CHANGES revision 306343
1 2 OpenSSL CHANGES 3 _______________ 4 5 Changes between 1.0.2i and 1.0.2j [26 Sep 2016] 6 7 *) Missing CRL sanity check 8 9 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0 10 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use 11 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception. 12 13 This issue only affects the OpenSSL 1.0.2i 14 (CVE-2016-7052) 15 [Matt Caswell] 16 17 Changes between 1.0.2h and 1.0.2i [22 Sep 2016] 18 19 *) OCSP Status Request extension unbounded memory growth 20 21 A malicious client can send an excessively large OCSP Status Request 22 extension. If that client continually requests renegotiation, sending a 23 large OCSP Status Request extension each time, then there will be unbounded 24 memory growth on the server. This will eventually lead to a Denial Of 25 Service attack through memory exhaustion. Servers with a default 26 configuration are vulnerable even if they do not support OCSP. Builds using 27 the "no-ocsp" build time option are not affected. 28 29 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 30 (CVE-2016-6304) 31 [Matt Caswell] 32 33 *) In order to mitigate the SWEET32 attack, the DES ciphers were moved from 34 HIGH to MEDIUM. 35 36 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan 37 Leurent (INRIA) 38 (CVE-2016-2183) 39 [Rich Salz] 40 41 *) OOB write in MDC2_Update() 42 43 An overflow can occur in MDC2_Update() either if called directly or 44 through the EVP_DigestUpdate() function using MDC2. If an attacker 45 is able to supply very large amounts of input data after a previous 46 call to EVP_EncryptUpdate() with a partial block then a length check 47 can overflow resulting in a heap corruption. 48 49 The amount of data needed is comparable to SIZE_MAX which is impractical 50 on most platforms. 51 52 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 53 (CVE-2016-6303) 54 [Stephen Henson] 55 56 *) Malformed SHA512 ticket DoS 57 58 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a 59 DoS attack where a malformed ticket will result in an OOB read which will 60 ultimately crash. 61 62 The use of SHA512 in TLS session tickets is comparatively rare as it requires 63 a custom server callback and ticket lookup mechanism. 64 65 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 66 (CVE-2016-6302) 67 [Stephen Henson] 68 69 *) OOB write in BN_bn2dec() 70 71 The function BN_bn2dec() does not check the return value of BN_div_word(). 72 This can cause an OOB write if an application uses this function with an 73 overly large BIGNUM. This could be a problem if an overly large certificate 74 or CRL is printed out from an untrusted source. TLS is not affected because 75 record limits will reject an oversized certificate before it is parsed. 76 77 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 78 (CVE-2016-2182) 79 [Stephen Henson] 80 81 *) OOB read in TS_OBJ_print_bio() 82 83 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is 84 the total length the OID text representation would use and not the amount 85 of data written. This will result in OOB reads when large OIDs are 86 presented. 87 88 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 89 (CVE-2016-2180) 90 [Stephen Henson] 91 92 *) Pointer arithmetic undefined behaviour 93 94 Avoid some undefined pointer arithmetic 95 96 A common idiom in the codebase is to check limits in the following manner: 97 "p + len > limit" 98 99 Where "p" points to some malloc'd data of SIZE bytes and 100 limit == p + SIZE 101 102 "len" here could be from some externally supplied data (e.g. from a TLS 103 message). 104 105 The rules of C pointer arithmetic are such that "p + len" is only well 106 defined where len <= SIZE. Therefore the above idiom is actually 107 undefined behaviour. 108 109 For example this could cause problems if some malloc implementation 110 provides an address for "p" such that "p + len" actually overflows for 111 values of len that are too big and therefore p + len < limit. 112 113 This issue was reported to OpenSSL by Guido Vranken 114 (CVE-2016-2177) 115 [Matt Caswell] 116 117 *) Constant time flag not preserved in DSA signing 118 119 Operations in the DSA signing algorithm should run in constant time in 120 order to avoid side channel attacks. A flaw in the OpenSSL DSA 121 implementation means that a non-constant time codepath is followed for 122 certain operations. This has been demonstrated through a cache-timing 123 attack to be sufficient for an attacker to recover the private DSA key. 124 125 This issue was reported by C��sar Pereida (Aalto University), Billy Brumley 126 (Tampere University of Technology), and Yuval Yarom (The University of 127 Adelaide and NICTA). 128 (CVE-2016-2178) 129 [C��sar Pereida] 130 131 *) DTLS buffered message DoS 132 133 In a DTLS connection where handshake messages are delivered out-of-order 134 those messages that OpenSSL is not yet ready to process will be buffered 135 for later use. Under certain circumstances, a flaw in the logic means that 136 those messages do not get removed from the buffer even though the handshake 137 has been completed. An attacker could force up to approx. 15 messages to 138 remain in the buffer when they are no longer required. These messages will 139 be cleared when the DTLS connection is closed. The default maximum size for 140 a message is 100k. Therefore the attacker could force an additional 1500k 141 to be consumed per connection. By opening many simulataneous connections an 142 attacker could cause a DoS attack through memory exhaustion. 143 144 This issue was reported to OpenSSL by Quan Luo. 145 (CVE-2016-2179) 146 [Matt Caswell] 147 148 *) DTLS replay protection DoS 149 150 A flaw in the DTLS replay attack protection mechanism means that records 151 that arrive for future epochs update the replay protection "window" before 152 the MAC for the record has been validated. This could be exploited by an 153 attacker by sending a record for the next epoch (which does not have to 154 decrypt or have a valid MAC), with a very large sequence number. This means 155 that all subsequent legitimate packets are dropped causing a denial of 156 service for a specific DTLS connection. 157 158 This issue was reported to OpenSSL by the OCAP audit team. 159 (CVE-2016-2181) 160 [Matt Caswell] 161 162 *) Certificate message OOB reads 163 164 In OpenSSL 1.0.2 and earlier some missing message length checks can result 165 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a 166 theoretical DoS risk but this has not been observed in practice on common 167 platforms. 168 169 The messages affected are client certificate, client certificate request 170 and server certificate. As a result the attack can only be performed 171 against a client or a server which enables client authentication. 172 173 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 174 (CVE-2016-6306) 175 [Stephen Henson] 176 177 Changes between 1.0.2g and 1.0.2h [3 May 2016] 178 179 *) Prevent padding oracle in AES-NI CBC MAC check 180 181 A MITM attacker can use a padding oracle attack to decrypt traffic 182 when the connection uses an AES CBC cipher and the server support 183 AES-NI. 184 185 This issue was introduced as part of the fix for Lucky 13 padding 186 attack (CVE-2013-0169). The padding check was rewritten to be in 187 constant time by making sure that always the same bytes are read and 188 compared against either the MAC or padding bytes. But it no longer 189 checked that there was enough data to have both the MAC and padding 190 bytes. 191 192 This issue was reported by Juraj Somorovsky using TLS-Attacker. 193 (CVE-2016-2107) 194 [Kurt Roeckx] 195 196 *) Fix EVP_EncodeUpdate overflow 197 198 An overflow can occur in the EVP_EncodeUpdate() function which is used for 199 Base64 encoding of binary data. If an attacker is able to supply very large 200 amounts of input data then a length check can overflow resulting in a heap 201 corruption. 202 203 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by 204 the PEM_write_bio* family of functions. These are mainly used within the 205 OpenSSL command line applications, so any application which processes data 206 from an untrusted source and outputs it as a PEM file should be considered 207 vulnerable to this issue. User applications that call these APIs directly 208 with large amounts of untrusted data may also be vulnerable. 209 210 This issue was reported by Guido Vranken. 211 (CVE-2016-2105) 212 [Matt Caswell] 213 214 *) Fix EVP_EncryptUpdate overflow 215 216 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker 217 is able to supply very large amounts of input data after a previous call to 218 EVP_EncryptUpdate() with a partial block then a length check can overflow 219 resulting in a heap corruption. Following an analysis of all OpenSSL 220 internal usage of the EVP_EncryptUpdate() function all usage is one of two 221 forms. The first form is where the EVP_EncryptUpdate() call is known to be 222 the first called function after an EVP_EncryptInit(), and therefore that 223 specific call must be safe. The second form is where the length passed to 224 EVP_EncryptUpdate() can be seen from the code to be some small value and 225 therefore there is no possibility of an overflow. Since all instances are 226 one of these two forms, it is believed that there can be no overflows in 227 internal code due to this problem. It should be noted that 228 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths. 229 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances 230 of these calls have also been analysed too and it is believed there are no 231 instances in internal usage where an overflow could occur. 232 233 This issue was reported by Guido Vranken. 234 (CVE-2016-2106) 235 [Matt Caswell] 236 237 *) Prevent ASN.1 BIO excessive memory allocation 238 239 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio() 240 a short invalid encoding can casuse allocation of large amounts of memory 241 potentially consuming excessive resources or exhausting memory. 242 243 Any application parsing untrusted data through d2i BIO functions is 244 affected. The memory based functions such as d2i_X509() are *not* affected. 245 Since the memory based functions are used by the TLS library, TLS 246 applications are not affected. 247 248 This issue was reported by Brian Carpenter. 249 (CVE-2016-2109) 250 [Stephen Henson] 251 252 *) EBCDIC overread 253 254 ASN1 Strings that are over 1024 bytes can cause an overread in applications 255 using the X509_NAME_oneline() function on EBCDIC systems. This could result 256 in arbitrary stack data being returned in the buffer. 257 258 This issue was reported by Guido Vranken. 259 (CVE-2016-2176) 260 [Matt Caswell] 261 262 *) Modify behavior of ALPN to invoke callback after SNI/servername 263 callback, such that updates to the SSL_CTX affect ALPN. 264 [Todd Short] 265 266 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the 267 default. 268 [Kurt Roeckx] 269 270 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the 271 methods are enabled and ssl2 is disabled the methods return NULL. 272 [Kurt Roeckx] 273 274 Changes between 1.0.2f and 1.0.2g [1 Mar 2016] 275 276 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL. 277 Builds that are not configured with "enable-weak-ssl-ciphers" will not 278 provide any "EXPORT" or "LOW" strength ciphers. 279 [Viktor Dukhovni] 280 281 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2 282 is by default disabled at build-time. Builds that are not configured with 283 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used, 284 users who want to negotiate SSLv2 via the version-flexible SSLv23_method() 285 will need to explicitly call either of: 286 287 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2); 288 or 289 SSL_clear_options(ssl, SSL_OP_NO_SSLv2); 290 291 as appropriate. Even if either of those is used, or the application 292 explicitly uses the version-specific SSLv2_method() or its client and 293 server variants, SSLv2 ciphers vulnerable to exhaustive search key 294 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT 295 ciphers, and SSLv2 56-bit DES are no longer available. 296 (CVE-2016-0800) 297 [Viktor Dukhovni] 298 299 *) Fix a double-free in DSA code 300 301 A double free bug was discovered when OpenSSL parses malformed DSA private 302 keys and could lead to a DoS attack or memory corruption for applications 303 that receive DSA private keys from untrusted sources. This scenario is 304 considered rare. 305 306 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using 307 libFuzzer. 308 (CVE-2016-0705) 309 [Stephen Henson] 310 311 *) Disable SRP fake user seed to address a server memory leak. 312 313 Add a new method SRP_VBASE_get1_by_user that handles the seed properly. 314 315 SRP_VBASE_get_by_user had inconsistent memory management behaviour. 316 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user 317 was changed to ignore the "fake user" SRP seed, even if the seed 318 is configured. 319 320 Users should use SRP_VBASE_get1_by_user instead. Note that in 321 SRP_VBASE_get1_by_user, caller must free the returned value. Note 322 also that even though configuring the SRP seed attempts to hide 323 invalid usernames by continuing the handshake with fake 324 credentials, this behaviour is not constant time and no strong 325 guarantees are made that the handshake is indistinguishable from 326 that of a valid user. 327 (CVE-2016-0798) 328 [Emilia K��sper] 329 330 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption 331 332 In the BN_hex2bn function the number of hex digits is calculated using an 333 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For 334 large values of |i| this can result in |bn_expand| not allocating any 335 memory because |i * 4| is negative. This can leave the internal BIGNUM data 336 field as NULL leading to a subsequent NULL ptr deref. For very large values 337 of |i|, the calculation |i * 4| could be a positive value smaller than |i|. 338 In this case memory is allocated to the internal BIGNUM data field, but it 339 is insufficiently sized leading to heap corruption. A similar issue exists 340 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn 341 is ever called by user applications with very large untrusted hex/dec data. 342 This is anticipated to be a rare occurrence. 343 344 All OpenSSL internal usage of these functions use data that is not expected 345 to be untrusted, e.g. config file data or application command line 346 arguments. If user developed applications generate config file data based 347 on untrusted data then it is possible that this could also lead to security 348 consequences. This is also anticipated to be rare. 349 350 This issue was reported to OpenSSL by Guido Vranken. 351 (CVE-2016-0797) 352 [Matt Caswell] 353 354 *) Fix memory issues in BIO_*printf functions 355 356 The internal |fmtstr| function used in processing a "%s" format string in 357 the BIO_*printf functions could overflow while calculating the length of a 358 string and cause an OOB read when printing very long strings. 359 360 Additionally the internal |doapr_outch| function can attempt to write to an 361 OOB memory location (at an offset from the NULL pointer) in the event of a 362 memory allocation failure. In 1.0.2 and below this could be caused where 363 the size of a buffer to be allocated is greater than INT_MAX. E.g. this 364 could be in processing a very long "%s" format string. Memory leaks can 365 also occur. 366 367 The first issue may mask the second issue dependent on compiler behaviour. 368 These problems could enable attacks where large amounts of untrusted data 369 is passed to the BIO_*printf functions. If applications use these functions 370 in this way then they could be vulnerable. OpenSSL itself uses these 371 functions when printing out human-readable dumps of ASN.1 data. Therefore 372 applications that print this data could be vulnerable if the data is from 373 untrusted sources. OpenSSL command line applications could also be 374 vulnerable where they print out ASN.1 data, or if untrusted data is passed 375 as command line arguments. 376 377 Libssl is not considered directly vulnerable. Additionally certificates etc 378 received via remote connections via libssl are also unlikely to be able to 379 trigger these issues because of message size limits enforced within libssl. 380 381 This issue was reported to OpenSSL Guido Vranken. 382 (CVE-2016-0799) 383 [Matt Caswell] 384 385 *) Side channel attack on modular exponentiation 386 387 A side-channel attack was found which makes use of cache-bank conflicts on 388 the Intel Sandy-Bridge microarchitecture which could lead to the recovery 389 of RSA keys. The ability to exploit this issue is limited as it relies on 390 an attacker who has control of code in a thread running on the same 391 hyper-threaded core as the victim thread which is performing decryptions. 392 393 This issue was reported to OpenSSL by Yuval Yarom, The University of 394 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and 395 Nadia Heninger, University of Pennsylvania with more information at 396 http://cachebleed.info. 397 (CVE-2016-0702) 398 [Andy Polyakov] 399 400 *) Change the req app to generate a 2048-bit RSA/DSA key by default, 401 if no keysize is specified with default_bits. This fixes an 402 omission in an earlier change that changed all RSA/DSA key generation 403 apps to use 2048 bits by default. 404 [Emilia K��sper] 405 406 Changes between 1.0.2e and 1.0.2f [28 Jan 2016] 407 408 *) DH small subgroups 409 410 Historically OpenSSL only ever generated DH parameters based on "safe" 411 primes. More recently (in version 1.0.2) support was provided for 412 generating X9.42 style parameter files such as those required for RFC 5114 413 support. The primes used in such files may not be "safe". Where an 414 application is using DH configured with parameters based on primes that are 415 not "safe" then an attacker could use this fact to find a peer's private 416 DH exponent. This attack requires that the attacker complete multiple 417 handshakes in which the peer uses the same private DH exponent. For example 418 this could be used to discover a TLS server's private DH exponent if it's 419 reusing the private DH exponent or it's using a static DH ciphersuite. 420 421 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in 422 TLS. It is not on by default. If the option is not set then the server 423 reuses the same private DH exponent for the life of the server process and 424 would be vulnerable to this attack. It is believed that many popular 425 applications do set this option and would therefore not be at risk. 426 427 The fix for this issue adds an additional check where a "q" parameter is 428 available (as is the case in X9.42 based parameters). This detects the 429 only known attack, and is the only possible defense for static DH 430 ciphersuites. This could have some performance impact. 431 432 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by 433 default and cannot be disabled. This could have some performance impact. 434 435 This issue was reported to OpenSSL by Antonio Sanso (Adobe). 436 (CVE-2016-0701) 437 [Matt Caswell] 438 439 *) SSLv2 doesn't block disabled ciphers 440 441 A malicious client can negotiate SSLv2 ciphers that have been disabled on 442 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have 443 been disabled, provided that the SSLv2 protocol was not also disabled via 444 SSL_OP_NO_SSLv2. 445 446 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram 447 and Sebastian Schinzel. 448 (CVE-2015-3197) 449 [Viktor Dukhovni] 450 451 *) Reject DH handshakes with parameters shorter than 1024 bits. 452 [Kurt Roeckx] 453 454 Changes between 1.0.2d and 1.0.2e [3 Dec 2015] 455 456 *) BN_mod_exp may produce incorrect results on x86_64 457 458 There is a carry propagating bug in the x86_64 Montgomery squaring 459 procedure. No EC algorithms are affected. Analysis suggests that attacks 460 against RSA and DSA as a result of this defect would be very difficult to 461 perform and are not believed likely. Attacks against DH are considered just 462 feasible (although very difficult) because most of the work necessary to 463 deduce information about a private key may be performed offline. The amount 464 of resources required for such an attack would be very significant and 465 likely only accessible to a limited number of attackers. An attacker would 466 additionally need online access to an unpatched system using the target 467 private key in a scenario with persistent DH parameters and a private 468 key that is shared between multiple clients. For example this can occur by 469 default in OpenSSL DHE based SSL/TLS ciphersuites. 470 471 This issue was reported to OpenSSL by Hanno B��ck. 472 (CVE-2015-3193) 473 [Andy Polyakov] 474 475 *) Certificate verify crash with missing PSS parameter 476 477 The signature verification routines will crash with a NULL pointer 478 dereference if presented with an ASN.1 signature using the RSA PSS 479 algorithm and absent mask generation function parameter. Since these 480 routines are used to verify certificate signature algorithms this can be 481 used to crash any certificate verification operation and exploited in a 482 DoS attack. Any application which performs certificate verification is 483 vulnerable including OpenSSL clients and servers which enable client 484 authentication. 485 486 This issue was reported to OpenSSL by Lo��c Jonas Etienne (Qnective AG). 487 (CVE-2015-3194) 488 [Stephen Henson] 489 490 *) X509_ATTRIBUTE memory leak 491 492 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak 493 memory. This structure is used by the PKCS#7 and CMS routines so any 494 application which reads PKCS#7 or CMS data from untrusted sources is 495 affected. SSL/TLS is not affected. 496 497 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using 498 libFuzzer. 499 (CVE-2015-3195) 500 [Stephen Henson] 501 502 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs. 503 This changes the decoding behaviour for some invalid messages, 504 though the change is mostly in the more lenient direction, and 505 legacy behaviour is preserved as much as possible. 506 [Emilia K��sper] 507 508 *) In DSA_generate_parameters_ex, if the provided seed is too short, 509 use a random seed, as already documented. 510 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>] 511 512 Changes between 1.0.2c and 1.0.2d [9 Jul 2015] 513 514 *) Alternate chains certificate forgery 515 516 During certificate verfification, OpenSSL will attempt to find an 517 alternative certificate chain if the first attempt to build such a chain 518 fails. An error in the implementation of this logic can mean that an 519 attacker could cause certain checks on untrusted certificates to be 520 bypassed, such as the CA flag, enabling them to use a valid leaf 521 certificate to act as a CA and "issue" an invalid certificate. 522 523 This issue was reported to OpenSSL by Adam Langley/David Benjamin 524 (Google/BoringSSL). 525 (CVE-2015-1793) 526 [Matt Caswell] 527 528 *) Race condition handling PSK identify hint 529 530 If PSK identity hints are received by a multi-threaded client then 531 the values are wrongly updated in the parent SSL_CTX structure. This can 532 result in a race condition potentially leading to a double free of the 533 identify hint data. 534 (CVE-2015-3196) 535 [Stephen Henson] 536 537 Changes between 1.0.2b and 1.0.2c [12 Jun 2015] 538 539 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI 540 incompatibility in the handling of HMAC. The previous ABI has now been 541 restored. 542 543 Changes between 1.0.2a and 1.0.2b [11 Jun 2015] 544 545 *) Malformed ECParameters causes infinite loop 546 547 When processing an ECParameters structure OpenSSL enters an infinite loop 548 if the curve specified is over a specially malformed binary polynomial 549 field. 550 551 This can be used to perform denial of service against any 552 system which processes public keys, certificate requests or 553 certificates. This includes TLS clients and TLS servers with 554 client authentication enabled. 555 556 This issue was reported to OpenSSL by Joseph Barr-Pixton. 557 (CVE-2015-1788) 558 [Andy Polyakov] 559 560 *) Exploitable out-of-bounds read in X509_cmp_time 561 562 X509_cmp_time does not properly check the length of the ASN1_TIME 563 string and can read a few bytes out of bounds. In addition, 564 X509_cmp_time accepts an arbitrary number of fractional seconds in the 565 time string. 566 567 An attacker can use this to craft malformed certificates and CRLs of 568 various sizes and potentially cause a segmentation fault, resulting in 569 a DoS on applications that verify certificates or CRLs. TLS clients 570 that verify CRLs are affected. TLS clients and servers with client 571 authentication enabled may be affected if they use custom verification 572 callbacks. 573 574 This issue was reported to OpenSSL by Robert Swiecki (Google), and 575 independently by Hanno B��ck. 576 (CVE-2015-1789) 577 [Emilia K��sper] 578 579 *) PKCS7 crash with missing EnvelopedContent 580 581 The PKCS#7 parsing code does not handle missing inner EncryptedContent 582 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs 583 with missing content and trigger a NULL pointer dereference on parsing. 584 585 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7 586 structures from untrusted sources are affected. OpenSSL clients and 587 servers are not affected. 588 589 This issue was reported to OpenSSL by Michal Zalewski (Google). 590 (CVE-2015-1790) 591 [Emilia K��sper] 592 593 *) CMS verify infinite loop with unknown hash function 594 595 When verifying a signedData message the CMS code can enter an infinite loop 596 if presented with an unknown hash function OID. This can be used to perform 597 denial of service against any system which verifies signedData messages using 598 the CMS code. 599 This issue was reported to OpenSSL by Johannes Bauer. 