8380 loop rather than for the current invocation of the inner loop. 8381 DSA_generate_parameters additionally can call the callback 8382 function with an 'iteration count' of -1, meaning that a 8383 candidate has passed the trial division test (when q is generated 8384 from an application-provided seed, trial division is skipped). 8385 [Bodo Moeller] 8386 8387 *) New function BN_is_prime_fasttest that optionally does trial 8388 division before starting the Rabin-Miller test and has 8389 an additional BN_CTX * argument (whereas BN_is_prime always 8390 has to allocate at least one BN_CTX). 8391 'callback(1, -1, cb_arg)' is called when a number has passed the 8392 trial division stage. 8393 [Bodo Moeller] 8394 8395 *) Fix for bug in CRL encoding. The validity dates weren't being handled 8396 as ASN1_TIME. 8397 [Steve Henson] 8398 8399 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 8400 [Steve Henson] 8401 8402 *) New function BN_pseudo_rand(). 8403 [Ulf M��ller] 8404 8405 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 8406 bignum version of BN_from_montgomery() with the working code from 8407 SSLeay 0.9.0 (the word based version is faster anyway), and clean up 8408 the comments. 8409 [Ulf M��ller] 8410 8411 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 8412 made it impossible to use the same SSL_SESSION data structure in 8413 SSL2 clients in multiple threads. 8414 [Bodo Moeller] 8415 8416 *) The return value of RAND_load_file() no longer counts bytes obtained 8417 by stat(). RAND_load_file(..., -1) is new and uses the complete file 8418 to seed the PRNG (previously an explicit byte count was required). 8419 [Ulf M��ller, Bodo M��ller] 8420 8421 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 8422 used (char *) instead of (void *) and had casts all over the place. 8423 [Steve Henson] 8424 8425 *) Make BN_generate_prime() return NULL on error if ret!=NULL. 8426 [Ulf M��ller] 8427 8428 *) Retain source code compatibility for BN_prime_checks macro: 8429 BN_is_prime(..., BN_prime_checks, ...) now uses 8430 BN_prime_checks_for_size to determine the appropriate number of 8431 Rabin-Miller iterations. 8432 [Ulf M��ller] 8433 8434 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 8435 DH_CHECK_P_NOT_SAFE_PRIME. 8436 (Check if this is true? OpenPGP calls them "strong".) 8437 [Ulf M��ller] 8438 8439 *) Merge the functionality of "dh" and "gendh" programs into a new program 8440 "dhparam". The old programs are retained for now but will handle DH keys 8441 (instead of parameters) in future. 8442 [Steve Henson] 8443 8444 *) Make the ciphers, s_server and s_client programs check the return values 8445 when a new cipher list is set. 8446 [Steve Henson] 8447 8448 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 8449 ciphers. Before when the 56bit ciphers were enabled the sorting was 8450 wrong. 8451 8452 The syntax for the cipher sorting has been extended to support sorting by 8453 cipher-strength (using the strength_bits hard coded in the tables). 8454 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 8455 8456 Fix a bug in the cipher-command parser: when supplying a cipher command 8457 string with an "undefined" symbol (neither command nor alphanumeric 8458 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 8459 an error is flagged. 8460 8461 Due to the strength-sorting extension, the code of the 8462 ssl_create_cipher_list() function was completely rearranged. I hope that 8463 the readability was also increased :-) 8464 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 8465 8466 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 8467 for the first serial number and places 2 in the serial number file. This 8468 avoids problems when the root CA is created with serial number zero and 8469 the first user certificate has the same issuer name and serial number 8470 as the root CA. 8471 [Steve Henson] 8472 8473 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 8474 the new code. Add documentation for this stuff. 8475 [Steve Henson] 8476 8477 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 8478 X509_*() to X509at_*() on the grounds that they don't handle X509 8479 structures and behave in an analagous way to the X509v3 functions: 8480 they shouldn't be called directly but wrapper functions should be used 8481 instead. 8482 8483 So we also now have some wrapper functions that call the X509at functions 8484 when passed certificate requests. (TO DO: similar things can be done with 8485 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 8486 things. Some of these need some d2i or i2d and print functionality 8487 because they handle more complex structures.) 8488 [Steve Henson] 8489 8490 *) Add missing #ifndefs that caused missing symbols when building libssl 8491 as a shared library without RSA. Use #ifndef NO_SSL2 instead of 8492 NO_RSA in ssl/s2*.c. 8493 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M��ller] 8494 8495 *) Precautions against using the PRNG uninitialized: RAND_bytes() now 8496 has a return value which indicates the quality of the random data 8497 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 8498 error queue. New function RAND_pseudo_bytes() generates output that is 8499 guaranteed to be unique but not unpredictable. RAND_add is like 8500 RAND_seed, but takes an extra argument for an entropy estimate 8501 (RAND_seed always assumes full entropy). 8502 [Ulf M��ller] 8503 8504 *) Do more iterations of Rabin-Miller probable prime test (specifically, 8505 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 8506 instead of only 2 for all lengths; see BN_prime_checks_for_size definition 8507 in crypto/bn/bn_prime.c for the complete table). This guarantees a 8508 false-positive rate of at most 2^-80 for random input. 8509 [Bodo Moeller] 8510 8511 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 8512 [Bodo Moeller] 8513 8514 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 8515 in the 0.9.5 release), this returns the chain 8516 from an X509_CTX structure with a dup of the stack and all 8517 the X509 reference counts upped: so the stack will exist 8518 after X509_CTX_cleanup() has been called. Modify pkcs12.c 8519 to use this. 8520 8521 Also make SSL_SESSION_print() print out the verify return 8522 code. 8523 [Steve Henson] 8524 8525 *) Add manpage for the pkcs12 command. Also change the default 8526 behaviour so MAC iteration counts are used unless the new 8527 -nomaciter option is used. This improves file security and 8528 only older versions of MSIE (4.0 for example) need it. 8529 [Steve Henson] 8530 8531 *) Honor the no-xxx Configure options when creating .DEF files. 8532 [Ulf M��ller] 8533 8534 *) Add PKCS#10 attributes to field table: challengePassword, 8535 unstructuredName and unstructuredAddress. These are taken from 8536 draft PKCS#9 v2.0 but are compatible with v1.2 provided no 8537 international characters are used. 8538 8539 More changes to X509_ATTRIBUTE code: allow the setting of types 8540 based on strings. Remove the 'loc' parameter when adding 8541 attributes because these will be a SET OF encoding which is sorted 8542 in ASN1 order. 8543 [Steve Henson] 8544 8545 *) Initial changes to the 'req' utility to allow request generation 8546 automation. This will allow an application to just generate a template 8547 file containing all the field values and have req construct the 8548 request. 8549 8550 Initial support for X509_ATTRIBUTE handling. Stacks of these are 8551 used all over the place including certificate requests and PKCS#7 8552 structures. They are currently handled manually where necessary with 8553 some primitive wrappers for PKCS#7. The new functions behave in a 8554 manner analogous to the X509 extension functions: they allow 8555 attributes to be looked up by NID and added. 8556 8557 Later something similar to the X509V3 code would be desirable to 8558 automatically handle the encoding, decoding and printing of the 8559 more complex types. The string types like challengePassword can 8560 be handled by the string table functions. 8561 8562 Also modified the multi byte string table handling. Now there is 8563 a 'global mask' which masks out certain types. The table itself 8564 can use the flag STABLE_NO_MASK to ignore the mask setting: this 8565 is useful when for example there is only one permissible type 8566 (as in countryName) and using the mask might result in no valid 8567 types at all. 8568 [Steve Henson] 8569 8570 *) Clean up 'Finished' handling, and add functions SSL_get_finished and 8571 SSL_get_peer_finished to allow applications to obtain the latest 8572 Finished messages sent to the peer or expected from the peer, 8573 respectively. (SSL_get_peer_finished is usually the Finished message 8574 actually received from the peer, otherwise the protocol will be aborted.) 8575 8576 As the Finished message are message digests of the complete handshake 8577 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 8578 be used for external authentication procedures when the authentication 8579 provided by SSL/TLS is not desired or is not enough. 8580 [Bodo Moeller] 8581 8582 *) Enhanced support for Alpha Linux is added. Now ./config checks if 8583 the host supports BWX extension and if Compaq C is present on the 8584 $PATH. Just exploiting of the BWX extension results in 20-30% 8585 performance kick for some algorithms, e.g. DES and RC4 to mention 8586 a couple. Compaq C in turn generates ~20% faster code for MD5 and 8587 SHA1. 8588 [Andy Polyakov] 8589 8590 *) Add support for MS "fast SGC". This is arguably a violation of the 8591 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 8592 weak crypto and after checking the certificate is SGC a second one 8593 with strong crypto. MS SGC stops the first handshake after receiving 8594 the server certificate message and sends a second client hello. Since 8595 a server will typically do all the time consuming operations before 8596 expecting any further messages from the client (server key exchange 8597 is the most expensive) there is little difference between the two. 8598 8599 To get OpenSSL to support MS SGC we have to permit a second client 8600 hello message after we have sent server done. In addition we have to 8601 reset the MAC if we do get this second client hello. 8602 [Steve Henson] 8603 8604 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 8605 if a DER encoded private key is RSA or DSA traditional format. Changed 8606 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 8607 format DER encoded private key. Newer code should use PKCS#8 format which 8608 has the key type encoded in the ASN1 structure. Added DER private key 8609 support to pkcs8 application. 8610 [Steve Henson] 8611 8612 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 8613 ciphersuites has been selected (as required by the SSL 3/TLS 1 8614 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 8615 is set, we interpret this as a request to violate the specification 8616 (the worst that can happen is a handshake failure, and 'correct' 8617 behaviour would result in a handshake failure anyway). 8618 [Bodo Moeller] 8619 8620 *) In SSL_CTX_add_session, take into account that there might be multiple 8621 SSL_SESSION structures with the same session ID (e.g. when two threads 8622 concurrently obtain them from an external cache). 8623 The internal cache can handle only one SSL_SESSION with a given ID, 8624 so if there's a conflict, we now throw out the old one to achieve 8625 consistency. 8626 [Bodo Moeller] 8627 8628 *) Add OIDs for idea and blowfish in CBC mode. This will allow both 8629 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 8630 some routines that use cipher OIDs: some ciphers do not have OIDs 8631 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 8632 example. 8633 [Steve Henson] 8634 8635 *) Simplify the trust setting structure and code. Now we just have 8636 two sequences of OIDs for trusted and rejected settings. These will 8637 typically have values the same as the extended key usage extension 8638 and any application specific purposes. 8639 8640 The trust checking code now has a default behaviour: it will just 8641 check for an object with the same NID as the passed id. Functions can 8642 be provided to override either the default behaviour or the behaviour 8643 for a given id. SSL client, server and email already have functions 8644 in place for compatibility: they check the NID and also return "trusted" 8645 if the certificate is self signed. 8646 [Steve Henson] 8647 8648 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 8649 traditional format into an EVP_PKEY structure. 8650 [Steve Henson] 8651 8652 *) Add a password callback function PEM_cb() which either prompts for 8653 a password if usr_data is NULL or otherwise assumes it is a null 8654 terminated password. Allow passwords to be passed on command line 8655 environment or config files in a few more utilities. 8656 [Steve Henson] 8657 8658 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 8659 keys. Add some short names for PKCS#8 PBE algorithms and allow them 8660 to be specified on the command line for the pkcs8 and pkcs12 utilities. 8661 Update documentation. 8662 [Steve Henson] 8663 8664 *) Support for ASN1 "NULL" type. This could be handled before by using 8665 ASN1_TYPE but there wasn't any function that would try to read a NULL 8666 and produce an error if it couldn't. For compatibility we also have 8667 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 8668 don't allocate anything because they don't need to. 8669 [Steve Henson] 8670 8671 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 8672 for details. 