600 (CVE-2015-1792) 601 [Stephen Henson] 602 603 *) Race condition handling NewSessionTicket 604 605 If a NewSessionTicket is received by a multi-threaded client when attempting to 606 reuse a previous ticket then a race condition can occur potentially leading to 607 a double free of the ticket data. 608 (CVE-2015-1791) 609 [Matt Caswell] 610 611 *) Removed support for the two export grade static DH ciphersuites 612 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites 613 were newly added (along with a number of other static DH ciphersuites) to 614 1.0.2. However the two export ones have *never* worked since they were 615 introduced. It seems strange in any case to be adding new export 616 ciphersuites, and given "logjam" it also does not seem correct to fix them. 617 [Matt Caswell] 618 619 *) Only support 256-bit or stronger elliptic curves with the 620 'ecdh_auto' setting (server) or by default (client). Of supported 621 curves, prefer P-256 (both). 622 [Emilia Kasper] 623 624 *) Reject DH handshakes with parameters shorter than 768 bits. 625 [Kurt Roeckx and Emilia Kasper] 626 627 Changes between 1.0.2 and 1.0.2a [19 Mar 2015] 628 629 *) ClientHello sigalgs DoS fix 630 631 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an 632 invalid signature algorithms extension a NULL pointer dereference will 633 occur. This can be exploited in a DoS attack against the server. 634 635 This issue was was reported to OpenSSL by David Ramos of Stanford 636 University. 637 (CVE-2015-0291) 638 [Stephen Henson and Matt Caswell] 639 640 *) Multiblock corrupted pointer fix 641 642 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This 643 feature only applies on 64 bit x86 architecture platforms that support AES 644 NI instructions. A defect in the implementation of "multiblock" can cause 645 OpenSSL's internal write buffer to become incorrectly set to NULL when 646 using non-blocking IO. Typically, when the user application is using a 647 socket BIO for writing, this will only result in a failed connection. 648 However if some other BIO is used then it is likely that a segmentation 649 fault will be triggered, thus enabling a potential DoS attack. 650 651 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller. 652 (CVE-2015-0290) 653 [Matt Caswell] 654 655 *) Segmentation fault in DTLSv1_listen fix 656 657 The DTLSv1_listen function is intended to be stateless and processes the 658 initial ClientHello from many peers. It is common for user code to loop 659 over the call to DTLSv1_listen until a valid ClientHello is received with 660 an associated cookie. A defect in the implementation of DTLSv1_listen means 661 that state is preserved in the SSL object from one invocation to the next 662 that can lead to a segmentation fault. Errors processing the initial 663 ClientHello can trigger this scenario. An example of such an error could be 664 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only 665 server. 666 667 This issue was reported to OpenSSL by Per Allansson. 668 (CVE-2015-0207) 669 [Matt Caswell] 670 671 *) Segmentation fault in ASN1_TYPE_cmp fix 672 673 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is 674 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check 675 certificate signature algorithm consistency this can be used to crash any 676 certificate verification operation and exploited in a DoS attack. Any 677 application which performs certificate verification is vulnerable including 678 OpenSSL clients and servers which enable client authentication. 679 (CVE-2015-0286) 680 [Stephen Henson] 681 682 *) Segmentation fault for invalid PSS parameters fix 683 684 The signature verification routines will crash with a NULL pointer 685 dereference if presented with an ASN.1 signature using the RSA PSS 686 algorithm and invalid parameters. Since these routines are used to verify 687 certificate signature algorithms this can be used to crash any 688 certificate verification operation and exploited in a DoS attack. Any 689 application which performs certificate verification is vulnerable including 690 OpenSSL clients and servers which enable client authentication. 691 692 This issue was was reported to OpenSSL by Brian Carpenter. 693 (CVE-2015-0208) 694 [Stephen Henson] 695 696 *) ASN.1 structure reuse memory corruption fix 697 698 Reusing a structure in ASN.1 parsing may allow an attacker to cause 699 memory corruption via an invalid write. Such reuse is and has been 700 strongly discouraged and is believed to be rare. 701 702 Applications that parse structures containing CHOICE or ANY DEFINED BY 703 components may be affected. Certificate parsing (d2i_X509 and related 704 functions) are however not affected. OpenSSL clients and servers are 705 not affected. 706 (CVE-2015-0287) 707 [Stephen Henson] 708 709 *) PKCS7 NULL pointer dereferences fix 710 711 The PKCS#7 parsing code does not handle missing outer ContentInfo 712 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with 713 missing content and trigger a NULL pointer dereference on parsing. 714 715 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or 716 otherwise parse PKCS#7 structures from untrusted sources are 717 affected. OpenSSL clients and servers are not affected. 718 719 This issue was reported to OpenSSL by Michal Zalewski (Google). 720 (CVE-2015-0289) 721 [Emilia K��sper] 722 723 *) DoS via reachable assert in SSLv2 servers fix 724 725 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in 726 servers that both support SSLv2 and enable export cipher suites by sending 727 a specially crafted SSLv2 CLIENT-MASTER-KEY message. 728 729 This issue was discovered by Sean Burford (Google) and Emilia K��sper 730 (OpenSSL development team). 731 (CVE-2015-0293) 732 [Emilia K��sper] 733 734 *) Empty CKE with client auth and DHE fix 735 736 If client auth is used then a server can seg fault in the event of a DHE 737 ciphersuite being selected and a zero length ClientKeyExchange message 738 being sent by the client. This could be exploited in a DoS attack. 739 (CVE-2015-1787) 740 [Matt Caswell] 741 742 *) Handshake with unseeded PRNG fix 743 744 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake 745 with an unseeded PRNG. The conditions are: 746 - The client is on a platform where the PRNG has not been seeded 747 automatically, and the user has not seeded manually 748 - A protocol specific client method version has been used (i.e. not 749 SSL_client_methodv23) 750 - A ciphersuite is used that does not require additional random data from 751 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA). 752 753 If the handshake succeeds then the client random that has been used will 754 have been generated from a PRNG with insufficient entropy and therefore the 755 output may be predictable. 756 757 For example using the following command with an unseeded openssl will 758 succeed on an unpatched platform: 759 760 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA 761 (CVE-2015-0285) 762 [Matt Caswell] 763 764 *) Use After Free following d2i_ECPrivatekey error fix 765 766 A malformed EC private key file consumed via the d2i_ECPrivateKey function 767 could cause a use after free condition. This, in turn, could cause a double 768 free in several private key parsing functions (such as d2i_PrivateKey 769 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption 770 for applications that receive EC private keys from untrusted 771 sources. This scenario is considered rare. 772 773 This issue was discovered by the BoringSSL project and fixed in their 774 commit 517073cd4b. 775 (CVE-2015-0209) 776 [Matt Caswell] 777 778 *) X509_to_X509_REQ NULL pointer deref fix 779 780 The function X509_to_X509_REQ will crash with a NULL pointer dereference if 781 the certificate key is invalid. This function is rarely used in practice. 782 783 This issue was discovered by Brian Carpenter. 784 (CVE-2015-0288) 785 [Stephen Henson] 786 787 *) Removed the export ciphers from the DEFAULT ciphers 788 [Kurt Roeckx] 789 790 Changes between 1.0.1l and 1.0.2 [22 Jan 2015] 791 792 *) Change RSA and DH/DSA key generation apps to generate 2048-bit 793 keys by default. 794 [Kurt Roeckx] 795 796 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g. 797 ARMv5 through ARMv8, as opposite to "locking" it to single one. 798 So far those who have to target multiple plaforms would compromise 799 and argue that binary targeting say ARMv5 would still execute on 800 ARMv8. "Universal" build resolves this compromise by providing 801 near-optimal performance even on newer platforms. 802 [Andy Polyakov] 803 804 *) Accelerated NIST P-256 elliptic curve implementation for x86_64 805 (other platforms pending). 806 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov] 807 808 *) Add support for the SignedCertificateTimestampList certificate and 809 OCSP response extensions from RFC6962. 810 [Rob Stradling] 811 812 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) 813 for corner cases. (Certain input points at infinity could lead to 814 bogus results, with non-infinity inputs mapped to infinity too.) 815 [Bodo Moeller] 816 817 *) Initial support for PowerISA 2.0.7, first implemented in POWER8. 818 This covers AES, SHA256/512 and GHASH. "Initial" means that most 819 common cases are optimized and there still is room for further 820 improvements. Vector Permutation AES for Altivec is also added. 821 [Andy Polyakov] 822 823 *) Add support for little-endian ppc64 Linux target. 824 [Marcelo Cerri (IBM)] 825 826 *) Initial support for AMRv8 ISA crypto extensions. This covers AES, 827 SHA1, SHA256 and GHASH. "Initial" means that most common cases 828 are optimized and there still is room for further improvements. 829 Both 32- and 64-bit modes are supported. 830 [Andy Polyakov, Ard Biesheuvel (Linaro)] 831 832 *) Improved ARMv7 NEON support. 833 [Andy Polyakov] 834 835 *) Support for SPARC Architecture 2011 crypto extensions, first 836 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1, 837 SHA256/512, MD5, GHASH and modular exponentiation. 838 [Andy Polyakov, David Miller] 839 840 *) Accelerated modular exponentiation for Intel processors, a.k.a. 841 RSAZ. 842 [Shay Gueron & Vlad Krasnov (Intel Corp)] 843 844 *) Support for new and upcoming Intel processors, including AVX2, 845 BMI and SHA ISA extensions. This includes additional "stitched" 846 implementations, AESNI-SHA256 and GCM, and multi-buffer support 847 for TLS encrypt. 848 849 This work was sponsored by Intel Corp. 850 [Andy Polyakov] 851 852 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method() 853 supports both DTLS 1.2 and 1.0 and should use whatever version the peer 854 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only. 855 [Steve Henson] 856 857 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file(): 858 this fixes a limiation in previous versions of OpenSSL. 859 [Steve Henson] 860 861 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest, 862 MGF1 digest and OAEP label. 863 [Steve Henson] 864 865 *) Add EVP support for key wrapping algorithms, to avoid problems with 866 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in 867 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap 868 algorithms and include tests cases. 869 [Steve Henson] 870 871 *) Add functions to allocate and set the fields of an ECDSA_METHOD 872 structure. 873 [Douglas E. Engert, Steve Henson] 874 875 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the 876 difference in days and seconds between two tm or ASN1_TIME structures. 877 [Steve Henson] 878 879 *) Add -rev test option to s_server to just reverse order of characters 880 received by client and send back to server. Also prints an abbreviated 881 summary of the connection parameters. 882 [Steve Henson] 883 884 *) New option -brief for s_client and s_server to print out a brief summary 885 of connection parameters. 886 [Steve Henson] 887 888 *) Add callbacks for arbitrary TLS extensions. 889 [Trevor Perrin <trevp@trevp.net> and Ben Laurie] 890 891 *) New option -crl_download in several openssl utilities to download CRLs 892 from CRLDP extension in certificates. 893 [Steve Henson] 894 895 *) New options -CRL and -CRLform for s_client and s_server for CRLs. 896 [Steve Henson] 897 898 *) New function X509_CRL_diff to generate a delta CRL from the difference 899 of two full CRLs. Add support to "crl" utility. 900 [Steve Henson] 901 902 *) New functions to set lookup_crls function and to retrieve 903 X509_STORE from X509_STORE_CTX. 904 [Steve Henson] 905 906 *) Print out deprecated issuer and subject unique ID fields in 907 certificates. 908 [Steve Henson] 909 910 *) Extend OCSP I/O functions so they can be used for simple general purpose 911 HTTP as well as OCSP. New wrapper function which can be used to download 912 CRLs using the OCSP API. 913 [Steve Henson] 914 915 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs. 916 [Steve Henson] 917 918 *) SSL_CONF* functions. These provide a common framework for application 919 configuration using configuration files or command lines. 920 [Steve Henson] 921 922 *) SSL/TLS tracing code. This parses out SSL/TLS records using the 923 message callback and prints the results. Needs compile time option 924 "enable-ssl-trace". New options to s_client and s_server to enable 925 tracing. 926 [Steve Henson] 927 928 *) New ctrl and macro to retrieve supported points extensions. 929 Print out extension in s_server and s_client. 930 [Steve Henson] 931 932 *) New functions to retrieve certificate signature and signature 933 OID NID. 934 [Steve Henson] 935 936 *) Add functions to retrieve and manipulate the raw cipherlist sent by a 937 client to OpenSSL. 938 [Steve Henson] 939 940 *) New Suite B modes for TLS code. These use and enforce the requirements 941 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and 942 only use Suite B curves. The Suite B modes can be set by using the 943 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring. 944 [Steve Henson] 945 946 *) New chain verification flags for Suite B levels of security. Check 947 algorithms are acceptable when flags are set in X509_verify_cert. 948 [Steve Henson] 949 950 *) Make tls1_check_chain return a set of flags indicating checks passed 951 by a certificate chain. Add additional tests to handle client 952 certificates: checks for matching certificate type and issuer name 953 comparison. 954 [Steve Henson] 955 956 *) If an attempt is made to use a signature algorithm not in the peer 957 preference list abort the handshake. If client has no suitable 958 signature algorithms in response to a certificate request do not 959 use the certificate. 960 [Steve Henson] 961 962 *) If server EC tmp key is not in client preference list abort handshake. 963 [Steve Henson] 964 965 *) Add support for certificate stores in CERT structure. This makes it 966 possible to have different stores per SSL structure or one store in 967 the parent SSL_CTX. Include distint stores for certificate chain 968 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN 969 to build and store a certificate chain in CERT structure: returing 970 an error if the chain cannot be built: this will allow applications 971 to test if a chain is correctly configured. 972 973 Note: if the CERT based stores are not set then the parent SSL_CTX 974 store is used to retain compatibility with existing behaviour. 975 976 [Steve Henson] 977 978 *) New function ssl_set_client_disabled to set a ciphersuite disabled 979 mask based on the current session, check mask when sending client 980 hello and checking the requested ciphersuite. 981 [Steve Henson] 982 983 *) New ctrls to retrieve and set certificate types in a certificate 984 request message. Print out received values in s_client. If certificate 985 types is not set with custom values set sensible values based on 986 supported signature algorithms. 987 [Steve Henson] 988 989 *) Support for distinct client and server supported signature algorithms. 990 [Steve Henson] 991 992 *) Add certificate callback. If set this is called whenever a certificate 993 is required by client or server. An application can decide which 994 certificate chain to present based on arbitrary criteria: for example 995 supported signature algorithms. Add very simple example to s_server. 996 This fixes many of the problems and restrictions of the existing client 997 certificate callback: for example you can now clear an existing 998 certificate and specify the whole chain. 999 [Steve Henson] 1000 1001 *) Add new "valid_flags" field to CERT_PKEY structure which determines what 1002 the certificate can be used for (if anything). Set valid_flags field 1003 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used 1004 to have similar checks in it. 1005 1006 Add new "cert_flags" field to CERT structure and include a "strict mode". 1007 This enforces some TLS certificate requirements (such as only permitting 1008 certificate signature algorithms contained in the supported algorithms 1009 extension) which some implementations ignore: this option should be used 1010 with caution as it could cause interoperability issues. 1011 [Steve Henson] 1012 1013 *) Update and tidy signature algorithm extension processing. Work out 1014 shared signature algorithms based on preferences and peer algorithms 1015 and print them out in s_client and s_server. Abort handshake if no 1016 shared signature algorithms. 1017 [Steve Henson] 1018 1019 *) Add new functions to allow customised supported signature algorithms 1020 for SSL and SSL_CTX structures. Add options to s_client and s_server 1021 to support them. 1022 [Steve Henson] 1023 1024 *) New function SSL_certs_clear() to delete all references to certificates 1025 from an SSL structure. Before this once a certificate had been added 1026 it couldn't be removed. 1027 [Steve Henson] 1028 1029 *) Integrate hostname, email address and IP address checking with certificate 1030 verification. New verify options supporting checking in opensl utility. 1031 [Steve Henson] 1032 1033 *) Fixes and wildcard matching support to hostname and email checking 1034 functions. Add manual page. 1035 [Florian Weimer (Red Hat Product Security Team)] 1036 1037 *) New functions to check a hostname email or IP address against a 1038 certificate. Add options x509 utility to print results of checks against 1039 a certificate. 1040 [Steve Henson] 1041 1042 *) Fix OCSP checking. 1043 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie] 1044 1045 *) Initial experimental support for explicitly trusted non-root CAs. 1046 OpenSSL still tries to build a complete chain to a root but if an 1047 intermediate CA has a trust setting included that is used. The first 1048 setting is used: whether to trust (e.g., -addtrust option to the x509 1049 utility) or reject. 1050 [Steve Henson] 1051 1052 *) Add -trusted_first option which attempts to find certificates in the 1053 trusted store even if an untrusted chain is also supplied. 1054 [Steve Henson] 1055 1056 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE, 1057 platform support for Linux and Android. 1058 [Andy Polyakov] 1059 1060 *) Support for linux-x32, ILP32 environment in x86_64 framework. 1061 [Andy Polyakov] 1062 1063 *) Experimental multi-implementation support for FIPS capable OpenSSL. 1064 When in FIPS mode the approved implementations are used as normal, 1065 when not in FIPS mode the internal unapproved versions are used instead. 1066 This means that the FIPS capable OpenSSL isn't forced to use the 1067 (often lower perfomance) FIPS implementations outside FIPS mode. 1068 [Steve Henson] 1069 1070 *) Transparently support X9.42 DH parameters when calling 1071 PEM_read_bio_DHparameters. This means existing applications can handle 1072 the new parameter format automatically. 1073 [Steve Henson] 1074 1075 *) Initial experimental support for X9.42 DH parameter format: mainly 1076 to support use of 'q' parameter for RFC5114 parameters. 1077 [Steve Henson] 1078 1079 *) Add DH parameters from RFC5114 including test data to dhtest. 1080 [Steve Henson] 1081 1082 *) Support for automatic EC temporary key parameter selection. If enabled 1083 the most preferred EC parameters are automatically used instead of 1084 hardcoded fixed parameters. Now a server just has to call: 1085 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically 1086 support ECDH and use the most appropriate parameters. 1087 [Steve Henson] 1088 1089 *) Enhance and tidy EC curve and point format TLS extension code. Use 1090 static structures instead of allocation if default values are used. 1091 New ctrls to set curves we wish to support and to retrieve shared curves. 1092 Print out shared curves in s_server. New options to s_server and s_client 1093 to set list of supported curves. 1094 [Steve Henson] 1095 1096 *) New ctrls to retrieve supported signature algorithms and 1097 supported curve values as an array of NIDs. Extend openssl utility 1098 to print out received values. 1099 [Steve Henson] 1100 1101 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert 1102 between NIDs and the more common NIST names such as "P-256". Enhance 1103 ecparam utility and ECC method to recognise the NIST names for curves. 1104 [Steve Henson] 1105 1106 *) Enhance SSL/TLS certificate chain handling to support different 1107 chains for each certificate instead of one chain in the parent SSL_CTX. 1108 [Steve Henson] 1109 1110 *) Support for fixed DH ciphersuite client authentication: where both 1111 server and client use DH certificates with common parameters. 1112 [Steve Henson] 1113 1114 *) Support for fixed DH ciphersuites: those requiring DH server 1115 certificates. 1116 [Steve Henson] 1117 1118 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of 1119 the certificate. 1120 Note: Related 1.0.2-beta specific macros X509_get_cert_info, 1121 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and 1122 X509_CINF_get_signature were reverted post internal team review. 1123 1124 Changes between 1.0.1k and 1.0.1l [15 Jan 2015] 1125 1126 *) Build fixes for the Windows and OpenVMS platforms 1127 [Matt Caswell and Richard Levitte] 1128 1129 Changes between 1.0.1j and 1.0.1k [8 Jan 2015] 1130 1131 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS 1132 message can cause a segmentation fault in OpenSSL due to a NULL pointer 1133 dereference. This could lead to a Denial Of Service attack. Thanks to 1134 Markus Stenberg of Cisco Systems, Inc. for reporting this issue. 1135 (CVE-2014-3571) 1136 [Steve Henson] 1137 1138 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the 1139 dtls1_buffer_record function under certain conditions. In particular this 1140 could occur if an attacker sent repeated DTLS records with the same 1141 sequence number but for the next epoch. The memory leak could be exploited 1142 by an attacker in a Denial of Service attack through memory exhaustion. 1143 Thanks to Chris Mueller for reporting this issue. 1144 (CVE-2015-0206) 1145 [Matt Caswell] 1146 1147 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is 1148 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl 1149 method would be set to NULL which could later result in a NULL pointer 1150 dereference. Thanks to Frank Schmirler for reporting this issue. 1151 (CVE-2014-3569) 1152 [Kurt Roeckx] 1153 1154 *) Abort handshake if server key exchange message is omitted for ephemeral 1155 ECDH ciphersuites. 1156 1157 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for 1158 reporting this issue. 1159 (CVE-2014-3572) 1160 [Steve Henson] 1161 1162 *) Remove non-export ephemeral RSA code on client and server. This code 1163 violated the TLS standard by allowing the use of temporary RSA keys in 1164 non-export ciphersuites and could be used by a server to effectively 1165 downgrade the RSA key length used to a value smaller than the server 1166 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at 1167 INRIA or reporting this issue. 1168 (CVE-2015-0204) 1169 [Steve Henson] 1170 1171 *) Fixed issue where DH client certificates are accepted without verification. 1172 An OpenSSL server will accept a DH certificate for client authentication 1173 without the certificate verify message. This effectively allows a client to 1174 authenticate without the use of a private key. This only affects servers 1175 which trust a client certificate authority which issues certificates 1176 containing DH keys: these are extremely rare and hardly ever encountered. 1177 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting 1178 this issue. 1179 (CVE-2015-0205) 1180 [Steve Henson] 1181 1182 *) Ensure that the session ID context of an SSL is updated when its 1183 SSL_CTX is updated via SSL_set_SSL_CTX. 1184 1185 The session ID context is typically set from the parent SSL_CTX, 1186 and can vary with the CTX. 1187 [Adam Langley] 1188 1189 *) Fix various certificate fingerprint issues. 1190 1191 By using non-DER or invalid encodings outside the signed portion of a 1192 certificate the fingerprint can be changed without breaking the signature. 1193 Although no details of the signed portion of the certificate can be changed 1194 this can cause problems with some applications: e.g. those using the 1195 certificate fingerprint for blacklists. 1196 1197 1. Reject signatures with non zero unused bits. 1198 1199 If the BIT STRING containing the signature has non zero unused bits reject 1200 the signature. All current signature algorithms require zero unused bits. 1201 1202 2. Check certificate algorithm consistency. 1203 1204 Check the AlgorithmIdentifier inside TBS matches the one in the 1205 certificate signature. NB: this will result in signature failure 1206 errors for some broken certificates. 1207 1208 Thanks to Konrad Kraszewski from Google for reporting this issue. 1209 1210 3. Check DSA/ECDSA signatures use DER. 1211 1212 Reencode DSA/ECDSA signatures and compare with the original received 1213 signature. Return an error if there is a mismatch. 