8673 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 8674 8675 *) Rebuild of the memory allocation routines used by OpenSSL code and 8676 possibly others as well. The purpose is to make an interface that 8677 provide hooks so anyone can build a separate set of allocation and 8678 deallocation routines to be used by OpenSSL, for example memory 8679 pool implementations, or something else, which was previously hard 8680 since Malloc(), Realloc() and Free() were defined as macros having 8681 the values malloc, realloc and free, respectively (except for Win32 8682 compilations). The same is provided for memory debugging code. 8683 OpenSSL already comes with functionality to find memory leaks, but 8684 this gives people a chance to debug other memory problems. 8685 8686 With these changes, a new set of functions and macros have appeared: 8687 8688 CRYPTO_set_mem_debug_functions() [F] 8689 CRYPTO_get_mem_debug_functions() [F] 8690 CRYPTO_dbg_set_options() [F] 8691 CRYPTO_dbg_get_options() [F] 8692 CRYPTO_malloc_debug_init() [M] 8693 8694 The memory debug functions are NULL by default, unless the library 8695 is compiled with CRYPTO_MDEBUG or friends is defined. If someone 8696 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 8697 gives the standard debugging functions that come with OpenSSL) or 8698 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 8699 provided by the library user) must be used. When the standard 8700 debugging functions are used, CRYPTO_dbg_set_options can be used to 8701 request additional information: 8702 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 8703 the CRYPTO_MDEBUG_xxx macro when compiling the library. 8704 8705 Also, things like CRYPTO_set_mem_functions will always give the 8706 expected result (the new set of functions is used for allocation 8707 and deallocation) at all times, regardless of platform and compiler 8708 options. 8709 8710 To finish it up, some functions that were never use in any other 8711 way than through macros have a new API and new semantic: 8712 8713 CRYPTO_dbg_malloc() 8714 CRYPTO_dbg_realloc() 8715 CRYPTO_dbg_free() 8716 8717 All macros of value have retained their old syntax. 8718 [Richard Levitte and Bodo Moeller] 8719 8720 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 8721 ordering of SMIMECapabilities wasn't in "strength order" and there 8722 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 8723 algorithm. 8724 [Steve Henson] 8725 8726 *) Some ASN1 types with illegal zero length encoding (INTEGER, 8727 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 8728 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 8729 8730 *) Merge in my S/MIME library for OpenSSL. This provides a simple 8731 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 8732 functionality to handle multipart/signed properly) and a utility 8733 called 'smime' to call all this stuff. This is based on code I 8734 originally wrote for Celo who have kindly allowed it to be 8735 included in OpenSSL. 8736 [Steve Henson] 8737 8738 *) Add variants des_set_key_checked and des_set_key_unchecked of 8739 des_set_key (aka des_key_sched). Global variable des_check_key 8740 decides which of these is called by des_set_key; this way 8741 des_check_key behaves as it always did, but applications and 8742 the library itself, which was buggy for des_check_key == 1, 8743 have a cleaner way to pick the version they need. 8744 [Bodo Moeller] 8745 8746 *) New function PKCS12_newpass() which changes the password of a 8747 PKCS12 structure. 8748 [Steve Henson] 8749 8750 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 8751 dynamic mix. In both cases the ids can be used as an index into the 8752 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 8753 functions so they accept a list of the field values and the 8754 application doesn't need to directly manipulate the X509_TRUST 8755 structure. 8756 [Steve Henson] 8757 8758 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 8759 need initialising. 8760 [Steve Henson] 8761 8762 *) Modify the way the V3 extension code looks up extensions. This now 8763 works in a similar way to the object code: we have some "standard" 8764 extensions in a static table which is searched with OBJ_bsearch() 8765 and the application can add dynamic ones if needed. The file 8766 crypto/x509v3/ext_dat.h now has the info: this file needs to be 8767 updated whenever a new extension is added to the core code and kept 8768 in ext_nid order. There is a simple program 'tabtest.c' which checks 8769 this. New extensions are not added too often so this file can readily 8770 be maintained manually. 8771 8772 There are two big advantages in doing things this way. The extensions 8773 can be looked up immediately and no longer need to be "added" using 8774 X509V3_add_standard_extensions(): this function now does nothing. 8775 [Side note: I get *lots* of email saying the extension code doesn't 8776 work because people forget to call this function] 8777 Also no dynamic allocation is done unless new extensions are added: 8778 so if we don't add custom extensions there is no need to call 8779 X509V3_EXT_cleanup(). 8780 [Steve Henson] 8781 8782 *) Modify enc utility's salting as follows: make salting the default. Add a 8783 magic header, so unsalted files fail gracefully instead of just decrypting 8784 to garbage. This is because not salting is a big security hole, so people 8785 should be discouraged from doing it. 8786 [Ben Laurie] 8787 8788 *) Fixes and enhancements to the 'x509' utility. It allowed a message 8789 digest to be passed on the command line but it only used this 8790 parameter when signing a certificate. Modified so all relevant 8791 operations are affected by the digest parameter including the 8792 -fingerprint and -x509toreq options. Also -x509toreq choked if a 8793 DSA key was used because it didn't fix the digest. 8794 [Steve Henson] 8795 8796 *) Initial certificate chain verify code. Currently tests the untrusted 8797 certificates for consistency with the verify purpose (which is set 8798 when the X509_STORE_CTX structure is set up) and checks the pathlength. 8799 8800 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 8801 this is because it will reject chains with invalid extensions whereas 8802 every previous version of OpenSSL and SSLeay made no checks at all. 8803 8804 Trust code: checks the root CA for the relevant trust settings. Trust 8805 settings have an initial value consistent with the verify purpose: e.g. 8806 if the verify purpose is for SSL client use it expects the CA to be 8807 trusted for SSL client use. However the default value can be changed to 8808 permit custom trust settings: one example of this would be to only trust 8809 certificates from a specific "secure" set of CAs. 8810 8811 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 8812 which should be used for version portability: especially since the 8813 verify structure is likely to change more often now. 8814 8815 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 8816 to set them. If not set then assume SSL clients will verify SSL servers 8817 and vice versa. 8818 8819 Two new options to the verify program: -untrusted allows a set of 8820 untrusted certificates to be passed in and -purpose which sets the 8821 intended purpose of the certificate. If a purpose is set then the 8822 new chain verify code is used to check extension consistency. 8823 [Steve Henson] 8824 8825 *) Support for the authority information access extension. 8826 [Steve Henson] 8827 8828 *) Modify RSA and DSA PEM read routines to transparently handle 8829 PKCS#8 format private keys. New *_PUBKEY_* functions that handle 8830 public keys in a format compatible with certificate 8831 SubjectPublicKeyInfo structures. Unfortunately there were already 8832 functions called *_PublicKey_* which used various odd formats so 8833 these are retained for compatibility: however the DSA variants were 8834 never in a public release so they have been deleted. Changed dsa/rsa 8835 utilities to handle the new format: note no releases ever handled public 8836 keys so we should be OK. 8837 8838 The primary motivation for this change is to avoid the same fiasco 8839 that dogs private keys: there are several incompatible private key 8840 formats some of which are standard and some OpenSSL specific and 8841 require various evil hacks to allow partial transparent handling and 8842 even then it doesn't work with DER formats. Given the option anything 8843 other than PKCS#8 should be dumped: but the other formats have to 8844 stay in the name of compatibility. 8845 8846 With public keys and the benefit of hindsight one standard format 8847 is used which works with EVP_PKEY, RSA or DSA structures: though 8848 it clearly returns an error if you try to read the wrong kind of key. 8849 8850 Added a -pubkey option to the 'x509' utility to output the public key. 8851 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 8852 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 8853 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 8854 that do the same as the EVP_PKEY_assign_*() except they up the 8855 reference count of the added key (they don't "swallow" the 8856 supplied key). 8857 [Steve Henson] 8858 8859 *) Fixes to crypto/x509/by_file.c the code to read in certificates and 8860 CRLs would fail if the file contained no certificates or no CRLs: 8861 added a new function to read in both types and return the number 8862 read: this means that if none are read it will be an error. The 8863 DER versions of the certificate and CRL reader would always fail 8864 because it isn't possible to mix certificates and CRLs in DER format 8865 without choking one or the other routine. Changed this to just read 8866 a certificate: this is the best we can do. Also modified the code 8867 in apps/verify.c to take notice of return codes: it was previously 8868 attempting to read in certificates from NULL pointers and ignoring 8869 any errors: this is one reason why the cert and CRL reader seemed 8870 to work. It doesn't check return codes from the default certificate 8871 routines: these may well fail if the certificates aren't installed. 8872 [Steve Henson] 8873 8874 *) Code to support otherName option in GeneralName. 8875 [Steve Henson] 8876 8877 *) First update to verify code. Change the verify utility 8878 so it warns if it is passed a self signed certificate: 8879 for consistency with the normal behaviour. X509_verify 8880 has been modified to it will now verify a self signed 8881 certificate if *exactly* the same certificate appears 8882 in the store: it was previously impossible to trust a 8883 single self signed certificate. This means that: 8884 openssl verify ss.pem 8885 now gives a warning about a self signed certificate but 8886 openssl verify -CAfile ss.pem ss.pem 8887 is OK. 8888 [Steve Henson] 8889 8890 *) For servers, store verify_result in SSL_SESSION data structure 8891 (and add it to external session representation). 8892 This is needed when client certificate verifications fails, 8893 but an application-provided verification callback (set by 8894 SSL_CTX_set_cert_verify_callback) allows accepting the session 8895 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 8896 but returns 1): When the session is reused, we have to set 8897 ssl->verify_result to the appropriate error code to avoid 8898 security holes. 8899 [Bodo Moeller, problem pointed out by Lutz Jaenicke] 8900 8901 *) Fix a bug in the new PKCS#7 code: it didn't consider the 8902 case in PKCS7_dataInit() where the signed PKCS7 structure 8903 didn't contain any existing data because it was being created. 8904 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 8905 8906 *) Add a salt to the key derivation routines in enc.c. This 8907 forms the first 8 bytes of the encrypted file. Also add a 8908 -S option to allow a salt to be input on the command line. 8909 [Steve Henson] 8910 8911 *) New function X509_cmp(). Oddly enough there wasn't a function 8912 to compare two certificates. We do this by working out the SHA1 8913 hash and comparing that. X509_cmp() will be needed by the trust 8914 code. 8915 [Steve Henson] 8916 8917 *) SSL_get1_session() is like SSL_get_session(), but increments 8918 the reference count in the SSL_SESSION returned. 8919 [Geoff Thorpe <geoff@eu.c2.net>] 8920 8921 *) Fix for 'req': it was adding a null to request attributes. 8922 Also change the X509_LOOKUP and X509_INFO code to handle 8923 certificate auxiliary information. 8924 [Steve Henson] 8925 8926 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 8927 the 'enc' command. 8928 [Steve Henson] 8929 8930 *) Add the possibility to add extra information to the memory leak 8931 detecting output, to form tracebacks, showing from where each 8932 allocation was originated: CRYPTO_push_info("constant string") adds 8933 the string plus current file name and line number to a per-thread 8934 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 8935 is like calling CYRPTO_pop_info() until the stack is empty. 8936 Also updated memory leak detection code to be multi-thread-safe. 8937 [Richard Levitte] 8938 8939 *) Add options -text and -noout to pkcs7 utility and delete the 8940 encryption options which never did anything. Update docs. 8941 [Steve Henson] 8942 8943 *) Add options to some of the utilities to allow the pass phrase 8944 to be included on either the command line (not recommended on 8945 OSes like Unix) or read from the environment. Update the 8946 manpages and fix a few bugs. 8947 [Steve Henson] 8948 8949 *) Add a few manpages for some of the openssl commands. 8950 [Steve Henson] 8951 8952 *) Fix the -revoke option in ca. It was freeing up memory twice, 8953 leaking and not finding already revoked certificates. 