1214 1215 This will reject various cases including garbage after signature 1216 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS 1217 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs 1218 (negative or with leading zeroes). 1219 1220 Further analysis was conducted and fixes were developed by Stephen Henson 1221 of the OpenSSL core team. 1222 1223 (CVE-2014-8275) 1224 [Steve Henson] 1225 1226 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect 1227 results on some platforms, including x86_64. This bug occurs at random 1228 with a very low probability, and is not known to be exploitable in any 1229 way, though its exact impact is difficult to determine. Thanks to Pieter 1230 Wuille (Blockstream) who reported this issue and also suggested an initial 1231 fix. Further analysis was conducted by the OpenSSL development team and 1232 Adam Langley of Google. The final fix was developed by Andy Polyakov of 1233 the OpenSSL core team. 1234 (CVE-2014-3570) 1235 [Andy Polyakov] 1236 1237 *) Do not resume sessions on the server if the negotiated protocol 1238 version does not match the session's version. Resuming with a different 1239 version, while not strictly forbidden by the RFC, is of questionable 1240 sanity and breaks all known clients. 1241 [David Benjamin, Emilia K��sper] 1242 1243 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject 1244 early CCS messages during renegotiation. (Note that because 1245 renegotiation is encrypted, this early CCS was not exploitable.) 1246 [Emilia K��sper] 1247 1248 *) Tighten client-side session ticket handling during renegotiation: 1249 ensure that the client only accepts a session ticket if the server sends 1250 the extension anew in the ServerHello. Previously, a TLS client would 1251 reuse the old extension state and thus accept a session ticket if one was 1252 announced in the initial ServerHello. 1253 1254 Similarly, ensure that the client requires a session ticket if one 1255 was advertised in the ServerHello. Previously, a TLS client would 1256 ignore a missing NewSessionTicket message. 1257 [Emilia K��sper] 1258 1259 Changes between 1.0.1i and 1.0.1j [15 Oct 2014] 1260 1261 *) SRTP Memory Leak. 1262 1263 A flaw in the DTLS SRTP extension parsing code allows an attacker, who 1264 sends a carefully crafted handshake message, to cause OpenSSL to fail 1265 to free up to 64k of memory causing a memory leak. This could be 1266 exploited in a Denial Of Service attack. This issue affects OpenSSL 1267 1.0.1 server implementations for both SSL/TLS and DTLS regardless of 1268 whether SRTP is used or configured. Implementations of OpenSSL that 1269 have been compiled with OPENSSL_NO_SRTP defined are not affected. 1270 1271 The fix was developed by the OpenSSL team. 1272 (CVE-2014-3513) 1273 [OpenSSL team] 1274 1275 *) Session Ticket Memory Leak. 1276 1277 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the 1278 integrity of that ticket is first verified. In the event of a session 1279 ticket integrity check failing, OpenSSL will fail to free memory 1280 causing a memory leak. By sending a large number of invalid session 1281 tickets an attacker could exploit this issue in a Denial Of Service 1282 attack. 1283 (CVE-2014-3567) 1284 [Steve Henson] 1285 1286 *) Build option no-ssl3 is incomplete. 1287 1288 When OpenSSL is configured with "no-ssl3" as a build option, servers 1289 could accept and complete a SSL 3.0 handshake, and clients could be 1290 configured to send them. 1291 (CVE-2014-3568) 1292 [Akamai and the OpenSSL team] 1293 1294 *) Add support for TLS_FALLBACK_SCSV. 1295 Client applications doing fallback retries should call 1296 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV). 1297 (CVE-2014-3566) 1298 [Adam Langley, Bodo Moeller] 1299 1300 *) Add additional DigestInfo checks. 1301 1302 Reencode DigestInto in DER and check against the original when 1303 verifying RSA signature: this will reject any improperly encoded 1304 DigestInfo structures. 1305 1306 Note: this is a precautionary measure and no attacks are currently known. 1307 1308 [Steve Henson] 1309 1310 Changes between 1.0.1h and 1.0.1i [6 Aug 2014] 1311 1312 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the 1313 SRP code can be overrun an internal buffer. Add sanity check that 1314 g, A, B < N to SRP code. 1315 1316 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC 1317 Group for discovering this issue. 1318 (CVE-2014-3512) 1319 [Steve Henson] 1320 1321 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate 1322 TLS 1.0 instead of higher protocol versions when the ClientHello message 1323 is badly fragmented. This allows a man-in-the-middle attacker to force a 1324 downgrade to TLS 1.0 even if both the server and the client support a 1325 higher protocol version, by modifying the client's TLS records. 1326 1327 Thanks to David Benjamin and Adam Langley (Google) for discovering and 1328 researching this issue. 1329 (CVE-2014-3511) 1330 [David Benjamin] 1331 1332 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject 1333 to a denial of service attack. A malicious server can crash the client 1334 with a null pointer dereference (read) by specifying an anonymous (EC)DH 1335 ciphersuite and sending carefully crafted handshake messages. 1336 1337 Thanks to Felix Gr��bert (Google) for discovering and researching this 1338 issue. 1339 (CVE-2014-3510) 1340 [Emilia K��sper] 1341 1342 *) By sending carefully crafted DTLS packets an attacker could cause openssl 1343 to leak memory. This can be exploited through a Denial of Service attack. 1344 Thanks to Adam Langley for discovering and researching this issue. 1345 (CVE-2014-3507) 1346 [Adam Langley] 1347 1348 *) An attacker can force openssl to consume large amounts of memory whilst 1349 processing DTLS handshake messages. This can be exploited through a 1350 Denial of Service attack. 1351 Thanks to Adam Langley for discovering and researching this issue. 1352 (CVE-2014-3506) 1353 [Adam Langley] 1354 1355 *) An attacker can force an error condition which causes openssl to crash 1356 whilst processing DTLS packets due to memory being freed twice. This 1357 can be exploited through a Denial of Service attack. 1358 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching 1359 this issue. 1360 (CVE-2014-3505) 1361 [Adam Langley] 1362 1363 *) If a multithreaded client connects to a malicious server using a resumed 1364 session and the server sends an ec point format extension it could write 1365 up to 255 bytes to freed memory. 1366 1367 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this 1368 issue. 1369 (CVE-2014-3509) 1370 [Gabor Tyukasz] 1371 1372 *) A malicious server can crash an OpenSSL client with a null pointer 1373 dereference (read) by specifying an SRP ciphersuite even though it was not 1374 properly negotiated with the client. This can be exploited through a 1375 Denial of Service attack. 1376 1377 Thanks to Joonas Kuorilehto and Riku Hietam��ki (Codenomicon) for 1378 discovering and researching this issue. 1379 (CVE-2014-5139) 1380 [Steve Henson] 1381 1382 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as 1383 X509_name_oneline, X509_name_print_ex et al. to leak some information 1384 from the stack. Applications may be affected if they echo pretty printing 1385 output to the attacker. 1386 1387 Thanks to Ivan Fratric (Google) for discovering this issue. 1388 (CVE-2014-3508) 1389 [Emilia K��sper, and Steve Henson] 1390 1391 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) 1392 for corner cases. (Certain input points at infinity could lead to 1393 bogus results, with non-infinity inputs mapped to infinity too.) 1394 [Bodo Moeller] 1395 1396 Changes between 1.0.1g and 1.0.1h [5 Jun 2014] 1397 1398 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted 1399 handshake can force the use of weak keying material in OpenSSL 1400 SSL/TLS clients and servers. 1401 1402 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and 1403 researching this issue. (CVE-2014-0224) 1404 [KIKUCHI Masashi, Steve Henson] 1405 1406 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an 1407 OpenSSL DTLS client the code can be made to recurse eventually crashing 1408 in a DoS attack. 1409 1410 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue. 1411 (CVE-2014-0221) 1412 [Imre Rad, Steve Henson] 1413 1414 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can 1415 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS 1416 client or server. This is potentially exploitable to run arbitrary 1417 code on a vulnerable client or server. 1418 1419 Thanks to J��ri Aedla for reporting this issue. (CVE-2014-0195) 1420 [J��ri Aedla, Steve Henson] 1421 1422 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites 1423 are subject to a denial of service attack. 1424 1425 Thanks to Felix Gr��bert and Ivan Fratric at Google for discovering 1426 this issue. (CVE-2014-3470) 1427 [Felix Gr��bert, Ivan Fratric, Steve Henson] 1428 1429 *) Harmonize version and its documentation. -f flag is used to display 1430 compilation flags. 1431 [mancha <mancha1@zoho.com>] 1432 1433 *) Fix eckey_priv_encode so it immediately returns an error upon a failure 1434 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue. 1435 [mancha <mancha1@zoho.com>] 1436 1437 *) Fix some double frees. These are not thought to be exploitable. 1438 [mancha <mancha1@zoho.com>] 1439 1440 Changes between 1.0.1f and 1.0.1g [7 Apr 2014] 1441 1442 *) A missing bounds check in the handling of the TLS heartbeat extension 1443 can be used to reveal up to 64k of memory to a connected client or 1444 server. 1445 1446 Thanks for Neel Mehta of Google Security for discovering this bug and to 1447 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for 1448 preparing the fix (CVE-2014-0160) 1449 [Adam Langley, Bodo Moeller] 1450 1451 *) Fix for the attack described in the paper "Recovering OpenSSL 1452 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack" 1453 by Yuval Yarom and Naomi Benger. Details can be obtained from: 1454 http://eprint.iacr.org/2014/140 1455 1456 Thanks to Yuval Yarom and Naomi Benger for discovering this 1457 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076) 1458 [Yuval Yarom and Naomi Benger] 1459 1460 *) TLS pad extension: draft-agl-tls-padding-03 1461 1462 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the 1463 TLS client Hello record length value would otherwise be > 255 and 1464 less that 512 pad with a dummy extension containing zeroes so it 1465 is at least 512 bytes long. 1466 1467 [Adam Langley, Steve Henson] 1468 1469 Changes between 1.0.1e and 1.0.1f [6 Jan 2014] 1470 1471 *) Fix for TLS record tampering bug. A carefully crafted invalid 1472 handshake could crash OpenSSL with a NULL pointer exception. 1473 Thanks to Anton Johansson for reporting this issues. 1474 (CVE-2013-4353) 1475 1476 *) Keep original DTLS digest and encryption contexts in retransmission 1477 structures so we can use the previous session parameters if they need 1478 to be resent. (CVE-2013-6450) 1479 [Steve Henson] 1480 1481 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which 1482 avoids preferring ECDHE-ECDSA ciphers when the client appears to be 1483 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for 1484 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug 1485 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing 1486 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer. 1487 [Rob Stradling, Adam Langley] 1488 1489 Changes between 1.0.1d and 1.0.1e [11 Feb 2013] 1490 1491 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI 1492 supporting platforms or when small records were transferred. 1493 [Andy Polyakov, Steve Henson] 1494 1495 Changes between 1.0.1c and 1.0.1d [5 Feb 2013] 1496 1497 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 1498 1499 This addresses the flaw in CBC record processing discovered by 1500 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 1501 at: http://www.isg.rhul.ac.uk/tls/ 1502 1503 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 1504 Security Group at Royal Holloway, University of London 1505 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 1506 Emilia K��sper for the initial patch. 1507 (CVE-2013-0169) 1508 [Emilia K��sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 1509 1510 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode 1511 ciphersuites which can be exploited in a denial of service attack. 1512 Thanks go to and to Adam Langley <agl@chromium.org> for discovering 1513 and detecting this bug and to Wolfgang Ettlinger 1514 <wolfgang.ettlinger@gmail.com> for independently discovering this issue. 1515 (CVE-2012-2686) 1516 [Adam Langley] 1517 1518 *) Return an error when checking OCSP signatures when key is NULL. 1519 This fixes a DoS attack. (CVE-2013-0166) 1520 [Steve Henson] 1521 1522 *) Make openssl verify return errors. 1523 [Chris Palmer <palmer@google.com> and Ben Laurie] 1524 1525 *) Call OCSP Stapling callback after ciphersuite has been chosen, so 1526 the right response is stapled. Also change SSL_get_certificate() 1527 so it returns the certificate actually sent. 1528 See http://rt.openssl.org/Ticket/Display.html?id=2836. 1529 [Rob Stradling <rob.stradling@comodo.com>] 1530 1531 *) Fix possible deadlock when decoding public keys. 1532 [Steve Henson] 1533 1534 *) Don't use TLS 1.0 record version number in initial client hello 1535 if renegotiating. 1536 [Steve Henson] 1537 1538 Changes between 1.0.1b and 1.0.1c [10 May 2012] 1539 1540 *) Sanity check record length before skipping explicit IV in TLS 1541 1.2, 1.1 and DTLS to fix DoS attack. 1542 1543 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 1544 fuzzing as a service testing platform. 1545 (CVE-2012-2333) 1546 [Steve Henson] 1547 1548 *) Initialise tkeylen properly when encrypting CMS messages. 1549 Thanks to Solar Designer of Openwall for reporting this issue. 1550 [Steve Henson] 1551 1552 *) In FIPS mode don't try to use composite ciphers as they are not 1553 approved. 1554 [Steve Henson] 1555 1556 Changes between 1.0.1a and 1.0.1b [26 Apr 2012] 1557 1558 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and 1559 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately 1560 mean any application compiled against OpenSSL 1.0.0 headers setting 1561 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng 1562 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to 1563 0x10000000L Any application which was previously compiled against 1564 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1 1565 will need to be recompiled as a result. Letting be results in 1566 inability to disable specifically TLS 1.1 and in client context, 1567 in unlike event, limit maximum offered version to TLS 1.0 [see below]. 1568 [Steve Henson] 1569 1570 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not 1571 disable just protocol X, but all protocols above X *if* there are 1572 protocols *below* X still enabled. In more practical terms it means 1573 that if application wants to disable TLS1.0 in favor of TLS1.1 and 1574 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass 1575 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to 1576 client side. 1577 [Andy Polyakov] 1578 1579 Changes between 1.0.1 and 1.0.1a [19 Apr 2012] 1580 1581 *) Check for potentially exploitable overflows in asn1_d2i_read_bio 1582 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 1583 in CRYPTO_realloc_clean. 1584 1585 Thanks to Tavis Ormandy, Google Security Team, for discovering this 1586 issue and to Adam Langley <agl@chromium.org> for fixing it. 1587 (CVE-2012-2110) 1588 [Adam Langley (Google), Tavis Ormandy, Google Security Team] 1589 1590 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections. 1591 [Adam Langley] 1592 1593 *) Workarounds for some broken servers that "hang" if a client hello 1594 record length exceeds 255 bytes. 1595 1596 1. Do not use record version number > TLS 1.0 in initial client 1597 hello: some (but not all) hanging servers will now work. 1598 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate 1599 the number of ciphers sent in the client hello. This should be 1600 set to an even number, such as 50, for example by passing: 1601 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure. 1602 Most broken servers should now work. 1603 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable 1604 TLS 1.2 client support entirely. 1605 [Steve Henson] 1606 1607 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH. 1608 [Andy Polyakov] 1609 1610 Changes between 1.0.0h and 1.0.1 [14 Mar 2012] 1611 1612 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET 1613 STRING form instead of a DigestInfo. 1614 [Steve Henson] 1615 1616 *) The format used for MDC2 RSA signatures is inconsistent between EVP 1617 and the RSA_sign/RSA_verify functions. This was made more apparent when 1618 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular 1619 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect 1620 the correct format in RSA_verify so both forms transparently work. 1621 [Steve Henson] 1622 1623 *) Some servers which support TLS 1.0 can choke if we initially indicate 1624 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA 1625 encrypted premaster secret. As a workaround use the maximum pemitted 1626 client version in client hello, this should keep such servers happy 1627 and still work with previous versions of OpenSSL. 1628 [Steve Henson] 1629 1630 *) Add support for TLS/DTLS heartbeats. 1631 [Robin Seggelmann <seggelmann@fh-muenster.de>] 1632 1633 *) Add support for SCTP. 1634 [Robin Seggelmann <seggelmann@fh-muenster.de>] 1635 1636 *) Improved PRNG seeding for VOS. 1637 [Paul Green <Paul.Green@stratus.com>] 1638 1639 *) Extensive assembler packs updates, most notably: 1640 1641 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support; 1642 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES); 1643 - x86_64: bit-sliced AES implementation; 1644 - ARM: NEON support, contemporary platforms optimizations; 1645 - s390x: z196 support; 1646 - *: GHASH and GF(2^m) multiplication implementations; 1647 1648 [Andy Polyakov] 1649 1650 *) Make TLS-SRP code conformant with RFC 5054 API cleanup 1651 (removal of unnecessary code) 1652 [Peter Sylvester <peter.sylvester@edelweb.fr>] 1653 1654 *) Add TLS key material exporter from RFC 5705. 1655 [Eric Rescorla] 1656 1657 *) Add DTLS-SRTP negotiation from RFC 5764. 1658 [Eric Rescorla] 1659 1660 *) Add Next Protocol Negotiation, 1661 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be 1662 disabled with a no-npn flag to config or Configure. Code donated 1663 by Google. 1664 [Adam Langley <agl@google.com> and Ben Laurie] 1665 1666 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224, 1667 NIST-P256, NIST-P521, with constant-time single point multiplication on 1668 typical inputs. Compiler support for the nonstandard type __uint128_t is 1669 required to use this (present in gcc 4.4 and later, for 64-bit builds). 1670 Code made available under Apache License version 2.0. 1671 1672 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command 1673 line to include this in your build of OpenSSL, and run "make depend" (or 1674 "make update"). This enables the following EC_METHODs: 1675 1676 EC_GFp_nistp224_method() 1677 EC_GFp_nistp256_method() 1678 EC_GFp_nistp521_method() 1679 1680 EC_GROUP_new_by_curve_name() will automatically use these (while 1681 EC_GROUP_new_curve_GFp() currently prefers the more flexible 1682 implementations). 1683 [Emilia K��sper, Adam Langley, Bodo Moeller (Google)] 1684 1685 *) Use type ossl_ssize_t instad of ssize_t which isn't available on 1686 all platforms. Move ssize_t definition from e_os.h to the public 1687 header file e_os2.h as it now appears in public header file cms.h 1688 [Steve Henson] 1689 1690 *) New -sigopt option to the ca, req and x509 utilities. Additional 1691 signature parameters can be passed using this option and in 1692 particular PSS. 1693 [Steve Henson] 1694 1695 *) Add RSA PSS signing function. This will generate and set the 1696 appropriate AlgorithmIdentifiers for PSS based on those in the 1697 corresponding EVP_MD_CTX structure. No application support yet. 1698 [Steve Henson] 1699 1700 *) Support for companion algorithm specific ASN1 signing routines. 1701 New function ASN1_item_sign_ctx() signs a pre-initialised 1702 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on 1703 the appropriate parameters. 1704 [Steve Henson] 1705 1706 *) Add new algorithm specific ASN1 verification initialisation function 1707 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1 1708 handling will be the same no matter what EVP_PKEY_METHOD is used. 1709 Add a PSS handler to support verification of PSS signatures: checked 1710 against a number of sample certificates. 1711 [Steve Henson] 1712 1713 *) Add signature printing for PSS. Add PSS OIDs. 1714 [Steve Henson, Martin Kaiser <lists@kaiser.cx>] 1715 1716 *) Add algorithm specific signature printing. An individual ASN1 method 1717 can now print out signatures instead of the standard hex dump. 1718 1719 More complex signatures (e.g. PSS) can print out more meaningful 1720 information. Include DSA version that prints out the signature 1721 parameters r, s. 1722 [Steve Henson] 1723 1724 *) Password based recipient info support for CMS library: implementing 1725 RFC3211. 1726 [Steve Henson] 1727 1728 *) Split password based encryption into PBES2 and PBKDF2 functions. This 1729 neatly separates the code into cipher and PBE sections and is required 1730 for some algorithms that split PBES2 into separate pieces (such as 1731 password based CMS). 1732 [Steve Henson] 1733 1734 *) Session-handling fixes: 1735 - Fix handling of connections that are resuming with a session ID, 1736 but also support Session Tickets. 1737 - Fix a bug that suppressed issuing of a new ticket if the client 1738 presented a ticket with an expired session. 1739 - Try to set the ticket lifetime hint to something reasonable. 1740 - Make tickets shorter by excluding irrelevant information. 1741 - On the client side, don't ignore renewed tickets. 1742 [Adam Langley, Bodo Moeller (Google)] 1743 1744 *) Fix PSK session representation. 1745 [Bodo Moeller] 1746 1747 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations. 1748 1749 This work was sponsored by Intel. 1750 [Andy Polyakov] 1751 1752 *) Add GCM support to TLS library. Some custom code is needed to split 1753 the IV between the fixed (from PRF) and explicit (from TLS record) 1754 portions. This adds all GCM ciphersuites supported by RFC5288 and 1755 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and 1756 add a special AESGCM string for GCM only. 1757 [Steve Henson] 1758 1759 *) Expand range of ctrls for AES GCM. Permit setting invocation 1760 field on decrypt and retrieval of invocation field only on encrypt. 1761 [Steve Henson] 1762 1763 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support. 1764 As required by RFC5289 these ciphersuites cannot be used if for 1765 versions of TLS earlier than 1.2. 1766 [Steve Henson] 1767 1768 *) For FIPS capable OpenSSL interpret a NULL default public key method 1769 as unset and return the appopriate default but do *not* set the default. 1770 This means we can return the appopriate method in applications that 1771 swicth between FIPS and non-FIPS modes. 1772 [Steve Henson] 1773 1774 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an 1775 ENGINE is used then we cannot handle that in the FIPS module so we 1776 keep original code iff non-FIPS operations are allowed. 1777 [Steve Henson] 1778 1779 *) Add -attime option to openssl utilities. 1780 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson] 1781 1782 *) Redirect DSA and DH operations to FIPS module in FIPS mode. 1783 [Steve Henson] 1784 1785 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use 1786 FIPS EC methods unconditionally for now. 1787 [Steve Henson] 1788 1789 *) New build option no-ec2m to disable characteristic 2 code. 1790 [Steve Henson] 1791 1792 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not 1793 all cases can be covered as some introduce binary incompatibilities. 1794 [Steve Henson] 1795 1796 *) Redirect RSA operations to FIPS module including keygen, 1797 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods. 1798 [Steve Henson] 1799 1800 *) Add similar low level API blocking to ciphers. 1801 [Steve Henson] 1802 1803 *) Low level digest APIs are not approved in FIPS mode: any attempt 1804 to use these will cause a fatal error. Applications that *really* want 1805 to use them can use the private_* version instead. 1806 [Steve Henson] 1807 1808 *) Redirect cipher operations to FIPS module for FIPS builds. 1809 [Steve Henson] 1810 1811 *) Redirect digest operations to FIPS module for FIPS builds. 1812 [Steve Henson] 1813 1814 *) Update build system to add "fips" flag which will link in fipscanister.o 1815 for static and shared library builds embedding a signature if needed. 1816 [Steve Henson] 1817 1818 *) Output TLS supported curves in preference order instead of numerical 1819 order. This is currently hardcoded for the highest order curves first. 1820 This should be configurable so applications can judge speed vs strength. 1821 [Steve Henson] 1822 1823 *) Add TLS v1.2 server support for client authentication. 1824 [Steve Henson] 1825 1826 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers 1827 and enable MD5. 1828 [Steve Henson] 1829 1830 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying 1831 FIPS modules versions. 1832 [Steve Henson] 1833 1834 *) Add TLS v1.2 client side support for client authentication. Keep cache 1835 of handshake records longer as we don't know the hash algorithm to use 1836 until after the certificate request message is received. 1837 [Steve Henson] 1838 1839 *) Initial TLS v1.2 client support. Add a default signature algorithms 1840 extension including all the algorithms we support. Parse new signature 1841 format in client key exchange. Relax some ECC signing restrictions for 1842 TLS v1.2 as indicated in RFC5246. 