8954 [Steve Henson] 8955 8956 *) Extensive changes to support certificate auxiliary information. 8957 This involves the use of X509_CERT_AUX structure and X509_AUX 8958 functions. An X509_AUX function such as PEM_read_X509_AUX() 8959 can still read in a certificate file in the usual way but it 8960 will also read in any additional "auxiliary information". By 8961 doing things this way a fair degree of compatibility can be 8962 retained: existing certificates can have this information added 8963 using the new 'x509' options. 8964 8965 Current auxiliary information includes an "alias" and some trust 8966 settings. The trust settings will ultimately be used in enhanced 8967 certificate chain verification routines: currently a certificate 8968 can only be trusted if it is self signed and then it is trusted 8969 for all purposes. 8970 [Steve Henson] 8971 8972 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 8973 The problem was that one of the replacement routines had not been working 8974 since SSLeay releases. For now the offending routine has been replaced 8975 with non-optimised assembler. Even so, this now gives around 95% 8976 performance improvement for 1024 bit RSA signs. 8977 [Mark Cox] 8978 8979 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 8980 handling. Most clients have the effective key size in bits equal to 8981 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 8982 A few however don't do this and instead use the size of the decrypted key 8983 to determine the RC2 key length and the AlgorithmIdentifier to determine 8984 the effective key length. In this case the effective key length can still 8985 be 40 bits but the key length can be 168 bits for example. This is fixed 8986 by manually forcing an RC2 key into the EVP_PKEY structure because the 8987 EVP code can't currently handle unusual RC2 key sizes: it always assumes 8988 the key length and effective key length are equal. 8989 [Steve Henson] 8990 8991 *) Add a bunch of functions that should simplify the creation of 8992 X509_NAME structures. Now you should be able to do: 8993 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 8994 and have it automatically work out the correct field type and fill in 8995 the structures. The more adventurous can try: 8996 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 8997 and it will (hopefully) work out the correct multibyte encoding. 8998 [Steve Henson] 8999 9000 *) Change the 'req' utility to use the new field handling and multibyte 9001 copy routines. Before the DN field creation was handled in an ad hoc 9002 way in req, ca, and x509 which was rather broken and didn't support 9003 BMPStrings or UTF8Strings. Since some software doesn't implement 9004 BMPStrings or UTF8Strings yet, they can be enabled using the config file 9005 using the dirstring_type option. See the new comment in the default 9006 openssl.cnf for more info. 9007 [Steve Henson] 9008 9009 *) Make crypto/rand/md_rand.c more robust: 9010 - Assure unique random numbers after fork(). 9011 - Make sure that concurrent threads access the global counter and 9012 md serializably so that we never lose entropy in them 9013 or use exactly the same state in multiple threads. 9014 Access to the large state is not always serializable because 9015 the additional locking could be a performance killer, and 9016 md should be large enough anyway. 9017 [Bodo Moeller] 9018 9019 *) New file apps/app_rand.c with commonly needed functionality 9020 for handling the random seed file. 9021 9022 Use the random seed file in some applications that previously did not: 9023 ca, 9024 dsaparam -genkey (which also ignored its '-rand' option), 9025 s_client, 9026 s_server, 9027 x509 (when signing). 9028 Except on systems with /dev/urandom, it is crucial to have a random 9029 seed file at least for key creation, DSA signing, and for DH exchanges; 9030 for RSA signatures we could do without one. 9031 9032 gendh and gendsa (unlike genrsa) used to read only the first byte 9033 of each file listed in the '-rand' option. The function as previously 9034 found in genrsa is now in app_rand.c and is used by all programs 9035 that support '-rand'. 9036 [Bodo Moeller] 9037 9038 *) In RAND_write_file, use mode 0600 for creating files; 9039 don't just chmod when it may be too late. 9040 [Bodo Moeller] 9041 9042 *) Report an error from X509_STORE_load_locations 9043 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 9044 [Bill Perry] 9045 9046 *) New function ASN1_mbstring_copy() this copies a string in either 9047 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 9048 into an ASN1_STRING type. A mask of permissible types is passed 9049 and it chooses the "minimal" type to use or an error if not type 9050 is suitable. 9051 [Steve Henson] 9052 9053 *) Add function equivalents to the various macros in asn1.h. The old 9054 macros are retained with an M_ prefix. Code inside the library can 9055 use the M_ macros. External code (including the openssl utility) 9056 should *NOT* in order to be "shared library friendly". 9057 [Steve Henson] 9058 9059 *) Add various functions that can check a certificate's extensions 9060 to see if it usable for various purposes such as SSL client, 9061 server or S/MIME and CAs of these types. This is currently 9062 VERY EXPERIMENTAL but will ultimately be used for certificate chain 9063 verification. Also added a -purpose flag to x509 utility to 9064 print out all the purposes. 9065 [Steve Henson] 9066 9067 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 9068 functions. 9069 [Steve Henson] 9070 9071 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 9072 for, obtain and decode and extension and obtain its critical flag. 9073 This allows all the necessary extension code to be handled in a 9074 single function call. 9075 [Steve Henson] 9076 9077 *) RC4 tune-up featuring 30-40% performance improvement on most RISC 9078 platforms. See crypto/rc4/rc4_enc.c for further details. 9079 [Andy Polyakov] 9080 9081 *) New -noout option to asn1parse. This causes no output to be produced 9082 its main use is when combined with -strparse and -out to extract data 9083 from a file (which may not be in ASN.1 format). 9084 [Steve Henson] 9085 9086 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 9087 when producing the local key id. 9088 [Richard Levitte <levitte@stacken.kth.se>] 9089 9090 *) New option -dhparam in s_server. This allows a DH parameter file to be 9091 stated explicitly. If it is not stated then it tries the first server 9092 certificate file. The previous behaviour hard coded the filename 9093 "server.pem". 9094 [Steve Henson] 9095 9096 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 9097 a public key to be input or output. For example: 9098 openssl rsa -in key.pem -pubout -out pubkey.pem 9099 Also added necessary DSA public key functions to handle this. 9100 [Steve Henson] 9101 9102 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 9103 in the message. This was handled by allowing 9104 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 9105 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 9106 9107 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 9108 to the end of the strings whereas this didn't. This would cause problems 9109 if strings read with d2i_ASN1_bytes() were later modified. 9110 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 9111 9112 *) Fix for base64 decode bug. When a base64 bio reads only one line of 9113 data and it contains EOF it will end up returning an error. This is 9114 caused by input 46 bytes long. The cause is due to the way base64 9115 BIOs find the start of base64 encoded data. They do this by trying a 9116 trial decode on each line until they find one that works. When they 9117 do a flag is set and it starts again knowing it can pass all the 9118 data directly through the decoder. Unfortunately it doesn't reset 9119 the context it uses. This means that if EOF is reached an attempt 9120 is made to pass two EOFs through the context and this causes the 9121 resulting error. This can also cause other problems as well. As is 9122 usual with these problems it takes *ages* to find and the fix is 9123 trivial: move one line. 9124 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 9125 9126 *) Ugly workaround to get s_client and s_server working under Windows. The 9127 old code wouldn't work because it needed to select() on sockets and the 9128 tty (for keypresses and to see if data could be written). Win32 only 9129 supports select() on sockets so we select() with a 1s timeout on the 9130 sockets and then see if any characters are waiting to be read, if none 9131 are present then we retry, we also assume we can always write data to 9132 the tty. This isn't nice because the code then blocks until we've 9133 received a complete line of data and it is effectively polling the 9134 keyboard at 1s intervals: however it's quite a bit better than not 9135 working at all :-) A dedicated Windows application might handle this 9136 with an event loop for example. 9137 [Steve Henson] 9138 9139 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 9140 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 9141 will be called when RSA_sign() and RSA_verify() are used. This is useful 9142 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 9143 For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 9144 should *not* be used: RSA_sign() and RSA_verify() must be used instead. 9145 This necessitated the support of an extra signature type NID_md5_sha1 9146 for SSL signatures and modifications to the SSL library to use it instead 9147 of calling RSA_public_decrypt() and RSA_private_encrypt(). 9148 [Steve Henson] 9149 9150 *) Add new -verify -CAfile and -CApath options to the crl program, these 9151 will lookup a CRL issuers certificate and verify the signature in a 9152 similar way to the verify program. Tidy up the crl program so it 9153 no longer accesses structures directly. Make the ASN1 CRL parsing a bit 9154 less strict. It will now permit CRL extensions even if it is not 9155 a V2 CRL: this will allow it to tolerate some broken CRLs. 9156 [Steve Henson] 9157 9158 *) Initialize all non-automatic variables each time one of the openssl 9159 sub-programs is started (this is necessary as they may be started 9160 multiple times from the "OpenSSL>" prompt). 9161 [Lennart Bang, Bodo Moeller] 9162 9163 *) Preliminary compilation option RSA_NULL which disables RSA crypto without 9164 removing all other RSA functionality (this is what NO_RSA does). This 9165 is so (for example) those in the US can disable those operations covered 9166 by the RSA patent while allowing storage and parsing of RSA keys and RSA 9167 key generation. 9168 [Steve Henson] 9169 9170 *) Non-copying interface to BIO pairs. 9171 (still largely untested) 9172 [Bodo Moeller] 9173 9174 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 9175 ASCII string. This was handled independently in various places before. 9176 [Steve Henson] 9177 9178 *) New functions UTF8_getc() and UTF8_putc() that parse and generate 9179 UTF8 strings a character at a time. 9180 [Steve Henson] 9181 9182 *) Use client_version from client hello to select the protocol 9183 (s23_srvr.c) and for RSA client key exchange verification 9184 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 9185 [Bodo Moeller] 9186 9187 *) Add various utility functions to handle SPKACs, these were previously 9188 handled by poking round in the structure internals. Added new function 9189 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 9190 print, verify and generate SPKACs. Based on an original idea from 9191 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 9192 [Steve Henson] 9193 9194 *) RIPEMD160 is operational on all platforms and is back in 'make test'. 9195 [Andy Polyakov] 9196 9197 *) Allow the config file extension section to be overwritten on the 9198 command line. Based on an original idea from Massimiliano Pala 9199 <madwolf@comune.modena.it>. The new option is called -extensions 9200 and can be applied to ca, req and x509. Also -reqexts to override 9201 the request extensions in req and -crlexts to override the crl extensions 9202 in ca. 9203 [Steve Henson] 9204 9205 *) Add new feature to the SPKAC handling in ca. Now you can include 9206 the same field multiple times by preceding it by "XXXX." for example: 9207 1.OU="Unit name 1" 9208 2.OU="Unit name 2" 9209 this is the same syntax as used in the req config file. 9210 [Steve Henson] 9211 9212 *) Allow certificate extensions to be added to certificate requests. These 9213 are specified in a 'req_extensions' option of the req section of the 9214 config file. They can be printed out with the -text option to req but 9215 are otherwise ignored at present. 9216 [Steve Henson] 9217 9218 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 9219 data read consists of only the final block it would not decrypted because 9220 EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 9221 A misplaced 'break' also meant the decrypted final block might not be 9222 copied until the next read. 9223 [Steve Henson] 9224 9225 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 9226 a few extra parameters to the DH structure: these will be useful if 9227 for example we want the value of 'q' or implement X9.42 DH. 9228 [Steve Henson] 9229 9230 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 9231 provides hooks that allow the default DSA functions or functions on a 9232 "per key" basis to be replaced. This allows hardware acceleration and 9233 hardware key storage to be handled without major modification to the 9234 library. Also added low level modexp hooks and CRYPTO_EX structure and 9235 associated functions. 