1843 [Steve Henson] 1844 1845 *) Add server support for TLS v1.2 signature algorithms extension. Switch 1846 to new signature format when needed using client digest preference. 1847 All server ciphersuites should now work correctly in TLS v1.2. No client 1848 support yet and no support for client certificates. 1849 [Steve Henson] 1850 1851 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch 1852 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based 1853 ciphersuites. At present only RSA key exchange ciphersuites work with 1854 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete 1855 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods 1856 and version checking. 1857 [Steve Henson] 1858 1859 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled 1860 with this defined it will not be affected by any changes to ssl internal 1861 structures. Add several utility functions to allow openssl application 1862 to work with OPENSSL_NO_SSL_INTERN defined. 1863 [Steve Henson] 1864 1865 *) Add SRP support. 1866 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie] 1867 1868 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id. 1869 [Steve Henson] 1870 1871 *) Permit abbreviated handshakes when renegotiating using the function 1872 SSL_renegotiate_abbreviated(). 1873 [Robin Seggelmann <seggelmann@fh-muenster.de>] 1874 1875 *) Add call to ENGINE_register_all_complete() to 1876 ENGINE_load_builtin_engines(), so some implementations get used 1877 automatically instead of needing explicit application support. 1878 [Steve Henson] 1879 1880 *) Add support for TLS key exporter as described in RFC5705. 1881 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson] 1882 1883 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only 1884 a few changes are required: 1885 1886 Add SSL_OP_NO_TLSv1_1 flag. 1887 Add TLSv1_1 methods. 1888 Update version checking logic to handle version 1.1. 1889 Add explicit IV handling (ported from DTLS code). 1890 Add command line options to s_client/s_server. 1891 [Steve Henson] 1892 1893 Changes between 1.0.0g and 1.0.0h [12 Mar 2012] 1894 1895 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 1896 in CMS and PKCS7 code. When RSA decryption fails use a random key for 1897 content decryption and always return the same error. Note: this attack 1898 needs on average 2^20 messages so it only affects automated senders. The 1899 old behaviour can be reenabled in the CMS code by setting the 1900 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 1901 an MMA defence is not necessary. 1902 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 1903 this issue. (CVE-2012-0884) 1904 [Steve Henson] 1905 1906 *) Fix CVE-2011-4619: make sure we really are receiving a 1907 client hello before rejecting multiple SGC restarts. Thanks to 1908 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 1909 [Steve Henson] 1910 1911 Changes between 1.0.0f and 1.0.0g [18 Jan 2012] 1912 1913 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 1914 Thanks to Antonio Martin, Enterprise Secure Access Research and 1915 Development, Cisco Systems, Inc. for discovering this bug and 1916 preparing a fix. (CVE-2012-0050) 1917 [Antonio Martin] 1918 1919 Changes between 1.0.0e and 1.0.0f [4 Jan 2012] 1920 1921 *) Nadhem Alfardan and Kenny Paterson have discovered an extension 1922 of the Vaudenay padding oracle attack on CBC mode encryption 1923 which enables an efficient plaintext recovery attack against 1924 the OpenSSL implementation of DTLS. Their attack exploits timing 1925 differences arising during decryption processing. A research 1926 paper describing this attack can be found at: 1927 http://www.isg.rhul.ac.uk/~kp/dtls.pdf 1928 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 1929 Security Group at Royal Holloway, University of London 1930 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 1931 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 1932 for preparing the fix. (CVE-2011-4108) 1933 [Robin Seggelmann, Michael Tuexen] 1934 1935 *) Clear bytes used for block padding of SSL 3.0 records. 1936 (CVE-2011-4576) 1937 [Adam Langley (Google)] 1938 1939 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 1940 Kadianakis <desnacked@gmail.com> for discovering this issue and 1941 Adam Langley for preparing the fix. (CVE-2011-4619) 1942 [Adam Langley (Google)] 1943 1944 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027) 1945 [Andrey Kulikov <amdeich@gmail.com>] 1946 1947 *) Prevent malformed RFC3779 data triggering an assertion failure. 1948 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 1949 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 1950 [Rob Austein <sra@hactrn.net>] 1951 1952 *) Improved PRNG seeding for VOS. 1953 [Paul Green <Paul.Green@stratus.com>] 1954 1955 *) Fix ssl_ciph.c set-up race. 1956 [Adam Langley (Google)] 1957 1958 *) Fix spurious failures in ecdsatest.c. 1959 [Emilia K��sper (Google)] 1960 1961 *) Fix the BIO_f_buffer() implementation (which was mixing different 1962 interpretations of the '..._len' fields). 1963 [Adam Langley (Google)] 1964 1965 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 1966 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 1967 threads won't reuse the same blinding coefficients. 1968 1969 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 1970 lock to call BN_BLINDING_invert_ex, and avoids one use of 1971 BN_BLINDING_update for each BN_BLINDING structure (previously, 1972 the last update always remained unused). 1973 [Emilia K��sper (Google)] 1974 1975 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf. 1976 [Bob Buckholz (Google)] 1977 1978 Changes between 1.0.0d and 1.0.0e [6 Sep 2011] 1979 1980 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted 1981 by initialising X509_STORE_CTX properly. (CVE-2011-3207) 1982 [Kaspar Brand <ossl@velox.ch>] 1983 1984 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 1985 for multi-threaded use of ECDH. (CVE-2011-3210) 1986 [Adam Langley (Google)] 1987 1988 *) Fix x509_name_ex_d2i memory leak on bad inputs. 1989 [Bodo Moeller] 1990 1991 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check 1992 signature public key algorithm by using OID xref utilities instead. 1993 Before this you could only use some ECC ciphersuites with SHA1 only. 1994 [Steve Henson] 1995 1996 *) Add protection against ECDSA timing attacks as mentioned in the paper 1997 by Billy Bob Brumley and Nicola Tuveri, see: 1998 1999 http://eprint.iacr.org/2011/232.pdf 2000 2001 [Billy Bob Brumley and Nicola Tuveri] 2002 2003 Changes between 1.0.0c and 1.0.0d [8 Feb 2011] 2004 2005 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 2006 [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 2007 2008 *) Fix bug in string printing code: if *any* escaping is enabled we must 2009 escape the escape character (backslash) or the resulting string is 2010 ambiguous. 2011 [Steve Henson] 2012 2013 Changes between 1.0.0b and 1.0.0c [2 Dec 2010] 2014 2015 *) Disable code workaround for ancient and obsolete Netscape browsers 2016 and servers: an attacker can use it in a ciphersuite downgrade attack. 2017 Thanks to Martin Rex for discovering this bug. CVE-2010-4180 2018 [Steve Henson] 2019 2020 *) Fixed J-PAKE implementation error, originally discovered by 2021 Sebastien Martini, further info and confirmation from Stefan 2022 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 2023 [Ben Laurie] 2024 2025 Changes between 1.0.0a and 1.0.0b [16 Nov 2010] 2026 2027 *) Fix extension code to avoid race conditions which can result in a buffer 2028 overrun vulnerability: resumed sessions must not be modified as they can 2029 be shared by multiple threads. CVE-2010-3864 2030 [Steve Henson] 2031 2032 *) Fix WIN32 build system to correctly link an ENGINE directory into 2033 a DLL. 2034 [Steve Henson] 2035 2036 Changes between 1.0.0 and 1.0.0a [01 Jun 2010] 2037 2038 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover 2039 (CVE-2010-1633) 2040 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>] 2041 2042 Changes between 0.9.8n and 1.0.0 [29 Mar 2010] 2043 2044 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher 2045 context. The operation can be customised via the ctrl mechanism in 2046 case ENGINEs want to include additional functionality. 2047 [Steve Henson] 2048 2049 *) Tolerate yet another broken PKCS#8 key format: private key value negative. 2050 [Steve Henson] 2051 2052 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to 2053 output hashes compatible with older versions of OpenSSL. 2054 [Willy Weisz <weisz@vcpc.univie.ac.at>] 2055 2056 *) Fix compression algorithm handling: if resuming a session use the 2057 compression algorithm of the resumed session instead of determining 2058 it from client hello again. Don't allow server to change algorithm. 2059 [Steve Henson] 2060 2061 *) Add load_crls() function to apps tidying load_certs() too. Add option 2062 to verify utility to allow additional CRLs to be included. 2063 [Steve Henson] 2064 2065 *) Update OCSP request code to permit adding custom headers to the request: 2066 some responders need this. 2067 [Steve Henson] 2068 2069 *) The function EVP_PKEY_sign() returns <=0 on error: check return code 2070 correctly. 2071 [Julia Lawall <julia@diku.dk>] 2072 2073 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it 2074 needlessly dereferenced structures, used obsolete functions and 2075 didn't handle all updated verify codes correctly. 2076 [Steve Henson] 2077 2078 *) Disable MD2 in the default configuration. 2079 [Steve Henson] 2080 2081 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to 2082 indicate the initial BIO being pushed or popped. This makes it possible 2083 to determine whether the BIO is the one explicitly called or as a result 2084 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so 2085 it handles reference counts correctly and doesn't zero out the I/O bio 2086 when it is not being explicitly popped. WARNING: applications which 2087 included workarounds for the old buggy behaviour will need to be modified 2088 or they could free up already freed BIOs. 2089 [Steve Henson] 2090 2091 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni 2092 renaming to all platforms (within the 0.9.8 branch, this was 2093 done conditionally on Netware platforms to avoid a name clash). 2094 [Guenter <lists@gknw.net>] 2095 2096 *) Add ECDHE and PSK support to DTLS. 2097 [Michael Tuexen <tuexen@fh-muenster.de>] 2098 2099 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't 2100 be used on C++. 2101 [Steve Henson] 2102 2103 *) Add "missing" function EVP_MD_flags() (without this the only way to 2104 retrieve a digest flags is by accessing the structure directly. Update 2105 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest 2106 or cipher is registered as in the "from" argument. Print out all 2107 registered digests in the dgst usage message instead of manually 2108 attempting to work them out. 2109 [Steve Henson] 2110 2111 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello: 2112 this allows the use of compression and extensions. Change default cipher 2113 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2 2114 by default unless an application cipher string requests it. 2115 [Steve Henson] 2116 2117 *) Alter match criteria in PKCS12_parse(). It used to try to use local 2118 key ids to find matching certificates and keys but some PKCS#12 files 2119 don't follow the (somewhat unwritten) rules and this strategy fails. 2120 Now just gather all certificates together and the first private key 2121 then look for the first certificate that matches the key. 2122 [Steve Henson] 2123 2124 *) Support use of registered digest and cipher names for dgst and cipher 2125 commands instead of having to add each one as a special case. So now 2126 you can do: 2127 2128 openssl sha256 foo 2129 2130 as well as: 2131 2132 openssl dgst -sha256 foo 2133 2134 and this works for ENGINE based algorithms too. 2135 2136 [Steve Henson] 2137 2138 *) Update Gost ENGINE to support parameter files. 2139 [Victor B. Wagner <vitus@cryptocom.ru>] 2140 2141 *) Support GeneralizedTime in ca utility. 2142 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson] 2143 2144 *) Enhance the hash format used for certificate directory links. The new 2145 form uses the canonical encoding (meaning equivalent names will work 2146 even if they aren't identical) and uses SHA1 instead of MD5. This form 2147 is incompatible with the older format and as a result c_rehash should 2148 be used to rebuild symbolic links. 2149 [Steve Henson] 2150 2151 *) Make PKCS#8 the default write format for private keys, replacing the 2152 traditional format. This form is standardised, more secure and doesn't 2153 include an implicit MD5 dependency. 2154 [Steve Henson] 2155 2156 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code 2157 committed to OpenSSL should pass this lot as a minimum. 2158 [Steve Henson] 2159 2160 *) Add session ticket override functionality for use by EAP-FAST. 2161 [Jouni Malinen <j@w1.fi>] 2162 2163 *) Modify HMAC functions to return a value. Since these can be implemented 2164 in an ENGINE errors can occur. 2165 [Steve Henson] 2166 2167 *) Type-checked OBJ_bsearch_ex. 2168 [Ben Laurie] 2169 2170 *) Type-checked OBJ_bsearch. Also some constification necessitated 2171 by type-checking. Still to come: TXT_DB, bsearch(?), 2172 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING, 2173 CONF_VALUE. 2174 [Ben Laurie] 2175 2176 *) New function OPENSSL_gmtime_adj() to add a specific number of days and 2177 seconds to a tm structure directly, instead of going through OS 2178 specific date routines. This avoids any issues with OS routines such 2179 as the year 2038 bug. New *_adj() functions for ASN1 time structures 2180 and X509_time_adj_ex() to cover the extended range. The existing 2181 X509_time_adj() is still usable and will no longer have any date issues. 2182 [Steve Henson] 2183 2184 *) Delta CRL support. New use deltas option which will attempt to locate 2185 and search any appropriate delta CRLs available. 2186 2187 This work was sponsored by Google. 2188 [Steve Henson] 2189 2190 *) Support for CRLs partitioned by reason code. Reorganise CRL processing 2191 code and add additional score elements. Validate alternate CRL paths 2192 as part of the CRL checking and indicate a new error "CRL path validation 2193 error" in this case. Applications wanting additional details can use 2194 the verify callback and check the new "parent" field. If this is not 2195 NULL CRL path validation is taking place. Existing applications wont 2196 see this because it requires extended CRL support which is off by 2197 default. 2198 2199 This work was sponsored by Google. 2200 [Steve Henson] 2201 2202 *) Support for freshest CRL extension. 2203 2204 This work was sponsored by Google. 2205 [Steve Henson] 2206 2207 *) Initial indirect CRL support. Currently only supported in the CRLs 2208 passed directly and not via lookup. Process certificate issuer 2209 CRL entry extension and lookup CRL entries by bother issuer name 2210 and serial number. Check and process CRL issuer entry in IDP extension. 2211 2212 This work was sponsored by Google. 2213 [Steve Henson] 2214 2215 *) Add support for distinct certificate and CRL paths. The CRL issuer 2216 certificate is validated separately in this case. Only enabled if 2217 an extended CRL support flag is set: this flag will enable additional 2218 CRL functionality in future. 2219 2220 This work was sponsored by Google. 2221 [Steve Henson] 2222 2223 *) Add support for policy mappings extension. 2224 2225 This work was sponsored by Google. 2226 [Steve Henson] 2227 2228 *) Fixes to pathlength constraint, self issued certificate handling, 2229 policy processing to align with RFC3280 and PKITS tests. 2230 2231 This work was sponsored by Google. 2232 [Steve Henson] 2233 2234 *) Support for name constraints certificate extension. DN, email, DNS 2235 and URI types are currently supported. 2236 2237 This work was sponsored by Google. 2238 [Steve Henson] 2239 2240 *) To cater for systems that provide a pointer-based thread ID rather 2241 than numeric, deprecate the current numeric thread ID mechanism and 2242 replace it with a structure and associated callback type. This 2243 mechanism allows a numeric "hash" to be extracted from a thread ID in 2244 either case, and on platforms where pointers are larger than 'long', 2245 mixing is done to help ensure the numeric 'hash' is usable even if it 2246 can't be guaranteed unique. The default mechanism is to use "&errno" 2247 as a pointer-based thread ID to distinguish between threads. 2248 2249 Applications that want to provide their own thread IDs should now use 2250 CRYPTO_THREADID_set_callback() to register a callback that will call 2251 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer(). 2252 2253 Note that ERR_remove_state() is now deprecated, because it is tied 2254 to the assumption that thread IDs are numeric. ERR_remove_state(0) 2255 to free the current thread's error state should be replaced by 2256 ERR_remove_thread_state(NULL). 2257 2258 (This new approach replaces the functions CRYPTO_set_idptr_callback(), 2259 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in 2260 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an 2261 application was previously providing a numeric thread callback that 2262 was inappropriate for distinguishing threads, then uniqueness might 2263 have been obtained with &errno that happened immediately in the 2264 intermediate development versions of OpenSSL; this is no longer the 2265 case, the numeric thread callback will now override the automatic use 2266 of &errno.) 2267 [Geoff Thorpe, with help from Bodo Moeller] 2268 2269 *) Initial support for different CRL issuing certificates. This covers a 2270 simple case where the self issued certificates in the chain exist and 2271 the real CRL issuer is higher in the existing chain. 2272 2273 This work was sponsored by Google. 2274 [Steve Henson] 2275 2276 *) Removed effectively defunct crypto/store from the build. 2277 [Ben Laurie] 2278 2279 *) Revamp of STACK to provide stronger type-checking. Still to come: 2280 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE, 2281 ASN1_STRING, CONF_VALUE. 2282 [Ben Laurie] 2283 2284 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer 2285 RAM on SSL connections. This option can save about 34k per idle SSL. 2286 [Nick Mathewson] 2287 2288 *) Revamp of LHASH to provide stronger type-checking. Still to come: 2289 STACK, TXT_DB, bsearch, qsort. 2290 [Ben Laurie] 2291 2292 *) Initial support for Cryptographic Message Syntax (aka CMS) based 2293 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility, 2294 support for data, signedData, compressedData, digestedData and 2295 encryptedData, envelopedData types included. Scripts to check against 2296 RFC4134 examples draft and interop and consistency checks of many 2297 content types and variants. 2298 [Steve Henson] 2299 2300 *) Add options to enc utility to support use of zlib compression BIO. 2301 [Steve Henson] 2302 2303 *) Extend mk1mf to support importing of options and assembly language 2304 files from Configure script, currently only included in VC-WIN32. 2305 The assembly language rules can now optionally generate the source 2306 files from the associated perl scripts. 2307 [Steve Henson] 2308 2309 *) Implement remaining functionality needed to support GOST ciphersuites. 2310 Interop testing has been performed using CryptoPro implementations. 2311 [Victor B. Wagner <vitus@cryptocom.ru>] 2312 2313 *) s390x assembler pack. 2314 [Andy Polyakov] 2315 2316 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU 2317 "family." 2318 [Andy Polyakov] 2319 2320 *) Implement Opaque PRF Input TLS extension as specified in 2321 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an 2322 official specification yet and no extension type assignment by 2323 IANA exists, this extension (for now) will have to be explicitly 2324 enabled when building OpenSSL by providing the extension number 2325 to use. For example, specify an option 2326 2327 -DTLSEXT_TYPE_opaque_prf_input=0x9527 2328 2329 to the "config" or "Configure" script to enable the extension, 2330 assuming extension number 0x9527 (which is a completely arbitrary 2331 and unofficial assignment based on the MD5 hash of the Internet 2332 Draft). Note that by doing so, you potentially lose 2333 interoperability with other TLS implementations since these might 2334 be using the same extension number for other purposes. 2335 2336 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the 2337 opaque PRF input value to use in the handshake. This will create 2338 an interal copy of the length-'len' string at 'src', and will 2339 return non-zero for success. 2340 2341 To get more control and flexibility, provide a callback function 2342 by using 2343 2344 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb) 2345 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg) 2346 2347 where 2348 2349 int (*cb)(SSL *, void *peerinput, size_t len, void *arg); 2350 void *arg; 2351 2352 Callback function 'cb' will be called in handshakes, and is 2353 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate. 2354 Argument 'arg' is for application purposes (the value as given to 2355 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly 2356 be provided to the callback function). The callback function 2357 has to return non-zero to report success: usually 1 to use opaque 2358 PRF input just if possible, or 2 to enforce use of the opaque PRF 2359 input. In the latter case, the library will abort the handshake 2360 if opaque PRF input is not successfully negotiated. 2361 2362 Arguments 'peerinput' and 'len' given to the callback function 2363 will always be NULL and 0 in the case of a client. A server will 2364 see the client's opaque PRF input through these variables if 2365 available (NULL and 0 otherwise). Note that if the server 2366 provides an opaque PRF input, the length must be the same as the 2367 length of the client's opaque PRF input. 2368 2369 Note that the callback function will only be called when creating 2370 a new session (session resumption can resume whatever was 2371 previously negotiated), and will not be called in SSL 2.0 2372 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or 2373 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended 2374 for applications that need to enforce opaque PRF input. 2375 2376 [Bodo Moeller] 2377 2378 *) Update ssl code to support digests other than SHA1+MD5 for handshake 2379 MAC. 2380 2381 [Victor B. Wagner <vitus@cryptocom.ru>] 2382 2383 *) Add RFC4507 support to OpenSSL. This includes the corrections in 2384 RFC4507bis. The encrypted ticket format is an encrypted encoded 2385 SSL_SESSION structure, that way new session features are automatically 2386 supported. 2387 2388 If a client application caches session in an SSL_SESSION structure 2389 support is transparent because tickets are now stored in the encoded 2390 SSL_SESSION. 2391 2392 The SSL_CTX structure automatically generates keys for ticket 2393 protection in servers so again support should be possible 2394 with no application modification. 2395 2396 If a client or server wishes to disable RFC4507 support then the option 2397 SSL_OP_NO_TICKET can be set. 2398 2399 Add a TLS extension debugging callback to allow the contents of any client 2400 or server extensions to be examined. 2401 2402 This work was sponsored by Google. 2403 [Steve Henson] 2404 2405 *) Final changes to avoid use of pointer pointer casts in OpenSSL. 2406 OpenSSL should now compile cleanly on gcc 4.2 2407 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson] 2408 2409 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC 2410 support including streaming MAC support: this is required for GOST 2411 ciphersuite support. 2412 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson] 2413 2414 *) Add option -stream to use PKCS#7 streaming in smime utility. New 2415 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream() 2416 to output in BER and PEM format. 2417 [Steve Henson] 2418 2419 *) Experimental support for use of HMAC via EVP_PKEY interface. This 2420 allows HMAC to be handled via the EVP_DigestSign*() interface. The 2421 EVP_PKEY "key" in this case is the HMAC key, potentially allowing 2422 ENGINE support for HMAC keys which are unextractable. New -mac and 2423 -macopt options to dgst utility. 2424 [Steve Henson] 2425 2426 *) New option -sigopt to dgst utility. Update dgst to use 2427 EVP_Digest{Sign,Verify}*. These two changes make it possible to use 2428 alternative signing paramaters such as X9.31 or PSS in the dgst 2429 utility. 2430 [Steve Henson] 2431 2432 *) Change ssl_cipher_apply_rule(), the internal function that does 2433 the work each time a ciphersuite string requests enabling 2434 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or 2435 removing ("!foo+bar") a class of ciphersuites: Now it maintains 2436 the order of disabled ciphersuites such that those ciphersuites 2437 that most recently went from enabled to disabled not only stay 2438 in order with respect to each other, but also have higher priority 2439 than other disabled ciphersuites the next time ciphersuites are 2440 enabled again. 2441 2442 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable 2443 the same ciphersuites as with "HIGH" alone, but in a specific 2444 order where the PSK ciphersuites come first (since they are the 2445 most recently disabled ciphersuites when "HIGH" is parsed). 2446 2447 Also, change ssl_create_cipher_list() (using this new 2448 funcionality) such that between otherwise identical 2449 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in 2450 the default order. 