9236 [Steve Henson] 9237 9238 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 9239 as "read only": it can't be written to and the buffer it points to will 9240 not be freed. Reading from a read only BIO is much more efficient than 9241 a normal memory BIO. This was added because there are several times when 9242 an area of memory needs to be read from a BIO. The previous method was 9243 to create a memory BIO and write the data to it, this results in two 9244 copies of the data and an O(n^2) reading algorithm. There is a new 9245 function BIO_new_mem_buf() which creates a read only memory BIO from 9246 an area of memory. Also modified the PKCS#7 routines to use read only 9247 memory BIOs. 9248 [Steve Henson] 9249 9250 *) Bugfix: ssl23_get_client_hello did not work properly when called in 9251 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 9252 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 9253 but a retry condition occured while trying to read the rest. 9254 [Bodo Moeller] 9255 9256 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 9257 NID_pkcs7_encrypted by default: this was wrong since this should almost 9258 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 9259 the encrypted data type: this is a more sensible place to put it and it 9260 allows the PKCS#12 code to be tidied up that duplicated this 9261 functionality. 9262 [Steve Henson] 9263 9264 *) Changed obj_dat.pl script so it takes its input and output files on 9265 the command line. This should avoid shell escape redirection problems 9266 under Win32. 9267 [Steve Henson] 9268 9269 *) Initial support for certificate extension requests, these are included 9270 in things like Xenroll certificate requests. Included functions to allow 9271 extensions to be obtained and added. 9272 [Steve Henson] 9273 9274 *) -crlf option to s_client and s_server for sending newlines as 9275 CRLF (as required by many protocols). 9276 [Bodo Moeller] 9277 9278 Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 9279 9280 *) Install libRSAglue.a when OpenSSL is built with RSAref. 9281 [Ralf S. Engelschall] 9282 9283 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 9284 [Andrija Antonijevic <TheAntony2@bigfoot.com>] 9285 9286 *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 9287 program. 9288 [Steve Henson] 9289 9290 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 9291 DH parameters/keys (q is lost during that conversion, but the resulting 9292 DH parameters contain its length). 9293 9294 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 9295 much faster than DH_generate_parameters (which creates parameters 9296 where p = 2*q + 1), and also the smaller q makes DH computations 9297 much more efficient (160-bit exponentiation instead of 1024-bit 9298 exponentiation); so this provides a convenient way to support DHE 9299 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 9300 utter importance to use 9301 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 9302 or 9303 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 9304 when such DH parameters are used, because otherwise small subgroup 9305 attacks may become possible! 9306 [Bodo Moeller] 9307 9308 *) Avoid memory leak in i2d_DHparams. 9309 [Bodo Moeller] 9310 9311 *) Allow the -k option to be used more than once in the enc program: 9312 this allows the same encrypted message to be read by multiple recipients. 9313 [Steve Henson] 9314 9315 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 9316 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 9317 it will always use the numerical form of the OID, even if it has a short 9318 or long name. 9319 [Steve Henson] 9320 9321 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 9322 method only got called if p,q,dmp1,dmq1,iqmp components were present, 9323 otherwise bn_mod_exp was called. In the case of hardware keys for example 9324 no private key components need be present and it might store extra data 9325 in the RSA structure, which cannot be accessed from bn_mod_exp. 9326 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 9327 private key operations. 9328 [Steve Henson] 9329 9330 *) Added support for SPARC Linux. 9331 [Andy Polyakov] 9332 9333 *) pem_password_cb function type incompatibly changed from 9334 typedef int pem_password_cb(char *buf, int size, int rwflag); 9335 to 9336 ....(char *buf, int size, int rwflag, void *userdata); 9337 so that applications can pass data to their callbacks: 9338 The PEM[_ASN1]_{read,write}... functions and macros now take an 9339 additional void * argument, which is just handed through whenever 9340 the password callback is called. 9341 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 9342 9343 New function SSL_CTX_set_default_passwd_cb_userdata. 9344 9345 Compatibility note: As many C implementations push function arguments 9346 onto the stack in reverse order, the new library version is likely to 9347 interoperate with programs that have been compiled with the old 9348 pem_password_cb definition (PEM_whatever takes some data that 9349 happens to be on the stack as its last argument, and the callback 9350 just ignores this garbage); but there is no guarantee whatsoever that 9351 this will work. 9352 9353 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 9354 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 9355 problems not only on Windows, but also on some Unix platforms. 9356 To avoid problematic command lines, these definitions are now in an 9357 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 9358 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 9359 [Bodo Moeller] 9360 9361 *) MIPS III/IV assembler module is reimplemented. 9362 [Andy Polyakov] 9363 9364 *) More DES library cleanups: remove references to srand/rand and 9365 delete an unused file. 9366 [Ulf M��ller] 9367 9368 *) Add support for the the free Netwide assembler (NASM) under Win32, 9369 since not many people have MASM (ml) and it can be hard to obtain. 9370 This is currently experimental but it seems to work OK and pass all 9371 the tests. Check out INSTALL.W32 for info. 9372 [Steve Henson] 9373 9374 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 9375 without temporary keys kept an extra copy of the server key, 9376 and connections with temporary keys did not free everything in case 9377 of an error. 9378 [Bodo Moeller] 9379 9380 *) New function RSA_check_key and new openssl rsa option -check 9381 for verifying the consistency of RSA keys. 9382 [Ulf Moeller, Bodo Moeller] 9383 9384 *) Various changes to make Win32 compile work: 9385 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 9386 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 9387 comparison" warnings. 9388 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 9389 [Steve Henson] 9390 9391 *) Add a debugging option to PKCS#5 v2 key generation function: when 9392 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 9393 derived keys are printed to stderr. 9394 [Steve Henson] 9395 9396 *) Copy the flags in ASN1_STRING_dup(). 9397 [Roman E. Pavlov <pre@mo.msk.ru>] 9398 9399 *) The x509 application mishandled signing requests containing DSA 9400 keys when the signing key was also DSA and the parameters didn't match. 9401 9402 It was supposed to omit the parameters when they matched the signing key: 9403 the verifying software was then supposed to automatically use the CA's 9404 parameters if they were absent from the end user certificate. 9405 9406 Omitting parameters is no longer recommended. The test was also 9407 the wrong way round! This was probably due to unusual behaviour in 9408 EVP_cmp_parameters() which returns 1 if the parameters match. 9409 This meant that parameters were omitted when they *didn't* match and 9410 the certificate was useless. Certificates signed with 'ca' didn't have 9411 this bug. 9412 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 9413 9414 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 9415 The interface is as follows: 9416 Applications can use 9417 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 9418 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 9419 "off" is now the default. 9420 The library internally uses 9421 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 9422 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 9423 to disable memory-checking temporarily. 9424 9425 Some inconsistent states that previously were possible (and were 9426 even the default) are now avoided. 9427 9428 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 9429 with each memory chunk allocated; this is occasionally more helpful 9430 than just having a counter. 9431 9432 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 9433 9434 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 9435 extensions. 9436 [Bodo Moeller] 9437 9438 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 9439 which largely parallels "options", but is for changing API behaviour, 9440 whereas "options" are about protocol behaviour. 9441 Initial "mode" flags are: 9442 9443 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 9444 a single record has been written. 9445 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 9446 retries use the same buffer location. 9447 (But all of the contents must be 9448 copied!) 9449 [Bodo Moeller] 9450 9451 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 9452 worked. 9453 9454 *) Fix problems with no-hmac etc. 9455 [Ulf M��ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 9456 9457 *) New functions RSA_get_default_method(), RSA_set_method() and 9458 RSA_get_method(). These allows replacement of RSA_METHODs without having 9459 to mess around with the internals of an RSA structure. 9460 [Steve Henson] 9461 9462 *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 9463 Also really enable memory leak checks in openssl.c and in some 9464 test programs. 9465 [Chad C. Mulligan, Bodo Moeller] 9466 9467 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 9468 up the length of negative integers. This has now been simplified to just 9469 store the length when it is first determined and use it later, rather 9470 than trying to keep track of where data is copied and updating it to 9471 point to the end. 9472 [Steve Henson, reported by Brien Wheeler 9473 <bwheeler@authentica-security.com>] 9474 9475 *) Add a new function PKCS7_signatureVerify. This allows the verification 9476 of a PKCS#7 signature but with the signing certificate passed to the 9477 function itself. This contrasts with PKCS7_dataVerify which assumes the 9478 certificate is present in the PKCS#7 structure. This isn't always the 9479 case: certificates can be omitted from a PKCS#7 structure and be 9480 distributed by "out of band" means (such as a certificate database). 9481 [Steve Henson] 9482 9483 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 9484 function prototypes in pem.h, also change util/mkdef.pl to add the 9485 necessary function names. 9486 [Steve Henson] 9487 9488 *) mk1mf.pl (used by Windows builds) did not properly read the 9489 options set by Configure in the top level Makefile, and Configure 9490 was not even able to write more than one option correctly. 9491 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 9492 [Bodo Moeller] 9493 9494 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 9495 file to be loaded from a BIO or FILE pointer. The BIO version will 9496 for example allow memory BIOs to contain config info. 9497 [Steve Henson] 9498 9499 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 9500 Whoever hopes to achieve shared-library compatibility across versions 9501 must use this, not the compile-time macro. 9502 (Exercise 0.9.4: Which is the minimum library version required by 9503 such programs?) 9504 Note: All this applies only to multi-threaded programs, others don't 9505 need locks. 9506 [Bodo Moeller] 9507 9508 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 9509 through a BIO pair triggered the default case, i.e. 9510 SSLerr(...,SSL_R_UNKNOWN_STATE). 9511 [Bodo Moeller] 9512 9513 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 9514 can use the SSL library even if none of the specific BIOs is 9515 appropriate. 9516 [Bodo Moeller] 9517 9518 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 9519 for the encoded length. 9520 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 9521 9522 *) Add initial documentation of the X509V3 functions. 9523 [Steve Henson] 9524 9525 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 9526 PEM_write_bio_PKCS8PrivateKey() that are equivalent to 9527 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 9528 secure PKCS#8 private key format with a high iteration count. 9529 [Steve Henson] 9530 9531 *) Fix determination of Perl interpreter: A perl or perl5 9532 _directory_ in $PATH was also accepted as the interpreter. 9533 [Ralf S. Engelschall] 9534 9535 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 9536 wrong with it but it was very old and did things like calling 9537 PEM_ASN1_read() directly and used MD5 for the hash not to mention some 9538 unusual formatting. 9539 [Steve Henson] 9540 9541 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 9542 to use the new extension code. 9543 [Steve Henson] 9544 9545 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 9546 with macros. This should make it easier to change their form, add extra 9547 arguments etc. Fix a few PEM prototypes which didn't have cipher as a 9548 constant. 9549 [Steve Henson] 9550 9551 *) Add to configuration table a new entry that can specify an alternative 9552 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 9553 according to Mark Crispin <MRC@Panda.COM>. 9554 [Bodo Moeller] 9555 9556#if 0 9557 *) DES CBC did not update the IV. Weird. 9558 [Ben Laurie] 9559#else 9560 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 9561 Changing the behaviour of the former might break existing programs -- 9562 where IV updating is needed, des_ncbc_encrypt can be used. 9563#endif 9564 9565 *) When bntest is run from "make test" it drives bc to check its 9566 calculations, as well as internally checking them. If an internal check 9567 fails, it needs to cause bc to give a non-zero result or make test carries 9568 on without noticing the failure. Fixed. 9569 [Ben Laurie] 9570 9571 *) DES library cleanups. 9572 [Ulf M��ller] 9573 9574 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 9575 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 9576 ciphers. NOTE: although the key derivation function has been verified 9577 against some published test vectors it has not been extensively tested 9578 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 9579 of v2.0. 9580 [Steve Henson] 9581 9582 *) Instead of "mkdir -p", which is not fully portable, use new 9583 Perl script "util/mkdir-p.pl". 9584 [Bodo Moeller] 9585 9586 *) Rewrite the way password based encryption (PBE) is handled. It used to 9587 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 9588 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 9589 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 9590 the 'parameter' field of the AlgorithmIdentifier is passed to the 9591 underlying key generation function so it must do its own ASN1 parsing. 9592 This has also changed the EVP_PBE_CipherInit() function which now has a 9593 'parameter' argument instead of literal salt and iteration count values 9594 and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 9595 [Steve Henson] 9596 9597 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 9598 and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 9599 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 9600 KEY" because this clashed with PKCS#8 unencrypted string. Since this 9601 value was just used as a "magic string" and not used directly its 9602 value doesn't matter. 9603 [Steve Henson] 9604 9605 *) Introduce some semblance of const correctness to BN. Shame C doesn't 9606 support mutable. 9607 [Ben Laurie] 9608 9609 *) "linux-sparc64" configuration (ultrapenguin). 9610 [Ray Miller <ray.miller@oucs.ox.ac.uk>] 9611 "linux-sparc" configuration. 9612 [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 9613 9614 *) config now generates no-xxx options for missing ciphers. 9615 [Ulf M��ller] 9616 9617 *) Support the EBCDIC character set (work in progress). 9618 File ebcdic.c not yet included because it has a different license. 9619 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 9620 9621 *) Support BS2000/OSD-POSIX. 9622 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 9623 9624 *) Make callbacks for key generation use void * instead of char *. 9625 [Ben Laurie] 9626 9627 *) Make S/MIME samples compile (not yet tested). 9628 [Ben Laurie] 9629 9630 *) Additional typesafe stacks. 9631 [Ben Laurie] 9632 9633 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 9634 [Bodo Moeller] 9635 9636 9637 Changes between 0.9.3 and 0.9.3a [29 May 1999] 9638 9639 *) New configuration variant "sco5-gcc". 9640 9641 *) Updated some demos. 9642 [Sean O Riordain, Wade Scholine] 9643 9644 *) Add missing BIO_free at exit of pkcs12 application. 9645 [Wu Zhigang] 9646 9647 *) Fix memory leak in conf.c. 9648 [Steve Henson] 9649 9650 *) Updates for Win32 to assembler version of MD5. 9651 [Steve Henson] 9652 9653 *) Set #! path to perl in apps/der_chop to where we found it 9654 instead of using a fixed path. 9655 [Bodo Moeller] 9656 9657 *) SHA library changes for irix64-mips4-cc. 9658 [Andy Polyakov] 9659 9660 *) Improvements for VMS support. 9661 [Richard Levitte] 9662 9663 9664 Changes between 0.9.2b and 0.9.3 [24 May 1999] 9665 9666 *) Bignum library bug fix. IRIX 6 passes "make test" now! 9667 This also avoids the problems with SC4.2 and unpatched SC5. 9668 [Andy Polyakov <appro@fy.chalmers.se>] 9669 9670 *) New functions sk_num, sk_value and sk_set to replace the previous macros. 9671 These are required because of the typesafe stack would otherwise break 9672 existing code. If old code used a structure member which used to be STACK 9673 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 9674 sk_num or sk_value it would produce an error because the num, data members 9675 are not present in STACK_OF. Now it just produces a warning. sk_set 9676 replaces the old method of assigning a value to sk_value 9677 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 9678 that does this will no longer work (and should use sk_set instead) but 9679 this could be regarded as a "questionable" behaviour anyway. 9680 [Steve Henson] 9681 9682 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 9683 correctly handle encrypted S/MIME data. 9684 [Steve Henson] 9685 9686 *) Change type of various DES function arguments from des_cblock 9687 (which means, in function argument declarations, pointer to char) 9688 to des_cblock * (meaning pointer to array with 8 char elements), 9689 which allows the compiler to do more typechecking; it was like 9690 that back in SSLeay, but with lots of ugly casts. 9691 9692 Introduce new type const_des_cblock. 9693 [Bodo Moeller] 9694 9695 *) Reorganise the PKCS#7 library and get rid of some of the more obvious 9696 problems: find RecipientInfo structure that matches recipient certificate 9697 and initialise the ASN1 structures properly based on passed cipher. 9698 [Steve Henson] 9699 9700 *) Belatedly make the BN tests actually check the results. 9701 [Ben Laurie] 9702 9703 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 9704 to and from BNs: it was completely broken. New compilation option 9705 NEG_PUBKEY_BUG to allow for some broken certificates that encode public 9706 key elements as negative integers. 9707 [Steve Henson] 9708 9709 *) Reorganize and speed up MD5. 9710 [Andy Polyakov <appro@fy.chalmers.se>] 9711 9712 *) VMS support. 9713 [Richard Levitte <richard@levitte.org>] 9714 9715 *) New option -out to asn1parse to allow the parsed structure to be 9716 output to a file. This is most useful when combined with the -strparse 9717 option to examine the output of things like OCTET STRINGS. 9718 [Steve Henson] 9719 9720 *) Make SSL library a little more fool-proof by not requiring any longer 9721 that SSL_set_{accept,connect}_state be called before 9722 SSL_{accept,connect} may be used (SSL_set_..._state is omitted 9723 in many applications because usually everything *appeared* to work as 9724 intended anyway -- now it really works as intended). 9725 [Bodo Moeller] 9726 9727 *) Move openssl.cnf out of lib/. 9728 [Ulf M��ller] 9729 9730 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 9731 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 9732 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 9733 [Ralf S. Engelschall] 9734 9735 *) Various fixes to the EVP and PKCS#7 code. It may now be able to 9736 handle PKCS#7 enveloped data properly. 9737 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 9738 9739 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 9740 copying pointers. The cert_st handling is changed by this in 9741 various ways (and thus what used to be known as ctx->default_cert 9742 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 9743 any longer when s->cert does not give us what we need). 9744 ssl_cert_instantiate becomes obsolete by this change. 9745 As soon as we've got the new code right (possibly it already is?), 9746 we have solved a couple of bugs of the earlier code where s->cert 9747 was used as if it could not have been shared with other SSL structures. 9748 9749 Note that using the SSL API in certain dirty ways now will result 9750 in different behaviour than observed with earlier library versions: 9751 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 9752 does not influence s as it used to. 9753 9754 In order to clean up things more thoroughly, inside SSL_SESSION 9755 we don't use CERT any longer, but a new structure SESS_CERT 9756 that holds per-session data (if available); currently, this is 9757 the peer's certificate chain and, for clients, the server's certificate 9758 and temporary key. CERT holds only those values that can have 9759 meaningful defaults in an SSL_CTX. 9760 [Bodo Moeller] 9761 9762 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 9763 from the internal representation. Various PKCS#7 fixes: remove some 9764 evil casts and set the enc_dig_alg field properly based on the signing 9765 key type. 9766 [Steve Henson] 9767 9768 *) Allow PKCS#12 password to be set from the command line or the 9769 environment. Let 'ca' get its config file name from the environment 9770 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 9771 and 'x509'). 9772 [Steve Henson] 9773 9774 *) Allow certificate policies extension to use an IA5STRING for the 9775 organization field. This is contrary to the PKIX definition but 9776 VeriSign uses it and IE5 only recognises this form. Document 'x509' 9777 extension option. 9778 [Steve Henson] 9779 9780 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 9781 without disallowing inline assembler and the like for non-pedantic builds. 9782 [Ben Laurie] 9783 9784 *) Support Borland C++ builder. 9785 [Janez Jere <jj@void.si>, modified by Ulf M��ller] 9786 9787 *) Support Mingw32. 9788 [Ulf M��ller] 9789 9790 *) SHA-1 cleanups and performance enhancements. 9791 [Andy Polyakov <appro@fy.chalmers.se>] 9792 9793 *) Sparc v8plus assembler for the bignum library. 9794 [Andy Polyakov <appro@fy.chalmers.se>] 9795 9796 *) Accept any -xxx and +xxx compiler options in Configure. 9797 [Ulf M��ller] 9798 9799 *) Update HPUX configuration. 9800 [Anonymous] 9801 9802 *) Add missing sk_<type>_unshift() function to safestack.h 9803 [Ralf S. Engelschall] 9804 9805 *) New function SSL_CTX_use_certificate_chain_file that sets the 9806 "extra_cert"s in addition to the certificate. (This makes sense 9807 only for "PEM" format files, as chains as a whole are not 9808 DER-encoded.) 9809 [Bodo Moeller] 9810 9811 *) Support verify_depth from the SSL API. 9812 x509_vfy.c had what can be considered an off-by-one-error: 9813 Its depth (which was not part of the external interface) 9814 was actually counting the number of certificates in a chain; 9815 now it really counts the depth. 9816 [Bodo Moeller] 9817 9818 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 9819 instead of X509err, which often resulted in confusing error 9820 messages since the error codes are not globally unique 9821 (e.g. an alleged error in ssl3_accept when a certificate 9822 didn't match the private key). 9823 9824 *) New function SSL_CTX_set_session_id_context that allows to set a default 9825 value (so that you don't need SSL_set_session_id_context for each 9826 connection using the SSL_CTX). 9827 [Bodo Moeller] 9828 9829 *) OAEP decoding bug fix. 9830 [Ulf M��ller] 9831 9832 *) Support INSTALL_PREFIX for package builders, as proposed by 9833 David Harris. 9834 [Bodo Moeller] 9835 9836 *) New Configure options "threads" and "no-threads". For systems 9837 where the proper compiler options are known (currently Solaris 9838 and Linux), "threads" is the default. 9839 [Bodo Moeller] 9840 9841 *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 9842 [Bodo Moeller] 9843 9844 *) Install various scripts to $(OPENSSLDIR)/misc, not to 9845 $(INSTALLTOP)/bin -- they shouldn't clutter directories 9846 such as /usr/local/bin. 9847 [Bodo Moeller] 9848 9849 *) "make linux-shared" to build shared libraries. 9850 [Niels Poppe <niels@netbox.org>] 9851 9852 *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 9853 [Ulf M��ller] 9854 9855 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 9856 extension adding in x509 utility. 9857 [Steve Henson] 9858 9859 *) Remove NOPROTO sections and error code comments. 9860 [Ulf M��ller] 9861 9862 *) Partial rewrite of the DEF file generator to now parse the ANSI 9863 prototypes. 9864 [Steve Henson] 9865 9866 *) New Configure options --prefix=DIR and --openssldir=DIR. 9867 [Ulf M��ller] 9868 9869 *) Complete rewrite of the error code script(s). It is all now handled 9870 by one script at the top level which handles error code gathering, 9871 header rewriting and C source file generation. It should be much better 9872 than the old method: it now uses a modified version of Ulf's parser to 9873 read the ANSI prototypes in all header files (thus the old K&R definitions 9874 aren't needed for error creation any more) and do a better job of 9875 translating function codes into names. The old 'ASN1 error code imbedded 9876 in a comment' is no longer necessary and it doesn't use .err files which 9877 have now been deleted. Also the error code call doesn't have to appear all 9878 on one line (which resulted in some large lines...). 9879 [Steve Henson] 9880 9881 *) Change #include filenames from <foo.h> to <openssl/foo.h>. 9882 [Bodo Moeller] 9883 9884 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 9885 0 (which usually indicates a closed connection), but continue reading. 9886 [Bodo Moeller] 9887 9888 *) Fix some race conditions. 9889 [Bodo Moeller] 9890 9891 *) Add support for CRL distribution points extension. Add Certificate 9892 Policies and CRL distribution points documentation. 