2451 [Bodo Moeller] 2452 2453 *) Change ssl_create_cipher_list() so that it automatically 2454 arranges the ciphersuites in reasonable order before starting 2455 to process the rule string. Thus, the definition for "DEFAULT" 2456 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but 2457 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH". 2458 This makes it much easier to arrive at a reasonable default order 2459 in applications for which anonymous ciphers are OK (meaning 2460 that you can't actually use DEFAULT). 2461 [Bodo Moeller; suggested by Victor Duchovni] 2462 2463 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string 2464 processing) into multiple integers instead of setting 2465 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK", 2466 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer. 2467 (These masks as well as the individual bit definitions are hidden 2468 away into the non-exported interface ssl/ssl_locl.h, so this 2469 change to the definition of the SSL_CIPHER structure shouldn't 2470 affect applications.) This give us more bits for each of these 2471 categories, so there is no longer a need to coagulate AES128 and 2472 AES256 into a single algorithm bit, and to coagulate Camellia128 2473 and Camellia256 into a single algorithm bit, which has led to all 2474 kinds of kludges. 2475 2476 Thus, among other things, the kludge introduced in 0.9.7m and 2477 0.9.8e for masking out AES256 independently of AES128 or masking 2478 out Camellia256 independently of AES256 is not needed here in 0.9.9. 2479 2480 With the change, we also introduce new ciphersuite aliases that 2481 so far were missing: "AES128", "AES256", "CAMELLIA128", and 2482 "CAMELLIA256". 2483 [Bodo Moeller] 2484 2485 *) Add support for dsa-with-SHA224 and dsa-with-SHA256. 2486 Use the leftmost N bytes of the signature input if the input is 2487 larger than the prime q (with N being the size in bytes of q). 2488 [Nils Larsch] 2489 2490 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses 2491 it yet and it is largely untested. 2492 [Steve Henson] 2493 2494 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types. 2495 [Nils Larsch] 2496 2497 *) Initial incomplete changes to avoid need for function casts in OpenSSL 2498 some compilers (gcc 4.2 and later) reject their use. Safestack is 2499 reimplemented. Update ASN1 to avoid use of legacy functions. 2500 [Steve Henson] 2501 2502 *) Win32/64 targets are linked with Winsock2. 2503 [Andy Polyakov] 2504 2505 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected 2506 to external functions. This can be used to increase CRL handling 2507 efficiency especially when CRLs are very large by (for example) storing 2508 the CRL revoked certificates in a database. 2509 [Steve Henson] 2510 2511 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so 2512 new CRLs added to a directory can be used. New command line option 2513 -verify_return_error to s_client and s_server. This causes real errors 2514 to be returned by the verify callback instead of carrying on no matter 2515 what. This reflects the way a "real world" verify callback would behave. 2516 [Steve Henson] 2517 2518 *) GOST engine, supporting several GOST algorithms and public key formats. 2519 Kindly donated by Cryptocom. 2520 [Cryptocom] 2521 2522 *) Partial support for Issuing Distribution Point CRL extension. CRLs 2523 partitioned by DP are handled but no indirect CRL or reason partitioning 2524 (yet). Complete overhaul of CRL handling: now the most suitable CRL is 2525 selected via a scoring technique which handles IDP and AKID in CRLs. 2526 [Steve Henson] 2527 2528 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which 2529 will ultimately be used for all verify operations: this will remove the 2530 X509_STORE dependency on certificate verification and allow alternative 2531 lookup methods. X509_STORE based implementations of these two callbacks. 2532 [Steve Henson] 2533 2534 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names. 2535 Modify get_crl() to find a valid (unexpired) CRL if possible. 2536 [Steve Henson] 2537 2538 *) New function X509_CRL_match() to check if two CRLs are identical. Normally 2539 this would be called X509_CRL_cmp() but that name is already used by 2540 a function that just compares CRL issuer names. Cache several CRL 2541 extensions in X509_CRL structure and cache CRLDP in X509. 2542 [Steve Henson] 2543 2544 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name) 2545 this maps equivalent X509_NAME structures into a consistent structure. 2546 Name comparison can then be performed rapidly using memcmp(). 2547 [Steve Henson] 2548 2549 *) Non-blocking OCSP request processing. Add -timeout option to ocsp 2550 utility. 2551 [Steve Henson] 2552 2553 *) Allow digests to supply their own micalg string for S/MIME type using 2554 the ctrl EVP_MD_CTRL_MICALG. 2555 [Steve Henson] 2556 2557 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the 2558 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN 2559 ctrl. It can then customise the structure before and/or after signing 2560 if necessary. 2561 [Steve Henson] 2562 2563 *) New function OBJ_add_sigid() to allow application defined signature OIDs 2564 to be added to OpenSSLs internal tables. New function OBJ_sigid_free() 2565 to free up any added signature OIDs. 2566 [Steve Henson] 2567 2568 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(), 2569 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal 2570 digest and cipher tables. New options added to openssl utility: 2571 list-message-digest-algorithms and list-cipher-algorithms. 2572 [Steve Henson] 2573 2574 *) Change the array representation of binary polynomials: the list 2575 of degrees of non-zero coefficients is now terminated with -1. 2576 Previously it was terminated with 0, which was also part of the 2577 value; thus, the array representation was not applicable to 2578 polynomials where t^0 has coefficient zero. This change makes 2579 the array representation useful in a more general context. 2580 [Douglas Stebila] 2581 2582 *) Various modifications and fixes to SSL/TLS cipher string 2583 handling. For ECC, the code now distinguishes between fixed ECDH 2584 with RSA certificates on the one hand and with ECDSA certificates 2585 on the other hand, since these are separate ciphersuites. The 2586 unused code for Fortezza ciphersuites has been removed. 2587 2588 For consistency with EDH, ephemeral ECDH is now called "EECDH" 2589 (not "ECDHE"). For consistency with the code for DH 2590 certificates, use of ECDH certificates is now considered ECDH 2591 authentication, not RSA or ECDSA authentication (the latter is 2592 merely the CA's signing algorithm and not actively used in the 2593 protocol). 2594 2595 The temporary ciphersuite alias "ECCdraft" is no longer 2596 available, and ECC ciphersuites are no longer excluded from "ALL" 2597 and "DEFAULT". The following aliases now exist for RFC 4492 2598 ciphersuites, most of these by analogy with the DH case: 2599 2600 kECDHr - ECDH cert, signed with RSA 2601 kECDHe - ECDH cert, signed with ECDSA 2602 kECDH - ECDH cert (signed with either RSA or ECDSA) 2603 kEECDH - ephemeral ECDH 2604 ECDH - ECDH cert or ephemeral ECDH 2605 2606 aECDH - ECDH cert 2607 aECDSA - ECDSA cert 2608 ECDSA - ECDSA cert 2609 2610 AECDH - anonymous ECDH 2611 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH") 2612 2613 [Bodo Moeller] 2614 2615 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported. 2616 Use correct micalg parameters depending on digest(s) in signed message. 2617 [Steve Henson] 2618 2619 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process 2620 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code. 2621 [Steve Henson] 2622 2623 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit 2624 an engine to register a method. Add ENGINE lookups for methods and 2625 functional reference processing. 2626 [Steve Henson] 2627 2628 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of 2629 EVP_{Sign,Verify}* which allow an application to customise the signature 2630 process. 2631 [Steve Henson] 2632 2633 *) New -resign option to smime utility. This adds one or more signers 2634 to an existing PKCS#7 signedData structure. Also -md option to use an 2635 alternative message digest algorithm for signing. 2636 [Steve Henson] 2637 2638 *) Tidy up PKCS#7 routines and add new functions to make it easier to 2639 create PKCS7 structures containing multiple signers. Update smime 2640 application to support multiple signers. 2641 [Steve Henson] 2642 2643 *) New -macalg option to pkcs12 utility to allow setting of an alternative 2644 digest MAC. 2645 [Steve Henson] 2646 2647 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC. 2648 Reorganize PBE internals to lookup from a static table using NIDs, 2649 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl: 2650 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative 2651 PRF which will be automatically used with PBES2. 2652 [Steve Henson] 2653 2654 *) Replace the algorithm specific calls to generate keys in "req" with the 2655 new API. 2656 [Steve Henson] 2657 2658 *) Update PKCS#7 enveloped data routines to use new API. This is now 2659 supported by any public key method supporting the encrypt operation. A 2660 ctrl is added to allow the public key algorithm to examine or modify 2661 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is 2662 a no op. 2663 [Steve Henson] 2664 2665 *) Add a ctrl to asn1 method to allow a public key algorithm to express 2666 a default digest type to use. In most cases this will be SHA1 but some 2667 algorithms (such as GOST) need to specify an alternative digest. The 2668 return value indicates how strong the prefernce is 1 means optional and 2669 2 is mandatory (that is it is the only supported type). Modify 2670 ASN1_item_sign() to accept a NULL digest argument to indicate it should 2671 use the default md. Update openssl utilities to use the default digest 2672 type for signing if it is not explicitly indicated. 2673 [Steve Henson] 2674 2675 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New 2676 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant 2677 signing method from the key type. This effectively removes the link 2678 between digests and public key types. 2679 [Steve Henson] 2680 2681 *) Add an OID cross reference table and utility functions. Its purpose is to 2682 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1, 2683 rsaEncryption. This will allow some of the algorithm specific hackery 2684 needed to use the correct OID to be removed. 2685 [Steve Henson] 2686 2687 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO 2688 structures for PKCS7_sign(). They are now set up by the relevant public 2689 key ASN1 method. 2690 [Steve Henson] 2691 2692 *) Add provisional EC pkey method with support for ECDSA and ECDH. 2693 [Steve Henson] 2694 2695 *) Add support for key derivation (agreement) in the API, DH method and 2696 pkeyutl. 2697 [Steve Henson] 2698 2699 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support 2700 public and private key formats. As a side effect these add additional 2701 command line functionality not previously available: DSA signatures can be 2702 generated and verified using pkeyutl and DH key support and generation in 2703 pkey, genpkey. 2704 [Steve Henson] 2705 2706 *) BeOS support. 2707 [Oliver Tappe <zooey@hirschkaefer.de>] 2708 2709 *) New make target "install_html_docs" installs HTML renditions of the 2710 manual pages. 2711 [Oliver Tappe <zooey@hirschkaefer.de>] 2712 2713 *) New utility "genpkey" this is analagous to "genrsa" etc except it can 2714 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to 2715 support key and parameter generation and add initial key generation 2716 functionality for RSA. 2717 [Steve Henson] 2718 2719 *) Add functions for main EVP_PKEY_method operations. The undocumented 2720 functions EVP_PKEY_{encrypt,decrypt} have been renamed to 2721 EVP_PKEY_{encrypt,decrypt}_old. 2722 [Steve Henson] 2723 2724 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public 2725 key API, doesn't do much yet. 2726 [Steve Henson] 2727 2728 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about 2729 public key algorithms. New option to openssl utility: 2730 "list-public-key-algorithms" to print out info. 2731 [Steve Henson] 2732 2733 *) Implement the Supported Elliptic Curves Extension for 2734 ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 2735 [Douglas Stebila] 2736 2737 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or 2738 EVP_CIPHER structures to avoid later problems in EVP_cleanup(). 2739 [Steve Henson] 2740 2741 *) New utilities pkey and pkeyparam. These are similar to algorithm specific 2742 utilities such as rsa, dsa, dsaparam etc except they process any key 2743 type. 2744 [Steve Henson] 2745 2746 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New 2747 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(), 2748 EVP_PKEY_print_param() to print public key data from an EVP_PKEY 2749 structure. 2750 [Steve Henson] 2751 2752 *) Initial support for pluggable public key ASN1. 2753 De-spaghettify the public key ASN1 handling. Move public and private 2754 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate 2755 algorithm specific handling to a single module within the relevant 2756 algorithm directory. Add functions to allow (near) opaque processing 2757 of public and private key structures. 2758 [Steve Henson] 2759 2760 *) Implement the Supported Point Formats Extension for 2761 ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 2762 [Douglas Stebila] 2763 2764 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members 2765 for the psk identity [hint] and the psk callback functions to the 2766 SSL_SESSION, SSL and SSL_CTX structure. 2767 2768 New ciphersuites: 2769 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA, 2770 PSK-AES256-CBC-SHA 2771 2772 New functions: 2773 SSL_CTX_use_psk_identity_hint 2774 SSL_get_psk_identity_hint 2775 SSL_get_psk_identity 2776 SSL_use_psk_identity_hint 2777 2778 [Mika Kousa and Pasi Eronen of Nokia Corporation] 2779 2780 *) Add RFC 3161 compliant time stamp request creation, response generation 2781 and response verification functionality. 2782 [Zolt��n Gl��zik <zglozik@opentsa.org>, The OpenTSA Project] 2783 2784 *) Add initial support for TLS extensions, specifically for the server_name 2785 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 2786 have new members for a host name. The SSL data structure has an 2787 additional member SSL_CTX *initial_ctx so that new sessions can be 2788 stored in that context to allow for session resumption, even after the 2789 SSL has been switched to a new SSL_CTX in reaction to a client's 2790 server_name extension. 2791 2792 New functions (subject to change): 2793 2794 SSL_get_servername() 2795 SSL_get_servername_type() 2796 SSL_set_SSL_CTX() 2797 2798 New CTRL codes and macros (subject to change): 2799 2800 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 2801 - SSL_CTX_set_tlsext_servername_callback() 2802 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 2803 - SSL_CTX_set_tlsext_servername_arg() 2804 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 2805 2806 openssl s_client has a new '-servername ...' option. 2807 2808 openssl s_server has new options '-servername_host ...', '-cert2 ...', 2809 '-key2 ...', '-servername_fatal' (subject to change). This allows 2810 testing the HostName extension for a specific single host name ('-cert' 2811 and '-key' remain fallbacks for handshakes without HostName 2812 negotiation). If the unrecogninzed_name alert has to be sent, this by 2813 default is a warning; it becomes fatal with the '-servername_fatal' 2814 option. 2815 2816 [Peter Sylvester, Remy Allais, Christophe Renou] 2817 2818 *) Whirlpool hash implementation is added. 2819 [Andy Polyakov] 2820 2821 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to 2822 bn(64,32). Because of instruction set limitations it doesn't have 2823 any negative impact on performance. This was done mostly in order 2824 to make it possible to share assembler modules, such as bn_mul_mont 2825 implementations, between 32- and 64-bit builds without hassle. 2826 [Andy Polyakov] 2827 2828 *) Move code previously exiled into file crypto/ec/ec2_smpt.c 2829 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP 2830 macro. 2831 [Bodo Moeller] 2832 2833 *) New candidate for BIGNUM assembler implementation, bn_mul_mont, 2834 dedicated Montgomery multiplication procedure, is introduced. 2835 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher 2836 "64-bit" performance on certain 32-bit targets. 2837 [Andy Polyakov] 2838 2839 *) New option SSL_OP_NO_COMP to disable use of compression selectively 2840 in SSL structures. New SSL ctrl to set maximum send fragment size. 2841 Save memory by seeting the I/O buffer sizes dynamically instead of 2842 using the maximum available value. 2843 [Steve Henson] 2844 2845 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code 2846 in addition to the text details. 2847 [Bodo Moeller] 2848 2849 *) Very, very preliminary EXPERIMENTAL support for printing of general 2850 ASN1 structures. This currently produces rather ugly output and doesn't 2851 handle several customised structures at all. 2852 [Steve Henson] 2853 2854 *) Integrated support for PVK file format and some related formats such 2855 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support 2856 these in the 'rsa' and 'dsa' utilities. 2857 [Steve Henson] 2858 2859 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line. 2860 [Steve Henson] 2861 2862 *) Remove the ancient ASN1_METHOD code. This was only ever used in one 2863 place for the (very old) "NETSCAPE" format certificates which are now 2864 handled using new ASN1 code equivalents. 2865 [Steve Henson] 2866 2867 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD 2868 pointer and make the SSL_METHOD parameter in SSL_CTX_new, 2869 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'. 2870 [Nils Larsch] 2871 2872 *) Modify CRL distribution points extension code to print out previously 2873 unsupported fields. Enhance extension setting code to allow setting of 2874 all fields. 2875 [Steve Henson] 2876 2877 *) Add print and set support for Issuing Distribution Point CRL extension. 2878 [Steve Henson] 2879 2880 *) Change 'Configure' script to enable Camellia by default. 2881 [NTT] 2882 2883 Changes between 0.9.8m and 0.9.8n [24 Mar 2010] 2884 2885 *) When rejecting SSL/TLS records due to an incorrect version number, never 2886 update s->server with a new major version number. As of 2887 - OpenSSL 0.9.8m if 'short' is a 16-bit type, 2888 - OpenSSL 0.9.8f if 'short' is longer than 16 bits, 2889 the previous behavior could result in a read attempt at NULL when 2890 receiving specific incorrect SSL/TLS records once record payload 2891 protection is active. (CVE-2010-0740) 2892 [Bodo Moeller, Adam Langley <agl@chromium.org>] 2893 2894 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL 2895 could be crashed if the relevant tables were not present (e.g. chrooted). 2896 [Tomas Hoger <thoger@redhat.com>] 2897 2898 Changes between 0.9.8l and 0.9.8m [25 Feb 2010] 2899 2900 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245) 2901 [Martin Olsson, Neel Mehta] 2902 2903 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to 2904 accommodate for stack sorting, always a write lock!). 2905 [Bodo Moeller] 2906 2907 *) On some versions of WIN32 Heap32Next is very slow. This can cause 2908 excessive delays in the RAND_poll(): over a minute. As a workaround 2909 include a time check in the inner Heap32Next loop too. 2910 [Steve Henson] 2911 2912 *) The code that handled flushing of data in SSL/TLS originally used the 2913 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused 2914 the problem outlined in PR#1949. The fix suggested there however can 2915 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions 2916 of Apache). So instead simplify the code to flush unconditionally. 2917 This should be fine since flushing with no data to flush is a no op. 2918 [Steve Henson] 2919 2920 *) Handle TLS versions 2.0 and later properly and correctly use the 2921 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way 2922 off ancient servers have a habit of sticking around for a while... 2923 [Steve Henson] 2924 2925 *) Modify compression code so it frees up structures without using the 2926 ex_data callbacks. This works around a problem where some applications 2927 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when 2928 restarting) then use compression (e.g. SSL with compression) later. 2929 This results in significant per-connection memory leaks and 2930 has caused some security issues including CVE-2008-1678 and 2931 CVE-2009-4355. 2932 [Steve Henson] 2933 2934 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't 2935 change when encrypting or decrypting. 2936 [Bodo Moeller] 2937 2938 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to 2939 connect and renegotiate with servers which do not support RI. 2940 Until RI is more widely deployed this option is enabled by default. 2941 [Steve Henson] 2942 2943 *) Add "missing" ssl ctrls to clear options and mode. 2944 [Steve Henson] 2945 2946 *) If client attempts to renegotiate and doesn't support RI respond with 2947 a no_renegotiation alert as required by RFC5746. Some renegotiating 2948 TLS clients will continue a connection gracefully when they receive 2949 the alert. Unfortunately OpenSSL mishandled this alert and would hang 2950 waiting for a server hello which it will never receive. Now we treat a 2951 received no_renegotiation alert as a fatal error. This is because 2952 applications requesting a renegotiation might well expect it to succeed 2953 and would have no code in place to handle the server denying it so the 2954 only safe thing to do is to terminate the connection. 2955 [Steve Henson] 2956 2957 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if 2958 peer supports secure renegotiation and 0 otherwise. Print out peer 2959 renegotiation support in s_client/s_server. 2960 [Steve Henson] 2961 2962 *) Replace the highly broken and deprecated SPKAC certification method with 2963 the updated NID creation version. This should correctly handle UTF8. 2964 [Steve Henson] 2965 2966 *) Implement RFC5746. Re-enable renegotiation but require the extension 2967 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION 2968 turns out to be a bad idea. It has been replaced by 2969 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with 2970 SSL_CTX_set_options(). This is really not recommended unless you 2971 know what you are doing. 2972 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson] 2973 2974 *) Fixes to stateless session resumption handling. Use initial_ctx when 2975 issuing and attempting to decrypt tickets in case it has changed during 2976 servername handling. Use a non-zero length session ID when attempting 2977 stateless session resumption: this makes it possible to determine if 2978 a resumption has occurred immediately after receiving server hello 2979 (several places in OpenSSL subtly assume this) instead of later in 2980 the handshake. 2981 [Steve Henson] 2982 2983 *) The functions ENGINE_ctrl(), OPENSSL_isservice(), 2984 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error 2985 fixes for a few places where the return code is not checked 2986 correctly. 2987 [Julia Lawall <julia@diku.dk>] 2988 2989 *) Add --strict-warnings option to Configure script to include devteam 2990 warnings in other configurations. 2991 [Steve Henson] 2992 2993 *) Add support for --libdir option and LIBDIR variable in makefiles. This 2994 makes it possible to install openssl libraries in locations which 2995 have names other than "lib", for example "/usr/lib64" which some 2996 systems need. 2997 [Steve Henson, based on patch from Jeremy Utley] 2998 2999 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of 3000 X690 8.9.12 and can produce some misleading textual output of OIDs. 3001 [Steve Henson, reported by Dan Kaminsky] 3002 3003 *) Delete MD2 from algorithm tables. This follows the recommendation in 3004 several standards that it is not used in new applications due to 3005 several cryptographic weaknesses. For binary compatibility reasons 3006 the MD2 API is still compiled in by default. 3007 [Steve Henson] 3008 3009 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved 3010 and restored. 3011 [Steve Henson] 3012 3013 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and 3014 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name 3015 clash. 3016 [Guenter <lists@gknw.net>] 3017 3018 *) Fix the server certificate chain building code to use X509_verify_cert(), 3019 it used to have an ad-hoc builder which was unable to cope with anything 3020 other than a simple chain. 3021 [David Woodhouse <dwmw2@infradead.org>, Steve Henson] 3022 3023 *) Don't check self signed certificate signatures in X509_verify_cert() 3024 by default (a flag can override this): it just wastes time without 3025 adding any security. As a useful side effect self signed root CAs 3026 with non-FIPS digests are now usable in FIPS mode. 3027 [Steve Henson] 3028 3029 *) In dtls1_process_out_of_seq_message() the check if the current message 3030 is already buffered was missing. For every new message was memory 3031 allocated, allowing an attacker to perform an denial of service attack 3032 with sending out of seq handshake messages until there is no memory 3033 left. Additionally every future messege was buffered, even if the 3034 sequence number made no sense and would be part of another handshake. 3035 So only messages with sequence numbers less than 10 in advance will be 3036 buffered. (CVE-2009-1378) 3037 [Robin Seggelmann, discovered by Daniel Mentz] 3038 3039 *) Records are buffered if they arrive with a future epoch to be 3040 processed after finishing the corresponding handshake. There is 3041 currently no limitation to this buffer allowing an attacker to perform 3042 a DOS attack with sending records with future epochs until there is no 3043 memory left. This patch adds the pqueue_size() function to detemine 3044 the size of a buffer and limits the record buffer to 100 entries. 