9893 [Steve Henson] 9894 9895 *) Move the autogenerated header file parts to crypto/opensslconf.h. 9896 [Ulf M��ller] 9897 9898 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 9899 8 of keying material. Merlin has also confirmed interop with this fix 9900 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 9901 [Merlin Hughes <merlin@baltimore.ie>] 9902 9903 *) Fix lots of warnings. 9904 [Richard Levitte <levitte@stacken.kth.se>] 9905 9906 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 9907 the directory spec didn't end with a LIST_SEPARATOR_CHAR. 9908 [Richard Levitte <levitte@stacken.kth.se>] 9909 9910 *) Fix problems with sizeof(long) == 8. 9911 [Andy Polyakov <appro@fy.chalmers.se>] 9912 9913 *) Change functions to ANSI C. 9914 [Ulf M��ller] 9915 9916 *) Fix typos in error codes. 9917 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M��ller] 9918 9919 *) Remove defunct assembler files from Configure. 9920 [Ulf M��ller] 9921 9922 *) SPARC v8 assembler BIGNUM implementation. 9923 [Andy Polyakov <appro@fy.chalmers.se>] 9924 9925 *) Support for Certificate Policies extension: both print and set. 9926 Various additions to support the r2i method this uses. 9927 [Steve Henson] 9928 9929 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 9930 return a const string when you are expecting an allocated buffer. 9931 [Ben Laurie] 9932 9933 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 9934 types DirectoryString and DisplayText. 9935 [Steve Henson] 9936 9937 *) Add code to allow r2i extensions to access the configuration database, 9938 add an LHASH database driver and add several ctx helper functions. 9939 [Steve Henson] 9940 9941 *) Fix an evil bug in bn_expand2() which caused various BN functions to 9942 fail when they extended the size of a BIGNUM. 9943 [Steve Henson] 9944 9945 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 9946 support typesafe stack. 9947 [Steve Henson] 9948 9949 *) Fix typo in SSL_[gs]et_options(). 9950 [Nils Frostberg <nils@medcom.se>] 9951 9952 *) Delete various functions and files that belonged to the (now obsolete) 9953 old X509V3 handling code. 9954 [Steve Henson] 9955 9956 *) New Configure option "rsaref". 9957 [Ulf M��ller] 9958 9959 *) Don't auto-generate pem.h. 9960 [Bodo Moeller] 9961 9962 *) Introduce type-safe ASN.1 SETs. 9963 [Ben Laurie] 9964 9965 *) Convert various additional casted stacks to type-safe STACK_OF() variants. 9966 [Ben Laurie, Ralf S. Engelschall, Steve Henson] 9967 9968 *) Introduce type-safe STACKs. This will almost certainly break lots of code 9969 that links with OpenSSL (well at least cause lots of warnings), but fear 9970 not: the conversion is trivial, and it eliminates loads of evil casts. A 9971 few STACKed things have been converted already. Feel free to convert more. 9972 In the fullness of time, I'll do away with the STACK type altogether. 9973 [Ben Laurie] 9974 9975 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 9976 specified in <certfile> by updating the entry in the index.txt file. 9977 This way one no longer has to edit the index.txt file manually for 9978 revoking a certificate. The -revoke option does the gory details now. 9979 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 9980 9981 *) Fix `openssl crl -noout -text' combination where `-noout' killed the 9982 `-text' option at all and this way the `-noout -text' combination was 9983 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 9984 [Ralf S. Engelschall] 9985 9986 *) Make sure a corresponding plain text error message exists for the 9987 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 9988 verify callback function determined that a certificate was revoked. 9989 [Ralf S. Engelschall] 9990 9991 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 9992 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 9993 all available cipers including rc5, which was forgotten until now. 9994 In order to let the testing shell script know which algorithms 9995 are available, a new (up to now undocumented) command 9996 "openssl list-cipher-commands" is used. 9997 [Bodo Moeller] 9998 9999 *) Bugfix: s_client occasionally would sleep in select() when 10000 it should have checked SSL_pending() first. 10001 [Bodo Moeller] 10002 10003 *) New functions DSA_do_sign and DSA_do_verify to provide access to 10004 the raw DSA values prior to ASN.1 encoding. 10005 [Ulf M��ller] 10006 10007 *) Tweaks to Configure 10008 [Niels Poppe <niels@netbox.org>] 10009 10010 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 10011 yet... 10012 [Steve Henson] 10013 10014 *) New variables $(RANLIB) and $(PERL) in the Makefiles. 10015 [Ulf M��ller] 10016 10017 *) New config option to avoid instructions that are illegal on the 80386. 10018 The default code is faster, but requires at least a 486. 10019 [Ulf M��ller] 10020 10021 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 10022 SSL2_SERVER_VERSION (not used at all) macros, which are now the 10023 same as SSL2_VERSION anyway. 10024 [Bodo Moeller] 10025 10026 *) New "-showcerts" option for s_client. 10027 [Bodo Moeller] 10028 10029 *) Still more PKCS#12 integration. Add pkcs12 application to openssl 10030 application. Various cleanups and fixes. 10031 [Steve Henson] 10032 10033 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 10034 modify error routines to work internally. Add error codes and PBE init 10035 to library startup routines. 10036 [Steve Henson] 10037 10038 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 10039 packing functions to asn1 and evp. Changed function names and error 10040 codes along the way. 10041 [Steve Henson] 10042 10043 *) PKCS12 integration: and so it begins... First of several patches to 10044 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 10045 objects to objects.h 10046 [Steve Henson] 10047 10048 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 10049 and display support for Thawte strong extranet extension. 10050 [Steve Henson] 10051 10052 *) Add LinuxPPC support. 10053 [Jeff Dubrule <igor@pobox.org>] 10054 10055 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 10056 bn_div_words in alpha.s. 10057 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 10058 10059 *) Make sure the RSA OAEP test is skipped under -DRSAref because 10060 OAEP isn't supported when OpenSSL is built with RSAref. 10061 [Ulf Moeller <ulf@fitug.de>] 10062 10063 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 10064 so they no longer are missing under -DNOPROTO. 10065 [Soren S. Jorvang <soren@t.dk>] 10066 10067 10068 Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 10069 10070 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 10071 doesn't work when the session is reused. Coming soon! 10072 [Ben Laurie] 10073 10074 *) Fix a security hole, that allows sessions to be reused in the wrong 10075 context thus bypassing client cert protection! All software that uses 10076 client certs and session caches in multiple contexts NEEDS PATCHING to 10077 allow session reuse! A fuller solution is in the works. 10078 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 10079 10080 *) Some more source tree cleanups (removed obsolete files 10081 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 10082 permission on "config" script to be executable) and a fix for the INSTALL 10083 document. 10084 [Ulf Moeller <ulf@fitug.de>] 10085 10086 *) Remove some legacy and erroneous uses of malloc, free instead of 10087 Malloc, Free. 10088 [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 10089 10090 *) Make rsa_oaep_test return non-zero on error. 10091 [Ulf Moeller <ulf@fitug.de>] 10092 10093 *) Add support for native Solaris shared libraries. Configure 10094 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 10095 if someone would make that last step automatic. 10096 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 10097 10098 *) ctx_size was not built with the right compiler during "make links". Fixed. 10099 [Ben Laurie] 10100 10101 *) Change the meaning of 'ALL' in the cipher list. It now means "everything 10102 except NULL ciphers". This means the default cipher list will no longer 10103 enable NULL ciphers. They need to be specifically enabled e.g. with 10104 the string "DEFAULT:eNULL". 10105 [Steve Henson] 10106 10107 *) Fix to RSA private encryption routines: if p < q then it would 10108 occasionally produce an invalid result. This will only happen with 10109 externally generated keys because OpenSSL (and SSLeay) ensure p > q. 10110 [Steve Henson] 10111 10112 *) Be less restrictive and allow also `perl util/perlpath.pl 10113 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 10114 because this way one can also use an interpreter named `perl5' (which is 10115 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 10116 installed as `perl'). 10117 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 10118 10119 *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 10120 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 10121 10122 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 10123 advapi32.lib to Win32 build and change the pem test comparision 10124 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 10125 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 10126 and crypto/des/ede_cbcm_enc.c. 10127 [Steve Henson] 10128 10129 *) DES quad checksum was broken on big-endian architectures. Fixed. 10130 [Ben Laurie] 10131 10132 *) Comment out two functions in bio.h that aren't implemented. Fix up the 10133 Win32 test batch file so it (might) work again. The Win32 test batch file 10134 is horrible: I feel ill.... 10135 [Steve Henson] 10136 10137 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 10138 in e_os.h. Audit of header files to check ANSI and non ANSI 10139 sections: 10 functions were absent from non ANSI section and not exported 10140 from Windows DLLs. Fixed up libeay.num for new functions. 10141 [Steve Henson] 10142 10143 *) Make `openssl version' output lines consistent. 10144 [Ralf S. Engelschall] 10145 10146 *) Fix Win32 symbol export lists for BIO functions: Added 10147 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 10148 to ms/libeay{16,32}.def. 10149 [Ralf S. Engelschall] 10150 10151 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 10152 fine under Unix and passes some trivial tests I've now added. But the 10153 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 10154 added to make sure no one expects that this stuff really works in the 10155 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 10156 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 10157 openssl_bio.xs. 10158 [Ralf S. Engelschall] 10159 10160 *) Fix the generation of two part addresses in perl. 10161 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 10162 10163 *) Add config entry for Linux on MIPS. 10164 [John Tobey <jtobey@channel1.com>] 10165 10166 *) Make links whenever Configure is run, unless we are on Windoze. 10167 [Ben Laurie] 10168 10169 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 10170 Currently only issuerAltName and AuthorityKeyIdentifier make any sense 10171 in CRLs. 10172 [Steve Henson] 10173 10174 *) Add a useful kludge to allow package maintainers to specify compiler and 10175 other platforms details on the command line without having to patch the 10176 Configure script everytime: One now can use ``perl Configure 10177 <id>:<details>'', i.e. platform ids are allowed to have details appended 10178 to them (seperated by colons). This is treated as there would be a static 10179 pre-configured entry in Configure's %table under key <id> with value 10180 <details> and ``perl Configure <id>'' is called. So, when you want to 10181 perform a quick test-compile under FreeBSD 3.1 with pgcc and without 10182 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 10183 now, which overrides the FreeBSD-elf entry on-the-fly. 10184 [Ralf S. Engelschall] 10185 10186 *) Disable new TLS1 ciphersuites by default: they aren't official yet. 10187 [Ben Laurie] 10188 10189 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 10190 on the `perl Configure ...' command line. This way one can compile 10191 OpenSSL libraries with Position Independent Code (PIC) which is needed 10192 for linking it into DSOs. 10193 [Ralf S. Engelschall] 10194 10195 *) Remarkably, export ciphers were totally broken and no-one had noticed! 10196 Fixed. 10197 [Ben Laurie] 10198 10199 *) Cleaned up the LICENSE document: The official contact for any license 10200 questions now is the OpenSSL core team under openssl-core@openssl.org. 10201 And add a paragraph about the dual-license situation to make sure people 10202 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 10203 to the OpenSSL toolkit. 10204 [Ralf S. Engelschall] 10205 10206 *) General source tree makefile cleanups: Made `making xxx in yyy...' 10207 display consistent in the source tree and replaced `/bin/rm' by `rm'. 10208 Additonally cleaned up the `make links' target: Remove unnecessary 10209 semicolons, subsequent redundant removes, inline point.sh into mklink.sh 10210 to speed processing and no longer clutter the display with confusing 10211 stuff. Instead only the actually done links are displayed. 10212 [Ralf S. Engelschall] 10213 10214 *) Permit null encryption ciphersuites, used for authentication only. It used 10215 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 10216 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 10217 encryption. 10218 [Ben Laurie] 10219 10220 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 10221 signed attributes when verifying signatures (this would break them), 10222 the detached data encoding was wrong and public keys obtained using 10223 X509_get_pubkey() weren't freed. 