3045 (CVE-2009-1377) 3046 [Robin Seggelmann, discovered by Daniel Mentz] 3047 3048 *) Keep a copy of frag->msg_header.frag_len so it can be used after the 3049 parent structure is freed. (CVE-2009-1379) 3050 [Daniel Mentz] 3051 3052 *) Handle non-blocking I/O properly in SSL_shutdown() call. 3053 [Darryl Miles <darryl-mailinglists@netbauds.net>] 3054 3055 *) Add 2.5.4.* OIDs 3056 [Ilya O. <vrghost@gmail.com>] 3057 3058 Changes between 0.9.8k and 0.9.8l [5 Nov 2009] 3059 3060 *) Disable renegotiation completely - this fixes a severe security 3061 problem (CVE-2009-3555) at the cost of breaking all 3062 renegotiation. Renegotiation can be re-enabled by setting 3063 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at 3064 run-time. This is really not recommended unless you know what 3065 you're doing. 3066 [Ben Laurie] 3067 3068 Changes between 0.9.8j and 0.9.8k [25 Mar 2009] 3069 3070 *) Don't set val to NULL when freeing up structures, it is freed up by 3071 underlying code. If sizeof(void *) > sizeof(long) this can result in 3072 zeroing past the valid field. (CVE-2009-0789) 3073 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>] 3074 3075 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not 3076 checked correctly. This would allow some invalid signed attributes to 3077 appear to verify correctly. (CVE-2009-0591) 3078 [Ivan Nestlerode <inestlerode@us.ibm.com>] 3079 3080 *) Reject UniversalString and BMPString types with invalid lengths. This 3081 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have 3082 a legal length. (CVE-2009-0590) 3083 [Steve Henson] 3084 3085 *) Set S/MIME signing as the default purpose rather than setting it 3086 unconditionally. This allows applications to override it at the store 3087 level. 3088 [Steve Henson] 3089 3090 *) Permit restricted recursion of ASN1 strings. This is needed in practice 3091 to handle some structures. 3092 [Steve Henson] 3093 3094 *) Improve efficiency of mem_gets: don't search whole buffer each time 3095 for a '\n' 3096 [Jeremy Shapiro <jnshapir@us.ibm.com>] 3097 3098 *) New -hex option for openssl rand. 3099 [Matthieu Herrb] 3100 3101 *) Print out UTF8String and NumericString when parsing ASN1. 3102 [Steve Henson] 3103 3104 *) Support NumericString type for name components. 3105 [Steve Henson] 3106 3107 *) Allow CC in the environment to override the automatically chosen 3108 compiler. Note that nothing is done to ensure flags work with the 3109 chosen compiler. 3110 [Ben Laurie] 3111 3112 Changes between 0.9.8i and 0.9.8j [07 Jan 2009] 3113 3114 *) Properly check EVP_VerifyFinal() and similar return values 3115 (CVE-2008-5077). 3116 [Ben Laurie, Bodo Moeller, Google Security Team] 3117 3118 *) Enable TLS extensions by default. 3119 [Ben Laurie] 3120 3121 *) Allow the CHIL engine to be loaded, whether the application is 3122 multithreaded or not. (This does not release the developer from the 3123 obligation to set up the dynamic locking callbacks.) 3124 [Sander Temme <sander@temme.net>] 3125 3126 *) Use correct exit code if there is an error in dgst command. 3127 [Steve Henson; problem pointed out by Roland Dirlewanger] 3128 3129 *) Tweak Configure so that you need to say "experimental-jpake" to enable 3130 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. 3131 [Bodo Moeller] 3132 3133 *) Add experimental JPAKE support, including demo authentication in 3134 s_client and s_server. 3135 [Ben Laurie] 3136 3137 *) Set the comparison function in v3_addr_canonize(). 3138 [Rob Austein <sra@hactrn.net>] 3139 3140 *) Add support for XMPP STARTTLS in s_client. 3141 [Philip Paeps <philip@freebsd.org>] 3142 3143 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior 3144 to ensure that even with this option, only ciphersuites in the 3145 server's preference list will be accepted. (Note that the option 3146 applies only when resuming a session, so the earlier behavior was 3147 just about the algorithm choice for symmetric cryptography.) 3148 [Bodo Moeller] 3149 3150 Changes between 0.9.8h and 0.9.8i [15 Sep 2008] 3151 3152 *) Fix NULL pointer dereference if a DTLS server received 3153 ChangeCipherSpec as first record (CVE-2009-1386). 3154 [PR #1679] 3155 3156 *) Fix a state transitition in s3_srvr.c and d1_srvr.c 3157 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...). 3158 [Nagendra Modadugu] 3159 3160 *) The fix in 0.9.8c that supposedly got rid of unsafe 3161 double-checked locking was incomplete for RSA blinding, 3162 addressing just one layer of what turns out to have been 3163 doubly unsafe triple-checked locking. 3164 3165 So now fix this for real by retiring the MONT_HELPER macro 3166 in crypto/rsa/rsa_eay.c. 3167 3168 [Bodo Moeller; problem pointed out by Marius Schilder] 3169 3170 *) Various precautionary measures: 3171 3172 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). 3173 3174 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). 3175 (NB: This would require knowledge of the secret session ticket key 3176 to exploit, in which case you'd be SOL either way.) 3177 3178 - Change bn_nist.c so that it will properly handle input BIGNUMs 3179 outside the expected range. 3180 3181 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG 3182 builds. 3183 3184 [Neel Mehta, Bodo Moeller] 3185 3186 *) Allow engines to be "soft loaded" - i.e. optionally don't die if 3187 the load fails. Useful for distros. 3188 [Ben Laurie and the FreeBSD team] 3189 3190 *) Add support for Local Machine Keyset attribute in PKCS#12 files. 3191 [Steve Henson] 3192 3193 *) Fix BN_GF2m_mod_arr() top-bit cleanup code. 3194 [Huang Ying] 3195 3196 *) Expand ENGINE to support engine supplied SSL client certificate functions. 3197 3198 This work was sponsored by Logica. 3199 [Steve Henson] 3200 3201 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows 3202 keystores. Support for SSL/TLS client authentication too. 3203 Not compiled unless enable-capieng specified to Configure. 3204 3205 This work was sponsored by Logica. 3206 [Steve Henson] 3207 3208 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using 3209 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain 3210 attribute creation routines such as certifcate requests and PKCS#12 3211 files. 3212 [Steve Henson] 3213 3214 Changes between 0.9.8g and 0.9.8h [28 May 2008] 3215 3216 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS 3217 handshake which could lead to a cilent crash as found using the 3218 Codenomicon TLS test suite (CVE-2008-1672) 3219 [Steve Henson, Mark Cox] 3220 3221 *) Fix double free in TLS server name extensions which could lead to 3222 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891) 3223 [Joe Orton] 3224 3225 *) Clear error queue in SSL_CTX_use_certificate_chain_file() 3226 3227 Clear the error queue to ensure that error entries left from 3228 older function calls do not interfere with the correct operation. 3229 [Lutz Jaenicke, Erik de Castro Lopo] 3230 3231 *) Remove root CA certificates of commercial CAs: 3232 3233 The OpenSSL project does not recommend any specific CA and does not 3234 have any policy with respect to including or excluding any CA. 3235 Therefore it does not make any sense to ship an arbitrary selection 3236 of root CA certificates with the OpenSSL software. 3237 [Lutz Jaenicke] 3238 3239 *) RSA OAEP patches to fix two separate invalid memory reads. 3240 The first one involves inputs when 'lzero' is greater than 3241 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes 3242 before the beginning of from). The second one involves inputs where 3243 the 'db' section contains nothing but zeroes (there is a one-byte 3244 invalid read after the end of 'db'). 3245 [Ivan Nestlerode <inestlerode@us.ibm.com>] 3246 3247 *) Partial backport from 0.9.9-dev: 3248 3249 Introduce bn_mul_mont (dedicated Montgomery multiplication 3250 procedure) as a candidate for BIGNUM assembler implementation. 3251 While 0.9.9-dev uses assembler for various architectures, only 3252 x86_64 is available by default here in the 0.9.8 branch, and 3253 32-bit x86 is available through a compile-time setting. 3254 3255 To try the 32-bit x86 assembler implementation, use Configure 3256 option "enable-montasm" (which exists only for this backport). 3257 3258 As "enable-montasm" for 32-bit x86 disclaims code stability 3259 anyway, in this constellation we activate additional code 3260 backported from 0.9.9-dev for further performance improvements, 3261 namely BN_from_montgomery_word. (To enable this otherwise, 3262 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".) 3263 3264 [Andy Polyakov (backport partially by Bodo Moeller)] 3265 3266 *) Add TLS session ticket callback. This allows an application to set 3267 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed 3268 values. This is useful for key rollover for example where several key 3269 sets may exist with different names. 3270 [Steve Henson] 3271 3272 *) Reverse ENGINE-internal logic for caching default ENGINE handles. 3273 This was broken until now in 0.9.8 releases, such that the only way 3274 a registered ENGINE could be used (assuming it initialises 3275 successfully on the host) was to explicitly set it as the default 3276 for the relevant algorithms. This is in contradiction with 0.9.7 3277 behaviour and the documentation. With this fix, when an ENGINE is 3278 registered into a given algorithm's table of implementations, the 3279 'uptodate' flag is reset so that auto-discovery will be used next 3280 time a new context for that algorithm attempts to select an 3281 implementation. 3282 [Ian Lister (tweaked by Geoff Thorpe)] 3283 3284 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 3285 implemention in the following ways: 3286 3287 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be 3288 hard coded. 3289 3290 Lack of BER streaming support means one pass streaming processing is 3291 only supported if data is detached: setting the streaming flag is 3292 ignored for embedded content. 3293 3294 CMS support is disabled by default and must be explicitly enabled 3295 with the enable-cms configuration option. 3296 [Steve Henson] 3297 3298 *) Update the GMP engine glue to do direct copies between BIGNUM and 3299 mpz_t when openssl and GMP use the same limb size. Otherwise the 3300 existing "conversion via a text string export" trick is still used. 3301 [Paul Sheer <paulsheer@gmail.com>] 3302 3303 *) Zlib compression BIO. This is a filter BIO which compressed and 3304 uncompresses any data passed through it. 3305 [Steve Henson] 3306 3307 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement 3308 RFC3394 compatible AES key wrapping. 3309 [Steve Henson] 3310 3311 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): 3312 sets string data without copying. X509_ALGOR_set0() and 3313 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) 3314 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data 3315 from an X509_ATTRIBUTE structure optionally checking it occurs only 3316 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied 3317 data. 3318 [Steve Henson] 3319 3320 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() 3321 to get the expected BN_FLG_CONSTTIME behavior. 3322 [Bodo Moeller (Google)] 3323 3324 *) Netware support: 3325 3326 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets 3327 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) 3328 - added some more tests to do_tests.pl 3329 - fixed RunningProcess usage so that it works with newer LIBC NDKs too 3330 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency 3331 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, 3332 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc 3333 - various changes to netware.pl to enable gcc-cross builds on Win32 3334 platform 3335 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) 3336 - various changes to fix missing prototype warnings 3337 - fixed x86nasm.pl to create correct asm files for NASM COFF output 3338 - added AES, WHIRLPOOL and CPUID assembler code to build files 3339 - added missing AES assembler make rules to mk1mf.pl 3340 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply 3341 [Guenter Knauf <eflash@gmx.net>] 3342 3343 *) Implement certificate status request TLS extension defined in RFC3546. 3344 A client can set the appropriate parameters and receive the encoded 3345 OCSP response via a callback. A server can query the supplied parameters 3346 and set the encoded OCSP response in the callback. Add simplified examples 3347 to s_client and s_server. 3348 [Steve Henson] 3349 3350 Changes between 0.9.8f and 0.9.8g [19 Oct 2007] 3351 3352 *) Fix various bugs: 3353 + Binary incompatibility of ssl_ctx_st structure 3354 + DTLS interoperation with non-compliant servers 3355 + Don't call get_session_cb() without proposed session 3356 + Fix ia64 assembler code 3357 [Andy Polyakov, Steve Henson] 3358 3359 Changes between 0.9.8e and 0.9.8f [11 Oct 2007] 3360 3361 *) DTLS Handshake overhaul. There were longstanding issues with 3362 OpenSSL DTLS implementation, which were making it impossible for 3363 RFC 4347 compliant client to communicate with OpenSSL server. 3364 Unfortunately just fixing these incompatibilities would "cut off" 3365 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e 3366 server keeps tolerating non RFC compliant syntax. The opposite is 3367 not true, 0.9.8f client can not communicate with earlier server. 3368 This update even addresses CVE-2007-4995. 3369 [Andy Polyakov] 3370 3371 *) Changes to avoid need for function casts in OpenSSL: some compilers 3372 (gcc 4.2 and later) reject their use. 3373 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, 3374 Steve Henson] 3375 3376 *) Add RFC4507 support to OpenSSL. This includes the corrections in 3377 RFC4507bis. The encrypted ticket format is an encrypted encoded 3378 SSL_SESSION structure, that way new session features are automatically 3379 supported. 3380 3381 If a client application caches session in an SSL_SESSION structure 3382 support is transparent because tickets are now stored in the encoded 3383 SSL_SESSION. 3384 3385 The SSL_CTX structure automatically generates keys for ticket 3386 protection in servers so again support should be possible 3387 with no application modification. 3388 3389 If a client or server wishes to disable RFC4507 support then the option 3390 SSL_OP_NO_TICKET can be set. 3391 3392 Add a TLS extension debugging callback to allow the contents of any client 3393 or server extensions to be examined. 3394 3395 This work was sponsored by Google. 3396 [Steve Henson] 3397 3398 *) Add initial support for TLS extensions, specifically for the server_name 3399 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 3400 have new members for a host name. The SSL data structure has an 3401 additional member SSL_CTX *initial_ctx so that new sessions can be 3402 stored in that context to allow for session resumption, even after the 3403 SSL has been switched to a new SSL_CTX in reaction to a client's 3404 server_name extension. 3405 3406 New functions (subject to change): 3407 3408 SSL_get_servername() 3409 SSL_get_servername_type() 3410 SSL_set_SSL_CTX() 3411 3412 New CTRL codes and macros (subject to change): 3413 3414 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 3415 - SSL_CTX_set_tlsext_servername_callback() 3416 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 3417 - SSL_CTX_set_tlsext_servername_arg() 3418 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 3419 3420 openssl s_client has a new '-servername ...' option. 3421 3422 openssl s_server has new options '-servername_host ...', '-cert2 ...', 3423 '-key2 ...', '-servername_fatal' (subject to change). This allows 3424 testing the HostName extension for a specific single host name ('-cert' 3425 and '-key' remain fallbacks for handshakes without HostName 3426 negotiation). If the unrecogninzed_name alert has to be sent, this by 3427 default is a warning; it becomes fatal with the '-servername_fatal' 3428 option. 3429 3430 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson] 3431 3432 *) Add AES and SSE2 assembly language support to VC++ build. 3433 [Steve Henson] 3434 3435 *) Mitigate attack on final subtraction in Montgomery reduction. 3436 [Andy Polyakov] 3437 3438 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 3439 (which previously caused an internal error). 3440 [Bodo Moeller] 3441 3442 *) Squeeze another 10% out of IGE mode when in != out. 3443 [Ben Laurie] 3444 3445 *) AES IGE mode speedup. 3446 [Dean Gaudet (Google)] 3447 3448 *) Add the Korean symmetric 128-bit cipher SEED (see 3449 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and 3450 add SEED ciphersuites from RFC 4162: 3451 3452 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" 3453 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" 3454 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" 3455 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" 3456 3457 To minimize changes between patchlevels in the OpenSSL 0.9.8 3458 series, SEED remains excluded from compilation unless OpenSSL 3459 is configured with 'enable-seed'. 3460 [KISA, Bodo Moeller] 3461 3462 *) Mitigate branch prediction attacks, which can be practical if a 3463 single processor is shared, allowing a spy process to extract 3464 information. For detailed background information, see 3465 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron, 3466 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL 3467 and Necessary Software Countermeasures"). The core of the change 3468 are new versions BN_div_no_branch() and 3469 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), 3470 respectively, which are slower, but avoid the security-relevant 3471 conditional branches. These are automatically called by BN_div() 3472 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one 3473 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to 3474 remove a conditional branch. 3475 3476 BN_FLG_CONSTTIME is the new name for the previous 3477 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just 3478 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag 3479 in the exponent causes BN_mod_exp_mont() to use the alternative 3480 implementation in BN_mod_exp_mont_consttime().) The old name 3481 remains as a deprecated alias. 3482 3483 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general 3484 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses 3485 constant-time implementations for more than just exponentiation. 3486 Here too the old name is kept as a deprecated alias. 3487 3488 BN_BLINDING_new() will now use BN_dup() for the modulus so that 3489 the BN_BLINDING structure gets an independent copy of the 3490 modulus. This means that the previous "BIGNUM *m" argument to 3491 BN_BLINDING_new() and to BN_BLINDING_create_param() now 3492 essentially becomes "const BIGNUM *m", although we can't actually 3493 change this in the header file before 0.9.9. It allows 3494 RSA_setup_blinding() to use BN_with_flags() on the modulus to 3495 enable BN_FLG_CONSTTIME. 3496 3497 [Matthew D Wood (Intel Corp)] 3498 3499 *) In the SSL/TLS server implementation, be strict about session ID 3500 context matching (which matters if an application uses a single 3501 external cache for different purposes). Previously, 3502 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was 3503 set. This did ensure strict client verification, but meant that, 3504 with applications using a single external cache for quite 3505 different requirements, clients could circumvent ciphersuite 3506 restrictions for a given session ID context by starting a session 3507 in a different context. 3508 [Bodo Moeller] 3509 3510 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 3511 a ciphersuite string such as "DEFAULT:RSA" cannot enable 3512 authentication-only ciphersuites. 3513 [Bodo Moeller] 3514 3515 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was 3516 not complete and could lead to a possible single byte overflow 3517 (CVE-2007-5135) [Ben Laurie] 3518 3519 Changes between 0.9.8d and 0.9.8e [23 Feb 2007] 3520 3521 *) Since AES128 and AES256 (and similarly Camellia128 and 3522 Camellia256) share a single mask bit in the logic of 3523 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 3524 kludge to work properly if AES128 is available and AES256 isn't 3525 (or if Camellia128 is available and Camellia256 isn't). 3526 [Victor Duchovni] 3527 3528 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c 3529 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): 3530 When a point or a seed is encoded in a BIT STRING, we need to 3531 prevent the removal of trailing zero bits to get the proper DER 3532 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case 3533 of a NamedBitList, for which trailing 0 bits need to be removed.) 3534 [Bodo Moeller] 3535 3536 *) Have SSL/TLS server implementation tolerate "mismatched" record 3537 protocol version while receiving ClientHello even if the 3538 ClientHello is fragmented. (The server can't insist on the 3539 particular protocol version it has chosen before the ServerHello 3540 message has informed the client about his choice.) 3541 [Bodo Moeller] 3542 3543 *) Add RFC 3779 support. 3544 [Rob Austein for ARIN, Ben Laurie] 3545 3546 *) Load error codes if they are not already present instead of using a 3547 static variable. This allows them to be cleanly unloaded and reloaded. 3548 Improve header file function name parsing. 3549 [Steve Henson] 3550 3551 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO 3552 or CAPABILITY handshake as required by RFCs. 3553 [Goetz Babin-Ebell] 3554 3555 Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 3556 3557 *) Introduce limits to prevent malicious keys being able to 3558 cause a denial of service. (CVE-2006-2940) 3559 [Steve Henson, Bodo Moeller] 3560 3561 *) Fix ASN.1 parsing of certain invalid structures that can result 3562 in a denial of service. (CVE-2006-2937) [Steve Henson] 3563 3564 *) Fix buffer overflow in SSL_get_shared_ciphers() function. 3565 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 3566 3567 *) Fix SSL client code which could crash if connecting to a 3568 malicious SSLv2 server. (CVE-2006-4343) 3569 [Tavis Ormandy and Will Drewry, Google Security Team] 3570 3571 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 3572 match only those. Before that, "AES256-SHA" would be interpreted 3573 as a pattern and match "AES128-SHA" too (since AES128-SHA got 3574 the same strength classification in 0.9.7h) as we currently only 3575 have a single AES bit in the ciphersuite description bitmap. 3576 That change, however, also applied to ciphersuite strings such as 3577 "RC4-MD5" that intentionally matched multiple ciphersuites -- 3578 namely, SSL 2.0 ciphersuites in addition to the more common ones 3579 from SSL 3.0/TLS 1.0. 3580 3581 So we change the selection algorithm again: Naming an explicit 3582 ciphersuite selects this one ciphersuite, and any other similar 3583 ciphersuite (same bitmap) from *other* protocol versions. 3584 Thus, "RC4-MD5" again will properly select both the SSL 2.0 3585 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 3586 3587 Since SSL 2.0 does not have any ciphersuites for which the 3588 128/256 bit distinction would be relevant, this works for now. 3589 The proper fix will be to use different bits for AES128 and 3590 AES256, which would have avoided the problems from the beginning; 3591 however, bits are scarce, so we can only do this in a new release 3592 (not just a patchlevel) when we can change the SSL_CIPHER 3593 definition to split the single 'unsigned long mask' bitmap into 3594 multiple values to extend the available space. 3595 3596 [Bodo Moeller] 3597 3598 Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 3599 3600 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 3601 (CVE-2006-4339) [Ben Laurie and Google Security Team] 3602 3603 *) Add AES IGE and biIGE modes. 3604 [Ben Laurie] 3605 3606 *) Change the Unix randomness entropy gathering to use poll() when 3607 possible instead of select(), since the latter has some 3608 undesirable limitations. 3609 [Darryl Miles via Richard Levitte and Bodo Moeller] 3610 3611 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special 3612 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 3613 cannot be implicitly activated as part of, e.g., the "AES" alias. 3614 However, please upgrade to OpenSSL 0.9.9[-dev] for 3615 non-experimental use of the ECC ciphersuites to get TLS extension 3616 support, which is required for curve and point format negotiation 3617 to avoid potential handshake problems. 3618 [Bodo Moeller] 3619 3620 *) Disable rogue ciphersuites: 3621 3622 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 3623 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 3624 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 3625 3626 The latter two were purportedly from 3627 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 3628 appear there. 3629 3630 Also deactivate the remaining ciphersuites from 3631 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 3632 unofficial, and the ID has long expired. 3633 [Bodo Moeller] 3634 3635 *) Fix RSA blinding Heisenbug (problems sometimes occured on 3636 dual-core machines) and other potential thread-safety issues. 3637 [Bodo Moeller] 3638 3639 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 3640 versions), which is now available for royalty-free use 3641 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). 3642 Also, add Camellia TLS ciphersuites from RFC 4132. 3643 3644 To minimize changes between patchlevels in the OpenSSL 0.9.8 3645 series, Camellia remains excluded from compilation unless OpenSSL 3646 is configured with 'enable-camellia'. 3647 [NTT] 3648 3649 *) Disable the padding bug check when compression is in use. The padding 3650 bug check assumes the first packet is of even length, this is not 3651 necessarily true if compresssion is enabled and can result in false 3652 positives causing handshake failure. The actual bug test is ancient 3653 code so it is hoped that implementations will either have fixed it by 3654 now or any which still have the bug do not support compression. 3655 [Steve Henson] 3656 3657 Changes between 0.9.8a and 0.9.8b [04 May 2006] 3658 3659 *) When applying a cipher rule check to see if string match is an explicit 3660 cipher suite and only match that one cipher suite if it is. 3661 [Steve Henson] 3662 3663 *) Link in manifests for VC++ if needed. 3664 [Austin Ziegler <halostatue@gmail.com>] 3665 3666 *) Update support for ECC-based TLS ciphersuites according to 3667 draft-ietf-tls-ecc-12.txt with proposed changes (but without 3668 TLS extensions, which are supported starting with the 0.9.9 3669 branch, not in the OpenSSL 0.9.8 branch). 3670 [Douglas Stebila] 3671 3672 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 3673 opaque EVP_CIPHER_CTX handling. 3674 [Steve Henson] 3675 3676 *) Fixes and enhancements to zlib compression code. We now only use 3677 "zlib1.dll" and use the default __cdecl calling convention on Win32 3678 to conform with the standards mentioned here: 3679 http://www.zlib.net/DLL_FAQ.txt 3680 Static zlib linking now works on Windows and the new --with-zlib-include 3681 --with-zlib-lib options to Configure can be used to supply the location 3682 of the headers and library. Gracefully handle case where zlib library 3683 can't be loaded. 3684 [Steve Henson] 3685 3686 *) Several fixes and enhancements to the OID generation code. The old code 3687 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 3688 handle numbers larger than ULONG_MAX, truncated printing and had a 3689 non standard OBJ_obj2txt() behaviour. 3690 [Steve Henson] 3691 3692 *) Add support for building of engines under engine/ as shared libraries 3693 under VC++ build system. 3694 [Steve Henson] 3695 3696 *) Corrected the numerous bugs in the Win32 path splitter in DSO. 3697 Hopefully, we will not see any false combination of paths any more. 3698 [Richard Levitte] 3699 3700 Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 3701 3702 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 3703 (part of SSL_OP_ALL). This option used to disable the 3704 countermeasure against man-in-the-middle protocol-version 3705 rollback in the SSL 2.0 server implementation, which is a bad 3706 idea. (CVE-2005-2969) 3707 3708 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 3709 for Information Security, National Institute of Advanced Industrial 3710 Science and Technology [AIST], Japan)] 3711 3712 *) Add two function to clear and return the verify parameter flags. 3713 [Steve Henson] 3714 3715 *) Keep cipherlists sorted in the source instead of sorting them at 3716 runtime, thus removing the need for a lock. 3717 [Nils Larsch] 3718 3719 *) Avoid some small subgroup attacks in Diffie-Hellman. 3720 [Nick Mathewson and Ben Laurie] 3721 3722 *) Add functions for well-known primes. 3723 [Nick Mathewson] 3724 3725 *) Extended Windows CE support. 3726 [Satoshi Nakamura and Andy Polyakov] 3727 3728 *) Initialize SSL_METHOD structures at compile time instead of during 3729 runtime, thus removing the need for a lock. 3730 [Steve Henson] 3731 3732 *) Make PKCS7_decrypt() work even if no certificate is supplied by 3733 attempting to decrypt each encrypted key in turn. Add support to 3734 smime utility. 3735 [Steve Henson] 3736 3737 Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 3738 3739 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 3740 OpenSSL 0.9.8.] 3741 3742 *) Add libcrypto.pc and libssl.pc for those who feel they need them. 3743 [Richard Levitte] 3744 3745 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 3746 key into the same file any more. 3747 [Richard Levitte] 3748 3749 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 3750 [Andy Polyakov] 3751 3752 *) Add -utf8 command line and config file option to 'ca'. 3753 [Stefan <stf@udoma.org] 3754 3755 *) Removed the macro des_crypt(), as it seems to conflict with some 3756 libraries. Use DES_crypt(). 3757 [Richard Levitte] 3758 3759 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 3760 involves renaming the source and generated shared-libs for 3761 both. The engines will accept the corrected or legacy ids 3762 ('ncipher' and '4758_cca' respectively) when binding. NB, 3763 this only applies when building 'shared'. 3764 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 3765 3766 *) Add attribute functions to EVP_PKEY structure. Modify 3767 PKCS12_create() to recognize a CSP name attribute and 3768 use it. Make -CSP option work again in pkcs12 utility. 3769 [Steve Henson] 3770 3771 *) Add new functionality to the bn blinding code: 3772 - automatic re-creation of the BN_BLINDING parameters after 3773 a fixed number of uses (currently 32) 3774 - add new function for parameter creation 3775 - introduce flags to control the update behaviour of the 3776 BN_BLINDING parameters 3777 - hide BN_BLINDING structure 3778 Add a second BN_BLINDING slot to the RSA structure to improve 3779 performance when a single RSA object is shared among several 3780 threads. 3781 [Nils Larsch] 3782 3783 *) Add support for DTLS. 3784 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 3785 3786 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 3787 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 3788 [Walter Goulet] 3789 3790 *) Remove buggy and incompletet DH cert support from 3791 ssl/ssl_rsa.c and ssl/s3_both.c 3792 [Nils Larsch] 3793 3794 *) Use SHA-1 instead of MD5 as the default digest algorithm for 3795 the apps/openssl applications. 3796 [Nils Larsch] 3797 3798 *) Compile clean with "-Wall -Wmissing-prototypes 3799 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 3800 DEBUG_SAFESTACK must also be set. 3801 [Ben Laurie] 3802 3803 *) Change ./Configure so that certain algorithms can be disabled by default. 3804 The new counterpiece to "no-xxx" is "enable-xxx". 3805 3806 The patented RC5 and MDC2 algorithms will now be disabled unless 3807 "enable-rc5" and "enable-mdc2", respectively, are specified. 3808 3809 (IDEA remains enabled despite being patented. This is because IDEA 3810 is frequently required for interoperability, and there is no license 3811 fee for non-commercial use. As before, "no-idea" can be used to 3812 avoid this algorithm.) 3813 3814 [Bodo Moeller] 3815 3816 *) Add processing of proxy certificates (see RFC 3820). This work was 3817 sponsored by KTH (The Royal Institute of Technology in Stockholm) and 3818 EGEE (Enabling Grids for E-science in Europe). 3819 [Richard Levitte] 3820 3821 *) RC4 performance overhaul on modern architectures/implementations, such 3822 as Intel P4, IA-64 and AMD64. 3823 [Andy Polyakov] 3824 3825 *) New utility extract-section.pl. This can be used specify an alternative 3826 section number in a pod file instead of having to treat each file as 3827 a separate case in Makefile. This can be done by adding two lines to the 3828 pod file: 3829 3830 =for comment openssl_section:XXX 3831 3832 The blank line is mandatory. 3833 3834 [Steve Henson] 3835 3836 *) New arguments -certform, -keyform and -pass for s_client and s_server 3837 to allow alternative format key and certificate files and passphrase 3838 sources. 3839 [Steve Henson] 3840 3841 *) New structure X509_VERIFY_PARAM which combines current verify parameters, 3842 update associated structures and add various utility functions. 3843 3844 Add new policy related verify parameters, include policy checking in 3845 standard verify code. Enhance 'smime' application with extra parameters 3846 to support policy checking and print out. 3847 [Steve Henson] 3848 3849 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 3850 Nehemiah processors. These extensions support AES encryption in hardware 3851 as well as RNG (though RNG support is currently disabled). 3852 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 3853 3854 *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 3855 [Geoff Thorpe] 3856 3857 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 3858 [Andy Polyakov and a number of other people] 3859 3860 *) Improved PowerPC platform support. Most notably BIGNUM assembler 3861 implementation contributed by IBM. 3862 [Suresh Chari, Peter Waltenberg, Andy Polyakov] 3863 3864 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 3865 exponent rather than 'unsigned long'. There is a corresponding change to 3866 the new 'rsa_keygen' element of the RSA_METHOD structure. 3867 [Jelte Jansen, Geoff Thorpe] 3868 3869 *) Functionality for creating the initial serial number file is now 3870 moved from CA.pl to the 'ca' utility with a new option -create_serial. 3871 3872 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 3873 number file to 1, which is bound to cause problems. To avoid 3874 the problems while respecting compatibility between different 0.9.7 3875 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 3876 CA.pl for serial number initialization. With the new release 0.9.8, 3877 we can fix the problem directly in the 'ca' utility.) 3878 [Steve Henson] 3879 3880 *) Reduced header interdepencies by declaring more opaque objects in 3881 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 3882 give fewer recursive includes, which could break lazy source code - so 3883 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 3884 developers should define this symbol when building and using openssl to 3885 ensure they track the recommended behaviour, interfaces, [etc], but 3886 backwards-compatible behaviour prevails when this isn't defined. 3887 [Geoff Thorpe] 3888 3889 *) New function X509_POLICY_NODE_print() which prints out policy nodes. 3890 [Steve Henson] 3891 3892 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 3893 This will generate a random key of the appropriate length based on the 3894 cipher context. The EVP_CIPHER can provide its own random key generation 3895 routine to support keys of a specific form. This is used in the des and 3896 3des routines to generate a key of the correct parity. Update S/MIME 3897 code to use new functions and hence generate correct parity DES keys. 3898 Add EVP_CHECK_DES_KEY #define to return an error if the key is not 3899 valid (weak or incorrect parity). 3900 [Steve Henson] 3901 3902 *) Add a local set of CRLs that can be used by X509_verify_cert() as well 3903 as looking them up. This is useful when the verified structure may contain 3904 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 3905 present unless the new PKCS7_NO_CRL flag is asserted. 3906 [Steve Henson] 3907 3908 *) Extend ASN1 oid configuration module. It now additionally accepts the 3909 syntax: 3910 3911 shortName = some long name, 1.2.3.4 3912 [Steve Henson] 3913 3914 *) Reimplemented the BN_CTX implementation. There is now no more static 3915 limitation on the number of variables it can handle nor the depth of the 3916 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 3917 information can now expand as required, and rather than having a single 3918 static array of bignums, BN_CTX now uses a linked-list of such arrays 3919 allowing it to expand on demand whilst maintaining the usefulness of 3920 BN_CTX's "bundling". 3921 [Geoff Thorpe] 3922 3923 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 3924 to allow all RSA operations to function using a single BN_CTX. 3925 [Geoff Thorpe] 3926 3927 *) Preliminary support for certificate policy evaluation and checking. This 3928 is initially intended to pass the tests outlined in "Conformance Testing 3929 of Relying Party Client Certificate Path Processing Logic" v1.07. 3930 [Steve Henson] 3931 3932 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 3933 remained unused and not that useful. A variety of other little bignum 3934 tweaks and fixes have also been made continuing on from the audit (see 3935 below). 3936 [Geoff Thorpe] 3937 3938 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 3939 associated ASN1, EVP and SSL functions and old ASN1 macros. 3940 [Richard Levitte] 3941 3942 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 3943 and this should never fail. So the return value from the use of 3944 BN_set_word() (which can fail due to needless expansion) is now deprecated; 3945 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 3946 [Geoff Thorpe] 3947 3948 *) BN_CTX_get() should return zero-valued bignums, providing the same 3949 initialised value as BN_new(). 3950 [Geoff Thorpe, suggested by Ulf M��ller] 3951 3952 *) Support for inhibitAnyPolicy certificate extension. 3953 [Steve Henson] 3954 3955 *) An audit of the BIGNUM code is underway, for which debugging code is 3956 enabled when BN_DEBUG is defined. This makes stricter enforcements on what 3957 is considered valid when processing BIGNUMs, and causes execution to 3958 assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 3959 further steps are taken to deliberately pollute unused data in BIGNUM 3960 structures to try and expose faulty code further on. For now, openssl will 3961 (in its default mode of operation) continue to tolerate the inconsistent 3962 forms that it has tolerated in the past, but authors and packagers should 3963 consider trying openssl and their own applications when compiled with 3964 these debugging symbols defined. It will help highlight potential bugs in 3965 their own code, and will improve the test coverage for OpenSSL itself. At 3966 some point, these tighter rules will become openssl's default to improve 3967 maintainability, though the assert()s and other overheads will remain only 3968 in debugging configurations. See bn.h for more details. 3969 [Geoff Thorpe, Nils Larsch, Ulf M��ller] 3970 3971 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 3972 that can only be obtained through BN_CTX_new() (which implicitly 3973 initialises it). The presence of this function only made it possible 3974 to overwrite an existing structure (and cause memory leaks). 3975 [Geoff Thorpe] 3976 3977 *) Because of the callback-based approach for implementing LHASH as a 3978 template type, lh_insert() adds opaque objects to hash-tables and 3979 lh_doall() or lh_doall_arg() are typically used with a destructor callback 3980 to clean up those corresponding objects before destroying the hash table 3981 (and losing the object pointers). So some over-zealous constifications in 3982 LHASH have been relaxed so that lh_insert() does not take (nor store) the 3983 objects as "const" and the lh_doall[_arg] callback wrappers are not 3984 prototyped to have "const" restrictions on the object pointers they are 3985 given (and so aren't required to cast them away any more). 3986 [Geoff Thorpe] 3987 3988 *) The tmdiff.h API was so ugly and minimal that our own timing utility 3989 (speed) prefers to use its own implementation. The two implementations 3990 haven't been consolidated as yet (volunteers?) but the tmdiff API has had 3991 its object type properly exposed (MS_TM) instead of casting to/from "char 3992 *". This may still change yet if someone realises MS_TM and "ms_time_***" 3993 aren't necessarily the greatest nomenclatures - but this is what was used 3994 internally to the implementation so I've used that for now. 3995 [Geoff Thorpe] 3996 3997 *) Ensure that deprecated functions do not get compiled when 3998 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 3999 the self-tests were still using deprecated key-generation functions so 4000 these have been updated also. 4001 [Geoff Thorpe] 4002 4003 *) Reorganise PKCS#7 code to separate the digest location functionality 4004 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 4005 New function PKCS7_set_digest() to set the digest type for PKCS#7 4006 digestedData type. Add additional code to correctly generate the 4007 digestedData type and add support for this type in PKCS7 initialization 4008 functions. 4009 [Steve Henson] 4010 4011 *) New function PKCS7_set0_type_other() this initializes a PKCS7 4012 structure of type "other". 4013 [Steve Henson] 4014 4015 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 4016 sure the loop does correctly stop and breaking ("division by zero") 4017 modulus operations are not performed. The (pre-generated) prime 4018 table crypto/bn/bn_prime.h was already correct, but it could not be 4019 re-generated on some platforms because of the "division by zero" 4020 situation in the script. 4021 [Ralf S. Engelschall] 4022 4023 *) Update support for ECC-based TLS ciphersuites according to 4024 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 4025 SHA-1 now is only used for "small" curves (where the 4026 representation of a field element takes up to 24 bytes); for 4027 larger curves, the field element resulting from ECDH is directly 4028 used as premaster secret. 4029 [Douglas Stebila (Sun Microsystems Laboratories)] 4030 4031 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 4032 curve secp160r1 to the tests. 4033 [Douglas Stebila (Sun Microsystems Laboratories)] 4034 4035 *) Add the possibility to load symbols globally with DSO. 4036 [G��tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 4037 4038 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 4039 control of the error stack. 4040 [Richard Levitte] 4041 4042 *) Add support for STORE in ENGINE. 4043 [Richard Levitte] 4044 4045 *) Add the STORE type. The intention is to provide a common interface 4046 to certificate and key stores, be they simple file-based stores, or 4047 HSM-type store, or LDAP stores, or... 4048 NOTE: The code is currently UNTESTED and isn't really used anywhere. 4049 [Richard Levitte] 4050 4051 *) Add a generic structure called OPENSSL_ITEM. This can be used to 4052 pass a list of arguments to any function as well as provide a way 4053 for a function to pass data back to the caller. 4054 [Richard Levitte] 4055 4056 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 4057 works like BUF_strdup() but can be used to duplicate a portion of 4058 a string. The copy gets NUL-terminated. BUF_memdup() duplicates 4059 a memory area. 4060 [Richard Levitte] 4061 4062 *) Add the function sk_find_ex() which works like sk_find(), but will 4063 return an index to an element even if an exact match couldn't be 4064 found. The index is guaranteed to point at the element where the 4065 searched-for key would be inserted to preserve sorting order. 4066 [Richard Levitte] 4067 4068 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 4069 takes an extra flags argument for optional functionality. Currently, 4070 the following flags are defined: 4071 4072 OBJ_BSEARCH_VALUE_ON_NOMATCH 4073 This one gets OBJ_bsearch_ex() to return a pointer to the first 4074 element where the comparing function returns a negative or zero 4075 number. 4076 4077 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 4078 This one gets OBJ_bsearch_ex() to return a pointer to the first 4079 element where the comparing function returns zero. This is useful 4080 if there are more than one element where the comparing function 4081 returns zero. 4082 [Richard Levitte] 4083 4084 *) Make it possible to create self-signed certificates with 'openssl ca' 4085 in such a way that the self-signed certificate becomes part of the 4086 CA database and uses the same mechanisms for serial number generation 4087 as all other certificate signing. The new flag '-selfsign' enables 4088 this functionality. Adapt CA.sh and CA.pl.in. 4089 [Richard Levitte] 4090 4091 *) Add functionality to check the public key of a certificate request 4092 against a given private. This is useful to check that a certificate 4093 request can be signed by that key (self-signing). 4094 [Richard Levitte] 4095 4096 *) Make it possible to have multiple active certificates with the same 4097 subject in the CA index file. This is done only if the keyword 4098 'unique_subject' is set to 'no' in the main CA section (default 4099 if 'CA_default') of the configuration file. The value is saved 4100 with the database itself in a separate index attribute file, 4101 named like the index file with '.attr' appended to the name. 4102 [Richard Levitte] 4103 4104 *) Generate muti valued AVAs using '+' notation in config files for 4105 req and dirName. 4106 [Steve Henson] 4107 4108 *) Support for nameConstraints certificate extension. 4109 [Steve Henson] 4110 4111 *) Support for policyConstraints certificate extension. 4112 [Steve Henson] 4113 4114 *) Support for policyMappings certificate extension. 4115 [Steve Henson] 4116 4117 *) Make sure the default DSA_METHOD implementation only uses its 4118 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 4119 and change its own handlers to be NULL so as to remove unnecessary 4120 indirection. This lets alternative implementations fallback to the 4121 default implementation more easily. 4122 [Geoff Thorpe] 4123 4124 *) Support for directoryName in GeneralName related extensions 4125 in config files. 4126 [Steve Henson] 4127 4128 *) Make it possible to link applications using Makefile.shared. 4129 Make that possible even when linking against static libraries! 4130 [Richard Levitte] 4131 4132 *) Support for single pass processing for S/MIME signing. This now 4133 means that S/MIME signing can be done from a pipe, in addition 4134 cleartext signing (multipart/signed type) is effectively streaming 4135 and the signed data does not need to be all held in memory. 4136 4137 This is done with a new flag PKCS7_STREAM. When this flag is set 4138 PKCS7_sign() only initializes the PKCS7 structure and the actual signing 4139 is done after the data is output (and digests calculated) in 4140 SMIME_write_PKCS7(). 4141 [Steve Henson] 4142 4143 *) Add full support for -rpath/-R, both in shared libraries and 4144 applications, at least on the platforms where it's known how 4145 to do it. 4146 [Richard Levitte] 4147 4148 *) In crypto/ec/ec_mult.c, implement fast point multiplication with 4149 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 4150 will now compute a table of multiples of the generator that 4151 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 4152 faster (notably in the case of a single point multiplication, 4153 scalar * generator). 4154 [Nils Larsch, Bodo Moeller] 4155 4156 *) IPv6 support for certificate extensions. The various extensions 4157 which use the IP:a.b.c.d can now take IPv6 addresses using the 4158 formats of RFC1884 2.2 . IPv6 addresses are now also displayed 4159 correctly. 4160 [Steve Henson] 4161 4162 *) Added an ENGINE that implements RSA by performing private key 4163 exponentiations with the GMP library. The conversions to and from 4164 GMP's mpz_t format aren't optimised nor are any montgomery forms 4165 cached, and on x86 it appears OpenSSL's own performance has caught up. 4166 However there are likely to be other architectures where GMP could 4167 provide a boost. This ENGINE is not built in by default, but it can be 4168 specified at Configure time and should be accompanied by the necessary 4169 linker additions, eg; 4170 ./config -DOPENSSL_USE_GMP -lgmp 4171 [Geoff Thorpe] 4172 4173 *) "openssl engine" will not display ENGINE/DSO load failure errors when 4174 testing availability of engines with "-t" - the old behaviour is 4175 produced by increasing the feature's verbosity with "-tt". 4176 [Geoff Thorpe] 4177 4178 *) ECDSA routines: under certain error conditions uninitialized BN objects 4179 could be freed. Solution: make sure initialization is performed early 4180 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 4181 via PR#459) 4182 [Lutz Jaenicke] 4183 4184 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 4185 and DH_METHOD (eg. by ENGINE implementations) to override the normal 4186 software implementations. For DSA and DH, parameter generation can 4187 also be overriden by providing the appropriate method callbacks. 4188 [Geoff Thorpe] 4189 4190 *) Change the "progress" mechanism used in key-generation and 4191 primality testing to functions that take a new BN_GENCB pointer in 4192 place of callback/argument pairs. The new API functions have "_ex" 4193 postfixes and the older functions are reimplemented as wrappers for 4194 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 4195 declarations of the old functions to help (graceful) attempts to 4196 migrate to the new functions. Also, the new key-generation API 4197 functions operate on a caller-supplied key-structure and return 4198 success/failure rather than returning a key or NULL - this is to 4199 help make "keygen" another member function of RSA_METHOD etc. 4200 4201 Example for using the new callback interface: 4202 4203 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 4204 void *my_arg = ...; 4205 BN_GENCB my_cb; 4206 4207 BN_GENCB_set(&my_cb, my_callback, my_arg); 4208 4209 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 4210 /* For the meaning of a, b in calls to my_callback(), see the 4211 * documentation of the function that calls the callback. 4212 * cb will point to my_cb; my_arg can be retrieved as cb->arg. 4213 * my_callback should return 1 if it wants BN_is_prime_ex() 4214 * to continue, or 0 to stop. 4215 */ 4216 4217 [Geoff Thorpe] 4218 4219 *) Change the ZLIB compression method to be stateful, and make it 4220 available to TLS with the number defined in 4221 draft-ietf-tls-compression-04.txt. 4222 [Richard Levitte] 4223 4224 *) Add the ASN.1 structures and functions for CertificatePair, which 4225 is defined as follows (according to X.509_4thEditionDraftV6.pdf): 4226 4227 CertificatePair ::= SEQUENCE { 4228 forward [0] Certificate OPTIONAL, 4229 reverse [1] Certificate OPTIONAL, 4230 -- at least one of the pair shall be present -- } 4231 4232 Also implement the PEM functions to read and write certificate 4233 pairs, and defined the PEM tag as "CERTIFICATE PAIR". 