10224 [Steve Henson] 10225 10226 *) Add text documentation for the BUFFER functions. Also added a work around 10227 to a Win95 console bug. This was triggered by the password read stuff: the 10228 last character typed gets carried over to the next fread(). If you were 10229 generating a new cert request using 'req' for example then the last 10230 character of the passphrase would be CR which would then enter the first 10231 field as blank. 10232 [Steve Henson] 10233 10234 *) Added the new `Includes OpenSSL Cryptography Software' button as 10235 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 10236 button and can be used by applications based on OpenSSL to show the 10237 relationship to the OpenSSL project. 10238 [Ralf S. Engelschall] 10239 10240 *) Remove confusing variables in function signatures in files 10241 ssl/ssl_lib.c and ssl/ssl.h. 10242 [Lennart Bong <lob@kulthea.stacken.kth.se>] 10243 10244 *) Don't install bss_file.c under PREFIX/include/ 10245 [Lennart Bong <lob@kulthea.stacken.kth.se>] 10246 10247 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 10248 functions that return function pointers and has support for NT specific 10249 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 10250 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 10251 unsigned to signed types: this was killing the Win32 compile. 10252 [Steve Henson] 10253 10254 *) Add new certificate file to stack functions, 10255 SSL_add_dir_cert_subjects_to_stack() and 10256 SSL_add_file_cert_subjects_to_stack(). These largely supplant 10257 SSL_load_client_CA_file(), and can be used to add multiple certs easily 10258 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 10259 This means that Apache-SSL and similar packages don't have to mess around 10260 to add as many CAs as they want to the preferred list. 10261 [Ben Laurie] 10262 10263 *) Experiment with doxygen documentation. Currently only partially applied to 10264 ssl/ssl_lib.c. 10265 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 10266 openssl.doxy as the configuration file. 10267 [Ben Laurie] 10268 10269 *) Get rid of remaining C++-style comments which strict C compilers hate. 10270 [Ralf S. Engelschall, pointed out by Carlos Amengual] 10271 10272 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 10273 compiled in by default: it has problems with large keys. 10274 [Steve Henson] 10275 10276 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 10277 DH private keys and/or callback functions which directly correspond to 10278 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 10279 is needed for applications which have to configure certificates on a 10280 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 10281 (e.g. s_server). 10282 For the RSA certificate situation is makes no difference, but 10283 for the DSA certificate situation this fixes the "no shared cipher" 10284 problem where the OpenSSL cipher selection procedure failed because the 10285 temporary keys were not overtaken from the context and the API provided 10286 no way to reconfigure them. 10287 The new functions now let applications reconfigure the stuff and they 10288 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 10289 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 10290 non-public-API function ssl_cert_instantiate() is used as a helper 10291 function and also to reduce code redundancy inside ssl_rsa.c. 10292 [Ralf S. Engelschall] 10293 10294 *) Move s_server -dcert and -dkey options out of the undocumented feature 10295 area because they are useful for the DSA situation and should be 10296 recognized by the users. 10297 [Ralf S. Engelschall] 10298 10299 *) Fix the cipher decision scheme for export ciphers: the export bits are 10300 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 10301 SSL_EXP_MASK. So, the original variable has to be used instead of the 10302 already masked variable. 10303 [Richard Levitte <levitte@stacken.kth.se>] 10304 10305 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 10306 [Richard Levitte <levitte@stacken.kth.se>] 10307 10308 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 10309 from `int' to `unsigned int' because it's a length and initialized by 10310 EVP_DigestFinal() which expects an `unsigned int *'. 10311 [Richard Levitte <levitte@stacken.kth.se>] 10312 10313 *) Don't hard-code path to Perl interpreter on shebang line of Configure 10314 script. Instead use the usual Shell->Perl transition trick. 10315 [Ralf S. Engelschall] 10316 10317 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 10318 (in addition to RSA certificates) to match the behaviour of `openssl dsa 10319 -noout -modulus' as it's already the case for `openssl rsa -noout 10320 -modulus'. For RSA the -modulus is the real "modulus" while for DSA 10321 currently the public key is printed (a decision which was already done by 10322 `openssl dsa -modulus' in the past) which serves a similar purpose. 10323 Additionally the NO_RSA no longer completely removes the whole -modulus 10324 option; it now only avoids using the RSA stuff. Same applies to NO_DSA 10325 now, too. 10326 [Ralf S. Engelschall] 10327 10328 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 10329 BIO. See the source (crypto/evp/bio_ok.c) for more info. 10330 [Arne Ansper <arne@ats.cyber.ee>] 10331 10332 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 10333 to be added. Now both 'req' and 'ca' can use new objects defined in the 10334 config file. 10335 [Steve Henson] 10336 10337 *) Add cool BIO that does syslog (or event log on NT). 10338 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 10339 10340 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 10341 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 10342 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 10343 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 10344 [Ben Laurie] 10345 10346 *) Add preliminary config info for new extension code. 10347 [Steve Henson] 10348 10349 *) Make RSA_NO_PADDING really use no padding. 10350 [Ulf Moeller <ulf@fitug.de>] 10351 10352 *) Generate errors when private/public key check is done. 10353 [Ben Laurie] 10354 10355 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 10356 for some CRL extensions and new objects added. 10357 [Steve Henson] 10358 10359 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 10360 key usage extension and fuller support for authority key id. 10361 [Steve Henson] 10362 10363 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 10364 padding method for RSA, which is recommended for new applications in PKCS 10365 #1 v2.0 (RFC 2437, October 1998). 10366 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 10367 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 10368 against Bleichbacher's attack on RSA. 10369 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 10370 Ben Laurie] 10371 10372 *) Updates to the new SSL compression code 10373 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 10374 10375 *) Fix so that the version number in the master secret, when passed 10376 via RSA, checks that if TLS was proposed, but we roll back to SSLv3 10377 (because the server will not accept higher), that the version number 10378 is 0x03,0x01, not 0x03,0x00 10379 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 10380 10381 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 10382 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 10383 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 10384 [Steve Henson] 10385 10386 *) Support for RAW extensions where an arbitrary extension can be 10387 created by including its DER encoding. See apps/openssl.cnf for 10388 an example. 10389 [Steve Henson] 10390 10391 *) Make sure latest Perl versions don't interpret some generated C array 10392 code as Perl array code in the crypto/err/err_genc.pl script. 10393 [Lars Weber <3weber@informatik.uni-hamburg.de>] 10394 10395 *) Modify ms/do_ms.bat to not generate assembly language makefiles since 10396 not many people have the assembler. Various Win32 compilation fixes and 10397 update to the INSTALL.W32 file with (hopefully) more accurate Win32 10398 build instructions. 10399 [Steve Henson] 10400 10401 *) Modify configure script 'Configure' to automatically create crypto/date.h 10402 file under Win32 and also build pem.h from pem.org. New script 10403 util/mkfiles.pl to create the MINFO file on environments that can't do a 10404 'make files': perl util/mkfiles.pl >MINFO should work. 10405 [Steve Henson] 10406 10407 *) Major rework of DES function declarations, in the pursuit of correctness 10408 and purity. As a result, many evil casts evaporated, and some weirdness, 10409 too. You may find this causes warnings in your code. Zapping your evil 10410 casts will probably fix them. Mostly. 10411 [Ben Laurie] 10412 10413 *) Fix for a typo in asn1.h. Bug fix to object creation script 10414 obj_dat.pl. It considered a zero in an object definition to mean 10415 "end of object": none of the objects in objects.h have any zeros 10416 so it wasn't spotted. 10417 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 10418 10419 *) Add support for Triple DES Cipher Block Chaining with Output Feedback 10420 Masking (CBCM). In the absence of test vectors, the best I have been able 10421 to do is check that the decrypt undoes the encrypt, so far. Send me test 10422 vectors if you have them. 10423 [Ben Laurie] 10424 10425 *) Correct calculation of key length for export ciphers (too much space was 10426 allocated for null ciphers). This has not been tested! 10427 [Ben Laurie] 10428 10429 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 10430 message is now correct (it understands "crypto" and "ssl" on its 10431 command line). There is also now an "update" option. This will update 10432 the util/ssleay.num and util/libeay.num files with any new functions. 10433 If you do a: 10434 perl util/mkdef.pl crypto ssl update 10435 it will update them. 10436 [Steve Henson] 10437 10438 *) Overhauled the Perl interface (perl/*): 10439 - ported BN stuff to OpenSSL's different BN library 10440 - made the perl/ source tree CVS-aware 10441 - renamed the package from SSLeay to OpenSSL (the files still contain 10442 their history because I've copied them in the repository) 10443 - removed obsolete files (the test scripts will be replaced 10444 by better Test::Harness variants in the future) 10445 [Ralf S. Engelschall] 10446 10447 *) First cut for a very conservative source tree cleanup: 10448 1. merge various obsolete readme texts into doc/ssleay.txt 10449 where we collect the old documents and readme texts. 10450 2. remove the first part of files where I'm already sure that we no 10451 longer need them because of three reasons: either they are just temporary 10452 files which were left by Eric or they are preserved original files where 10453 I've verified that the diff is also available in the CVS via "cvs diff 10454 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 10455 the crypto/md/ stuff). 10456 [Ralf S. Engelschall] 10457 10458 *) More extension code. Incomplete support for subject and issuer alt 10459 name, issuer and authority key id. Change the i2v function parameters 10460 and add an extra 'crl' parameter in the X509V3_CTX structure: guess 10461 what that's for :-) Fix to ASN1 macro which messed up 10462 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 10463 [Steve Henson] 10464 10465 *) Preliminary support for ENUMERATED type. This is largely copied from the 10466 INTEGER code. 10467 [Steve Henson] 10468 10469 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 10470 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 10471 10472 *) Make sure `make rehash' target really finds the `openssl' program. 10473 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 10474 10475 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 10476 like to hear about it if this slows down other processors. 10477 [Ben Laurie] 10478 10479 *) Add CygWin32 platform information to Configure script. 10480 [Alan Batie <batie@aahz.jf.intel.com>] 10481 10482 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 10483 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 10484 10485 *) New program nseq to manipulate netscape certificate sequences 10486 [Steve Henson] 10487 10488 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 10489 few typos. 10490 [Steve Henson] 10491 10492 *) Fixes to BN code. Previously the default was to define BN_RECURSION 10493 but the BN code had some problems that would cause failures when 10494 doing certificate verification and some other functions. 10495 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 10496 10497 *) Add ASN1 and PEM code to support netscape certificate sequences. 10498 [Steve Henson] 10499 10500 *) Add ASN1 and PEM code to support netscape certificate sequences. 10501 [Steve Henson] 10502 10503 *) Add several PKIX and private extended key usage OIDs. 10504 [Steve Henson] 10505 10506 *) Modify the 'ca' program to handle the new extension code. Modify 10507 openssl.cnf for new extension format, add comments. 10508 [Steve Henson] 10509 10510 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 10511 and add a sample to openssl.cnf so req -x509 now adds appropriate 10512 CA extensions. 10513 [Steve Henson] 10514 10515 *) Continued X509 V3 changes. Add to other makefiles, integrate with the 10516 error code, add initial support to X509_print() and x509 application. 10517 [Steve Henson] 10518 10519 *) Takes a deep breath and start addding X509 V3 extension support code. Add 10520 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 10521 stuff is currently isolated and isn't even compiled yet. 10522 [Steve Henson] 10523 10524 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 10525 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 10526 Removed the versions check from X509 routines when loading extensions: 10527 this allows certain broken certificates that don't set the version 10528 properly to be processed. 10529 [Steve Henson] 10530 10531 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 10532 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 10533 can still be regenerated with "make depend". 10534 [Ben Laurie] 10535 10536 *) Spelling mistake in C version of CAST-128. 10537 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 10538 10539 *) Changes to the error generation code. The perl script err-code.pl 10540 now reads in the old error codes and retains the old numbers, only 10541 adding new ones if necessary. It also only changes the .err files if new 10542 codes are added. The makefiles have been modified to only insert errors 10543 when needed (to avoid needlessly modifying header files). This is done 10544 by only inserting errors if the .err file is newer than the auto generated 10545 C file. To rebuild all the error codes from scratch (the old behaviour) 10546 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 10547 or delete all the .err files. 10548 [Steve Henson] 10549 10550 *) CAST-128 was incorrectly implemented for short keys. The C version has 10551 been fixed, but is untested. The assembler versions are also fixed, but 10552 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 10553 to regenerate it if needed. 10554 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 10555 Hagino <itojun@kame.net>] 10556 10557 *) File was opened incorrectly in randfile.c. 10558 [Ulf M��ller <ulf@fitug.de>] 10559 10560 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 10561 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 10562 GeneralizedTime. ASN1_TIME is the proper type used in certificates et 10563 al: it's just almost always a UTCTime. Note this patch adds new error 10564 codes so do a "make errors" if there are problems. 10565 [Steve Henson] 10566 10567 *) Correct Linux 1 recognition in config. 10568 [Ulf M��ller <ulf@fitug.de>] 10569 10570 *) Remove pointless MD5 hash when using DSA keys in ca. 10571 [Anonymous <nobody@replay.com>] 10572 10573 *) Generate an error if given an empty string as a cert directory. Also 10574 generate an error if handed NULL (previously returned 0 to indicate an 10575 error, but didn't set one). 10576 [Ben Laurie, reported by Anonymous <nobody@replay.com>] 10577 10578 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 10579 [Ben Laurie] 10580 10581 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 10582 parameters. This was causing a warning which killed off the Win32 compile. 10583 [Steve Henson] 10584 10585 *) Remove C++ style comments from crypto/bn/bn_local.h. 10586 [Neil Costigan <neil.costigan@celocom.com>] 10587 10588 *) The function OBJ_txt2nid was broken. It was supposed to return a nid 10589 based on a text string, looking up short and long names and finally 10590 "dot" format. The "dot" format stuff didn't work. Added new function 10591 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 10592 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 10593 OID is not part of the table. 10594 [Steve Henson] 10595 10596 *) Add prototypes to X509 lookup/verify methods, fixing a bug in 10597 X509_LOOKUP_by_alias(). 10598 [Ben Laurie] 10599 10600 *) Sort openssl functions by name. 10601 [Ben Laurie] 10602 10603 *) Get the gendsa program working (hopefully) and add it to app list. Remove 10604 encryption from sample DSA keys (in case anyone is interested the password 10605 was "1234"). 10606 [Steve Henson] 10607 10608 *) Make _all_ *_free functions accept a NULL pointer. 10609 [Frans Heymans <fheymans@isaserver.be>] 10610 10611 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 10612 NULL pointers. 10613 [Anonymous <nobody@replay.com>] 10614 10615 *) s_server should send the CAfile as acceptable CAs, not its own cert. 10616 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 10617 10618 *) Don't blow it for numeric -newkey arguments to apps/req. 10619 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 10620 10621 *) Temp key "for export" tests were wrong in s3_srvr.c. 10622 [Anonymous <nobody@replay.com>] 10623 10624 *) Add prototype for temp key callback functions 10625 SSL_CTX_set_tmp_{rsa,dh}_callback(). 10626 [Ben Laurie] 10627 10628 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 10629 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 10630 [Steve Henson] 10631 10632 *) X509_name_add_entry() freed the wrong thing after an error. 10633 [Arne Ansper <arne@ats.cyber.ee>] 10634 10635 *) rsa_eay.c would attempt to free a NULL context. 10636 [Arne Ansper <arne@ats.cyber.ee>] 10637 10638 *) BIO_s_socket() had a broken should_retry() on Windoze. 10639 [Arne Ansper <arne@ats.cyber.ee>] 10640 10641 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 10642 [Arne Ansper <arne@ats.cyber.ee>] 10643 10644 *) Make sure the already existing X509_STORE->depth variable is initialized 10645 in X509_STORE_new(), but document the fact that this variable is still 10646 unused in the certificate verification process. 10647 [Ralf S. Engelschall] 10648 10649 *) Fix the various library and apps files to free up pkeys obtained from 10650 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 10651 [Steve Henson] 10652 10653 *) Fix reference counting in X509_PUBKEY_get(). This makes 10654 demos/maurice/example2.c work, amongst others, probably. 10655 [Steve Henson and Ben Laurie] 10656 10657 *) First cut of a cleanup for apps/. First the `ssleay' program is now named 10658 `openssl' and second, the shortcut symlinks for the `openssl <command>' 10659 are no longer created. This way we have a single and consistent command 10660 line interface `openssl <command>', similar to `cvs <command>'. 10661 [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 10662 10663 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 10664 BIT STRING wrapper always have zero unused bits. 10665 [Steve Henson] 10666 10667 *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 10668 [Steve Henson] 10669 10670 *) Make the top-level INSTALL documentation easier to understand. 10671 [Paul Sutton] 10672 10673 *) Makefiles updated to exit if an error occurs in a sub-directory 10674 make (including if user presses ^C) [Paul Sutton] 10675 10676 *) Make Montgomery context stuff explicit in RSA data structure. 10677 [Ben Laurie] 10678 10679 *) Fix build order of pem and err to allow for generated pem.h. 10680 [Ben Laurie] 10681 10682 *) Fix renumbering bug in X509_NAME_delete_entry(). 10683 [Ben Laurie] 10684 10685 *) Enhanced the err-ins.pl script so it makes the error library number 10686 global and can add a library name. This is needed for external ASN1 and 10687 other error libraries. 10688 [Steve Henson] 10689 10690 *) Fixed sk_insert which never worked properly. 10691 [Steve Henson] 10692 10693 *) Fix ASN1 macros so they can handle indefinite length construted 10694 EXPLICIT tags. Some non standard certificates use these: they can now 10695 be read in. 10696 [Steve Henson] 10697 10698 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 10699 into a single doc/ssleay.txt bundle. This way the information is still 10700 preserved but no longer messes up this directory. Now it's new room for 10701 the new set of documenation files. 10702 [Ralf S. Engelschall] 10703 10704 *) SETs were incorrectly DER encoded. This was a major pain, because they 10705 shared code with SEQUENCEs, which aren't coded the same. This means that 10706 almost everything to do with SETs or SEQUENCEs has either changed name or 10707 number of arguments. 10708 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 10709 10710 *) Fix test data to work with the above. 10711 [Ben Laurie] 10712 10713 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 10714 was already fixed by Eric for 0.9.1 it seems. 10715 [Ben Laurie - pointed out by Ulf M��ller <ulf@fitug.de>] 10716 10717 *) Autodetect FreeBSD3. 10718 [Ben Laurie] 10719 10720 *) Fix various bugs in Configure. This affects the following platforms: 10721 nextstep 10722 ncr-scde 10723 unixware-2.0 10724 unixware-2.0-pentium 10725 sco5-cc. 10726 [Ben Laurie] 10727 10728 *) Eliminate generated files from CVS. Reorder tests to regenerate files 10729 before they are needed. 10730 [Ben Laurie] 10731 10732 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 10733 [Ben Laurie] 10734 10735 10736 Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 10737 10738 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 10739 changed SSLeay to OpenSSL in version strings. 10740 [Ralf S. Engelschall] 10741 10742 *) Some fixups to the top-level documents. 10743 [Paul Sutton] 10744 10745 *) Fixed the nasty bug where rsaref.h was not found under compile-time 10746 because the symlink to include/ was missing. 10747 [Ralf S. Engelschall] 10748 10749 *) Incorporated the popular no-RSA/DSA-only patches 10750 which allow to compile a RSA-free SSLeay. 10751 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 10752 10753 *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 10754 when "ssleay" is still not found. 10755 [Ralf S. Engelschall] 10756 10757 *) Added more platforms to Configure: Cray T3E, HPUX 11, 10758 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 10759 10760 *) Updated the README file. 10761 [Ralf S. Engelschall] 10762 10763 *) Added various .cvsignore files in the CVS repository subdirs 10764 to make a "cvs update" really silent. 10765 [Ralf S. Engelschall] 10766 10767 *) Recompiled the error-definition header files and added 10768 missing symbols to the Win32 linker tables. 10769 [Ralf S. Engelschall] 10770 10771 *) Cleaned up the top-level documents; 10772 o new files: CHANGES and LICENSE 10773 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 10774 o merged COPYRIGHT into LICENSE 10775 o removed obsolete TODO file 10776 o renamed MICROSOFT to INSTALL.W32 10777 [Ralf S. Engelschall] 10778 10779 *) Removed dummy files from the 0.9.1b source tree: 10780 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 10781 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 10782 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 10783 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 10784 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 10785 [Ralf S. Engelschall] 10786 10787 *) Added various platform portability fixes. 10788 [Mark J. Cox] 10789 10790 *) The Genesis of the OpenSSL rpject: 10791 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 10792 Young and Tim J. Hudson created while they were working for C2Net until 10793 summer 1998. 10794 [The OpenSSL Project] 10795 10796 10797 Changes between 0.9.0b and 0.9.1b [not released] 10798 10799 *) Updated a few CA certificates under certs/ 10800 [Eric A. Young] 10801 10802 *) Changed some BIGNUM api stuff. 10803 [Eric A. Young] 10804 10805 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 10806 DGUX x86, Linux Alpha, etc. 10807 [Eric A. Young] 10808 10809 *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 10810 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 10811 available). 10812 [Eric A. Young] 10813 10814 *) Add -strparse option to asn1pars program which parses nested 10815 binary structures 10816 [Dr Stephen Henson <shenson@bigfoot.com>] 10817 10818 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 10819 [Eric A. Young] 10820 10821 *) DSA fix for "ca" program. 10822 [Eric A. Young] 10823 10824 *) Added "-genkey" option to "dsaparam" program. 10825 [Eric A. Young] 10826 10827 *) Added RIPE MD160 (rmd160) message digest. 10828 [Eric A. Young] 10829 10830 *) Added -a (all) option to "ssleay version" command. 10831 [Eric A. Young] 10832 10833 *) Added PLATFORM define which is the id given to Configure. 10834 [Eric A. Young] 10835 10836 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 10837 [Eric A. Young] 10838 10839 *) Extended the ASN.1 parser routines. 10840 [Eric A. Young] 10841 10842 *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 10843 [Eric A. Young] 10844 10845 *) Added a BN_CTX to the BN library. 10846 [Eric A. Young] 10847 10848 *) Fixed the weak key values in DES library 10849 [Eric A. Young] 10850 10851 *) Changed API in EVP library for cipher aliases. 10852 [Eric A. Young] 10853 10854 *) Added support for RC2/64bit cipher. 10855 [Eric A. Young] 10856 10857 *) Converted the lhash library to the crypto/mem.c functions. 10858 [Eric A. Young] 10859 10860 *) Added more recognized ASN.1 object ids. 10861 [Eric A. Young] 10862 10863 *) Added more RSA padding checks for SSL/TLS. 10864 [Eric A. Young] 10865 10866 *) Added BIO proxy/filter functionality. 10867 [Eric A. Young] 10868 10869 *) Added extra_certs to SSL_CTX which can be used 10870 send extra CA certificates to the client in the CA cert chain sending 10871 process. It can be configured with SSL_CTX_add_extra_chain_cert(). 10872 [Eric A. Young] 10873 10874 *) Now Fortezza is denied in the authentication phase because 10875 this is key exchange mechanism is not supported by SSLeay at all. 10876 [Eric A. Young] 10877 10878 *) Additional PKCS1 checks. 10879 [Eric A. Young] 10880 10881 *) Support the string "TLSv1" for all TLS v1 ciphers. 10882 [Eric A. Young] 10883 10884 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 10885 ex_data index of the SSL context in the X509_STORE_CTX ex_data. 10886 [Eric A. Young] 10887 10888 *) Fixed a few memory leaks. 10889 [Eric A. Young] 10890 10891 *) Fixed various code and comment typos. 10892 [Eric A. Young] 10893 10894 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 10895 bytes sent in the client random. 10896 [Edward Bishop <ebishop@spyglass.com>] 10897 10898