4234 4235 This needed to be defined, mostly for the sake of the LDAP 4236 attribute crossCertificatePair, but may prove useful elsewhere as 4237 well. 4238 [Richard Levitte] 4239 4240 *) Make it possible to inhibit symlinking of shared libraries in 4241 Makefile.shared, for Cygwin's sake. 4242 [Richard Levitte] 4243 4244 *) Extend the BIGNUM API by creating a function 4245 void BN_set_negative(BIGNUM *a, int neg); 4246 and a macro that behave like 4247 int BN_is_negative(const BIGNUM *a); 4248 4249 to avoid the need to access 'a->neg' directly in applications. 4250 [Nils Larsch] 4251 4252 *) Implement fast modular reduction for pseudo-Mersenne primes 4253 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 4254 EC_GROUP_new_curve_GFp() will now automatically use this 4255 if applicable. 4256 [Nils Larsch <nla@trustcenter.de>] 4257 4258 *) Add new lock type (CRYPTO_LOCK_BN). 4259 [Bodo Moeller] 4260 4261 *) Change the ENGINE framework to automatically load engines 4262 dynamically from specific directories unless they could be 4263 found to already be built in or loaded. Move all the 4264 current engines except for the cryptodev one to a new 4265 directory engines/. 4266 The engines in engines/ are built as shared libraries if 4267 the "shared" options was given to ./Configure or ./config. 4268 Otherwise, they are inserted in libcrypto.a. 4269 /usr/local/ssl/engines is the default directory for dynamic 4270 engines, but that can be overriden at configure time through 4271 the usual use of --prefix and/or --openssldir, and at run 4272 time with the environment variable OPENSSL_ENGINES. 4273 [Geoff Thorpe and Richard Levitte] 4274 4275 *) Add Makefile.shared, a helper makefile to build shared 4276 libraries. Addapt Makefile.org. 4277 [Richard Levitte] 4278 4279 *) Add version info to Win32 DLLs. 4280 [Peter 'Luna' Runestig" <peter@runestig.com>] 4281 4282 *) Add new 'medium level' PKCS#12 API. Certificates and keys 4283 can be added using this API to created arbitrary PKCS#12 4284 files while avoiding the low level API. 4285 4286 New options to PKCS12_create(), key or cert can be NULL and 4287 will then be omitted from the output file. The encryption 4288 algorithm NIDs can be set to -1 for no encryption, the mac 4289 iteration count can be set to 0 to omit the mac. 4290 4291 Enhance pkcs12 utility by making the -nokeys and -nocerts 4292 options work when creating a PKCS#12 file. New option -nomac 4293 to omit the mac, NONE can be set for an encryption algorithm. 4294 New code is modified to use the enhanced PKCS12_create() 4295 instead of the low level API. 4296 [Steve Henson] 4297 4298 *) Extend ASN1 encoder to support indefinite length constructed 4299 encoding. This can output sequences tags and octet strings in 4300 this form. Modify pk7_asn1.c to support indefinite length 4301 encoding. This is experimental and needs additional code to 4302 be useful, such as an ASN1 bio and some enhanced streaming 4303 PKCS#7 code. 4304 4305 Extend template encode functionality so that tagging is passed 4306 down to the template encoder. 4307 [Steve Henson] 4308 4309 *) Let 'openssl req' fail if an argument to '-newkey' is not 4310 recognized instead of using RSA as a default. 4311 [Bodo Moeller] 4312 4313 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 4314 As these are not official, they are not included in "ALL"; 4315 the "ECCdraft" ciphersuite group alias can be used to select them. 4316 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 4317 4318 *) Add ECDH engine support. 4319 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 4320 4321 *) Add ECDH in new directory crypto/ecdh/. 4322 [Douglas Stebila (Sun Microsystems Laboratories)] 4323 4324 *) Let BN_rand_range() abort with an error after 100 iterations 4325 without success (which indicates a broken PRNG). 4326 [Bodo Moeller] 4327 4328 *) Change BN_mod_sqrt() so that it verifies that the input value 4329 is really the square of the return value. (Previously, 4330 BN_mod_sqrt would show GIGO behaviour.) 4331 [Bodo Moeller] 4332 4333 *) Add named elliptic curves over binary fields from X9.62, SECG, 4334 and WAP/WTLS; add OIDs that were still missing. 4335 4336 [Sheueling Chang Shantz and Douglas Stebila 4337 (Sun Microsystems Laboratories)] 4338 4339 *) Extend the EC library for elliptic curves over binary fields 4340 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 4341 New EC_METHOD: 4342 4343 EC_GF2m_simple_method 4344 4345 New API functions: 4346 4347 EC_GROUP_new_curve_GF2m 4348 EC_GROUP_set_curve_GF2m 4349 EC_GROUP_get_curve_GF2m 4350 EC_POINT_set_affine_coordinates_GF2m 4351 EC_POINT_get_affine_coordinates_GF2m 4352 EC_POINT_set_compressed_coordinates_GF2m 4353 4354 Point compression for binary fields is disabled by default for 4355 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 4356 enable it). 4357 4358 As binary polynomials are represented as BIGNUMs, various members 4359 of the EC_GROUP and EC_POINT data structures can be shared 4360 between the implementations for prime fields and binary fields; 4361 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 4362 are essentially identical to their ..._GFp counterparts. 4363 (For simplicity, the '..._GFp' prefix has been dropped from 4364 various internal method names.) 4365 4366 An internal 'field_div' method (similar to 'field_mul' and 4367 'field_sqr') has been added; this is used only for binary fields. 4368 4369 [Sheueling Chang Shantz and Douglas Stebila 4370 (Sun Microsystems Laboratories)] 4371 4372 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 4373 through methods ('mul', 'precompute_mult'). 4374 4375 The generic implementations (now internally called 'ec_wNAF_mul' 4376 and 'ec_wNAF_precomputed_mult') remain the default if these 4377 methods are undefined. 4378 4379 [Sheueling Chang Shantz and Douglas Stebila 4380 (Sun Microsystems Laboratories)] 4381 4382 *) New function EC_GROUP_get_degree, which is defined through 4383 EC_METHOD. For curves over prime fields, this returns the bit 4384 length of the modulus. 4385 4386 [Sheueling Chang Shantz and Douglas Stebila 4387 (Sun Microsystems Laboratories)] 4388 4389 *) New functions EC_GROUP_dup, EC_POINT_dup. 4390 (These simply call ..._new and ..._copy). 4391 4392 [Sheueling Chang Shantz and Douglas Stebila 4393 (Sun Microsystems Laboratories)] 4394 4395 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 4396 Polynomials are represented as BIGNUMs (where the sign bit is not 4397 used) in the following functions [macros]: 4398 4399 BN_GF2m_add 4400 BN_GF2m_sub [= BN_GF2m_add] 4401 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 4402 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 4403 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 4404 BN_GF2m_mod_inv 4405 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 4406 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 4407 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 4408 BN_GF2m_cmp [= BN_ucmp] 4409 4410 (Note that only the 'mod' functions are actually for fields GF(2^m). 4411 BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 4412 4413 For some functions, an the irreducible polynomial defining a 4414 field can be given as an 'unsigned int[]' with strictly 4415 decreasing elements giving the indices of those bits that are set; 4416 i.e., p[] represents the polynomial 4417 f(t) = t^p[0] + t^p[1] + ... + t^p[k] 4418 where 4419 p[0] > p[1] > ... > p[k] = 0. 4420 This applies to the following functions: 4421 4422 BN_GF2m_mod_arr 4423 BN_GF2m_mod_mul_arr 4424 BN_GF2m_mod_sqr_arr 4425 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 4426 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 4427 BN_GF2m_mod_exp_arr 4428 BN_GF2m_mod_sqrt_arr 4429 BN_GF2m_mod_solve_quad_arr 4430 BN_GF2m_poly2arr 4431 BN_GF2m_arr2poly 4432 4433 Conversion can be performed by the following functions: 4434 4435 BN_GF2m_poly2arr 4436 BN_GF2m_arr2poly 4437 4438 bntest.c has additional tests for binary polynomial arithmetic. 4439 4440 Two implementations for BN_GF2m_mod_div() are available. 4441 The default algorithm simply uses BN_GF2m_mod_inv() and 4442 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 4443 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 4444 copyright notice in crypto/bn/bn_gf2m.c before enabling it). 4445 4446 [Sheueling Chang Shantz and Douglas Stebila 4447 (Sun Microsystems Laboratories)] 4448 4449 *) Add new error code 'ERR_R_DISABLED' that can be used when some 4450 functionality is disabled at compile-time. 4451 [Douglas Stebila <douglas.stebila@sun.com>] 4452 4453 *) Change default behaviour of 'openssl asn1parse' so that more 4454 information is visible when viewing, e.g., a certificate: 4455 4456 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 4457 mode the content of non-printable OCTET STRINGs is output in a 4458 style similar to INTEGERs, but with '[HEX DUMP]' prepended to 4459 avoid the appearance of a printable string. 4460 [Nils Larsch <nla@trustcenter.de>] 4461 4462 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 4463 functions 4464 EC_GROUP_set_asn1_flag() 4465 EC_GROUP_get_asn1_flag() 4466 EC_GROUP_set_point_conversion_form() 4467 EC_GROUP_get_point_conversion_form() 4468 These control ASN1 encoding details: 4469 - Curves (i.e., groups) are encoded explicitly unless asn1_flag 4470 has been set to OPENSSL_EC_NAMED_CURVE. 4471 - Points are encoded in uncompressed form by default; options for 4472 asn1_for are as for point2oct, namely 4473 POINT_CONVERSION_COMPRESSED 4474 POINT_CONVERSION_UNCOMPRESSED 4475 POINT_CONVERSION_HYBRID 4476 4477 Also add 'seed' and 'seed_len' members to EC_GROUP with access 4478 functions 4479 EC_GROUP_set_seed() 4480 EC_GROUP_get0_seed() 4481 EC_GROUP_get_seed_len() 4482 This is used only for ASN1 purposes (so far). 4483 [Nils Larsch <nla@trustcenter.de>] 4484 4485 *) Add 'field_type' member to EC_METHOD, which holds the NID 4486 of the appropriate field type OID. The new function 4487 EC_METHOD_get_field_type() returns this value. 4488 [Nils Larsch <nla@trustcenter.de>] 4489 4490 *) Add functions 4491 EC_POINT_point2bn() 4492 EC_POINT_bn2point() 4493 EC_POINT_point2hex() 4494 EC_POINT_hex2point() 4495 providing useful interfaces to EC_POINT_point2oct() and 4496 EC_POINT_oct2point(). 4497 [Nils Larsch <nla@trustcenter.de>] 4498 4499 *) Change internals of the EC library so that the functions 4500 EC_GROUP_set_generator() 4501 EC_GROUP_get_generator() 4502 EC_GROUP_get_order() 4503 EC_GROUP_get_cofactor() 4504 are implemented directly in crypto/ec/ec_lib.c and not dispatched 4505 to methods, which would lead to unnecessary code duplication when 4506 adding different types of curves. 4507 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 4508 4509 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 4510 arithmetic, and such that modified wNAFs are generated 4511 (which avoid length expansion in many cases). 4512 [Bodo Moeller] 4513 4514 *) Add a function EC_GROUP_check_discriminant() (defined via 4515 EC_METHOD) that verifies that the curve discriminant is non-zero. 4516 4517 Add a function EC_GROUP_check() that makes some sanity tests 4518 on a EC_GROUP, its generator and order. This includes 4519 EC_GROUP_check_discriminant(). 4520 [Nils Larsch <nla@trustcenter.de>] 4521 4522 *) Add ECDSA in new directory crypto/ecdsa/. 4523 4524 Add applications 'openssl ecparam' and 'openssl ecdsa' 4525 (these are based on 'openssl dsaparam' and 'openssl dsa'). 4526 4527 ECDSA support is also included in various other files across the 4528 library. Most notably, 4529 - 'openssl req' now has a '-newkey ecdsa:file' option; 4530 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 4531 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 4532 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 4533 them suitable for ECDSA where domain parameters must be 4534 extracted before the specific public key; 4535 - ECDSA engine support has been added. 4536 [Nils Larsch <nla@trustcenter.de>] 4537 4538 *) Include some named elliptic curves, and add OIDs from X9.62, 4539 SECG, and WAP/WTLS. Each curve can be obtained from the new 4540 function 4541 EC_GROUP_new_by_curve_name(), 4542 and the list of available named curves can be obtained with 4543 EC_get_builtin_curves(). 4544 Also add a 'curve_name' member to EC_GROUP objects, which can be 4545 accessed via 4546 EC_GROUP_set_curve_name() 4547 EC_GROUP_get_curve_name() 4548 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 4549 4550 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 4551 was actually never needed) and in BN_mul(). The removal in BN_mul() 4552 required a small change in bn_mul_part_recursive() and the addition 4553 of the functions bn_cmp_part_words(), bn_sub_part_words() and 4554 bn_add_part_words(), which do the same thing as bn_cmp_words(), 4555 bn_sub_words() and bn_add_words() except they take arrays with 4556 differing sizes. 4557 [Richard Levitte] 4558 4559 Changes between 0.9.7l and 0.9.7m [23 Feb 2007] 4560 4561 *) Cleanse PEM buffers before freeing them since they may contain 4562 sensitive data. 4563 [Benjamin Bennett <ben@psc.edu>] 4564 4565 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 4566 a ciphersuite string such as "DEFAULT:RSA" cannot enable 4567 authentication-only ciphersuites. 4568 [Bodo Moeller] 4569 4570 *) Since AES128 and AES256 share a single mask bit in the logic of 4571 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 4572 kludge to work properly if AES128 is available and AES256 isn't. 4573 [Victor Duchovni] 4574 4575 *) Expand security boundary to match 1.1.1 module. 4576 [Steve Henson] 4577 4578 *) Remove redundant features: hash file source, editing of test vectors 4579 modify fipsld to use external fips_premain.c signature. 4580 [Steve Henson] 4581 4582 *) New perl script mkfipsscr.pl to create shell scripts or batch files to 4583 run algorithm test programs. 4584 [Steve Henson] 4585 4586 *) Make algorithm test programs more tolerant of whitespace. 4587 [Steve Henson] 4588 4589 *) Have SSL/TLS server implementation tolerate "mismatched" record 4590 protocol version while receiving ClientHello even if the 4591 ClientHello is fragmented. (The server can't insist on the 4592 particular protocol version it has chosen before the ServerHello 4593 message has informed the client about his choice.) 4594 [Bodo Moeller] 4595 4596 *) Load error codes if they are not already present instead of using a 4597 static variable. This allows them to be cleanly unloaded and reloaded. 4598 [Steve Henson] 4599 4600 Changes between 0.9.7k and 0.9.7l [28 Sep 2006] 4601 4602 *) Introduce limits to prevent malicious keys being able to 4603 cause a denial of service. (CVE-2006-2940) 4604 [Steve Henson, Bodo Moeller] 4605 4606 *) Fix ASN.1 parsing of certain invalid structures that can result 4607 in a denial of service. (CVE-2006-2937) [Steve Henson] 4608 4609 *) Fix buffer overflow in SSL_get_shared_ciphers() function. 4610 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 4611 4612 *) Fix SSL client code which could crash if connecting to a 4613 malicious SSLv2 server. (CVE-2006-4343) 4614 [Tavis Ormandy and Will Drewry, Google Security Team] 4615 4616 *) Change ciphersuite string processing so that an explicit 4617 ciphersuite selects this one ciphersuite (so that "AES256-SHA" 4618 will no longer include "AES128-SHA"), and any other similar 4619 ciphersuite (same bitmap) from *other* protocol versions (so that 4620 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 4621 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 4622 changes from 0.9.8b and 0.9.8d. 4623 [Bodo Moeller] 4624 4625 Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 4626 4627 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 4628 (CVE-2006-4339) [Ben Laurie and Google Security Team] 4629 4630 *) Change the Unix randomness entropy gathering to use poll() when 4631 possible instead of select(), since the latter has some 4632 undesirable limitations. 4633 [Darryl Miles via Richard Levitte and Bodo Moeller] 4634 4635 *) Disable rogue ciphersuites: 4636 4637 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 4638 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 4639 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 4640 4641 The latter two were purportedly from 4642 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 4643 appear there. 4644 4645 Also deactive the remaining ciphersuites from 4646 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 4647 unofficial, and the ID has long expired. 4648 [Bodo Moeller] 4649 4650 *) Fix RSA blinding Heisenbug (problems sometimes occured on 4651 dual-core machines) and other potential thread-safety issues. 4652 [Bodo Moeller] 4653 4654 Changes between 0.9.7i and 0.9.7j [04 May 2006] 4655 4656 *) Adapt fipsld and the build system to link against the validated FIPS 4657 module in FIPS mode. 4658 [Steve Henson] 4659 4660 *) Fixes for VC++ 2005 build under Windows. 4661 [Steve Henson] 4662 4663 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 4664 from a Windows bash shell such as MSYS. It is autodetected from the 4665 "config" script when run from a VC++ environment. Modify standard VC++ 4666 build to use fipscanister.o from the GNU make build. 4667 [Steve Henson] 4668 4669 Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 4670 4671 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 4672 The value now differs depending on if you build for FIPS or not. 4673 BEWARE! A program linked with a shared FIPSed libcrypto can't be 4674 safely run with a non-FIPSed libcrypto, as it may crash because of 4675 the difference induced by this change. 4676 [Andy Polyakov] 4677 4678 Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 4679 4680 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 4681 (part of SSL_OP_ALL). This option used to disable the 4682 countermeasure against man-in-the-middle protocol-version 4683 rollback in the SSL 2.0 server implementation, which is a bad 4684 idea. (CVE-2005-2969) 4685 4686 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 4687 for Information Security, National Institute of Advanced Industrial 4688 Science and Technology [AIST], Japan)] 4689 4690 *) Minimal support for X9.31 signatures and PSS padding modes. This is 4691 mainly for FIPS compliance and not fully integrated at this stage. 4692 [Steve Henson] 4693 4694 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 4695 the exponentiation using a fixed-length exponent. (Otherwise, 4696 the information leaked through timing could expose the secret key 4697 after many signatures; cf. Bleichenbacher's attack on DSA with 4698 biased k.) 4699 [Bodo Moeller] 4700 4701 *) Make a new fixed-window mod_exp implementation the default for 4702 RSA, DSA, and DH private-key operations so that the sequence of 4703 squares and multiplies and the memory access pattern are 4704 independent of the particular secret key. This will mitigate 4705 cache-timing and potential related attacks. 4706 4707 BN_mod_exp_mont_consttime() is the new exponentiation implementation, 4708 and this is automatically used by BN_mod_exp_mont() if the new flag 4709 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 4710 will use this BN flag for private exponents unless the flag 4711 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 4712 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 4713 4714 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 4715 4716 *) Change the client implementation for SSLv23_method() and 4717 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 4718 Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 4719 (Previously, the SSL 2.0 backwards compatible Client Hello 4720 message format would be used even with SSL_OP_NO_SSLv2.) 4721 [Bodo Moeller] 4722 4723 *) Add support for smime-type MIME parameter in S/MIME messages which some 4724 clients need. 4725 [Steve Henson] 4726 4727 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 4728 a threadsafe manner. Modify rsa code to use new function and add calls 4729 to dsa and dh code (which had race conditions before). 4730 [Steve Henson] 4731 4732 *) Include the fixed error library code in the C error file definitions 4733 instead of fixing them up at runtime. This keeps the error code 4734 structures constant. 4735 [Steve Henson] 4736 4737 Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 4738 4739 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 4740 OpenSSL 0.9.8.] 4741 4742 *) Fixes for newer kerberos headers. NB: the casts are needed because 4743 the 'length' field is signed on one version and unsigned on another 4744 with no (?) obvious way to tell the difference, without these VC++ 4745 complains. Also the "definition" of FAR (blank) is no longer included 4746 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 4747 some needed definitions. 4748 [Steve Henson] 4749 4750 *) Undo Cygwin change. 4751 [Ulf M��ller] 4752 4753 *) Added support for proxy certificates according to RFC 3820. 4754 Because they may be a security thread to unaware applications, 4755 they must be explicitely allowed in run-time. See 4756 docs/HOWTO/proxy_certificates.txt for further information. 4757 [Richard Levitte] 4758 4759 Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 4760 4761 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 4762 server and client random values. Previously 4763 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 4764 less random data when sizeof(time_t) > 4 (some 64 bit platforms). 4765 4766 This change has negligible security impact because: 4767 4768 1. Server and client random values still have 24 bytes of pseudo random 4769 data. 4770 4771 2. Server and client random values are sent in the clear in the initial 4772 handshake. 4773 4774 3. The master secret is derived using the premaster secret (48 bytes in 4775 size for static RSA ciphersuites) as well as client server and random 4776 values. 4777 4778 The OpenSSL team would like to thank the UK NISCC for bringing this issue 4779 to our attention. 4780 4781 [Stephen Henson, reported by UK NISCC] 4782 4783 *) Use Windows randomness collection on Cygwin. 4784 [Ulf M��ller] 4785 4786 *) Fix hang in EGD/PRNGD query when communication socket is closed 4787 prematurely by EGD/PRNGD. 4788 [Darren Tucker <dtucker@zip.com.au> via Lutz J��nicke, resolves #1014] 4789 4790 *) Prompt for pass phrases when appropriate for PKCS12 input format. 4791 [Steve Henson] 4792 4793 *) Back-port of selected performance improvements from development 4794 branch, as well as improved support for PowerPC platforms. 4795 [Andy Polyakov] 4796 4797 *) Add lots of checks for memory allocation failure, error codes to indicate 4798 failure and freeing up memory if a failure occurs. 4799 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 4800 4801 *) Add new -passin argument to dgst. 4802 [Steve Henson] 4803 4804 *) Perform some character comparisons of different types in X509_NAME_cmp: 4805 this is needed for some certificates that reencode DNs into UTF8Strings 4806 (in violation of RFC3280) and can't or wont issue name rollover 4807 certificates. 4808 [Steve Henson] 4809 4810 *) Make an explicit check during certificate validation to see that 4811 the CA setting in each certificate on the chain is correct. As a 4812 side effect always do the following basic checks on extensions, 4813 not just when there's an associated purpose to the check: 4814 4815 - if there is an unhandled critical extension (unless the user 4816 has chosen to ignore this fault) 4817 - if the path length has been exceeded (if one is set at all) 4818 - that certain extensions fit the associated purpose (if one has 4819 been given) 4820 [Richard Levitte] 4821 4822 Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 4823 4824 *) Avoid a race condition when CRLs are checked in a multi threaded 4825 environment. This would happen due to the reordering of the revoked 4826 entries during signature checking and serial number lookup. Now the 4827 encoding is cached and the serial number sort performed under a lock. 4828 Add new STACK function sk_is_sorted(). 4829 [Steve Henson] 4830 4831 *) Add Delta CRL to the extension code. 4832 [Steve Henson] 4833 4834 *) Various fixes to s3_pkt.c so alerts are sent properly. 4835 [David Holmes <d.holmes@f5.com>] 4836 4837 *) Reduce the chances of duplicate issuer name and serial numbers (in 4838 violation of RFC3280) using the OpenSSL certificate creation utilities. 4839 This is done by creating a random 64 bit value for the initial serial 4840 number when a serial number file is created or when a self signed 4841 certificate is created using 'openssl req -x509'. The initial serial 4842 number file is created using 'openssl x509 -next_serial' in CA.pl 4843 rather than being initialized to 1. 4844 [Steve Henson] 4845 4846 Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 4847 4848 *) Fix null-pointer assignment in do_change_cipher_spec() revealed 4849 by using the Codenomicon TLS Test Tool (CVE-2004-0079) 4850 [Joe Orton, Steve Henson] 4851 4852 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 4853 (CVE-2004-0112) 4854 [Joe Orton, Steve Henson] 4855 4856 *) Make it possible to have multiple active certificates with the same 4857 subject in the CA index file. This is done only if the keyword 4858 'unique_subject' is set to 'no' in the main CA section (default 4859 if 'CA_default') of the configuration file. The value is saved 4860 with the database itself in a separate index attribute file, 4861 named like the index file with '.attr' appended to the name. 4862 [Richard Levitte] 4863 4864 *) X509 verify fixes. Disable broken certificate workarounds when 4865 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 4866 keyUsage extension present. Don't accept CRLs with unhandled critical 4867 extensions: since verify currently doesn't process CRL extensions this 4868 rejects a CRL with *any* critical extensions. Add new verify error codes 4869 for these cases. 4870 [Steve Henson] 4871 4872 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 4873 A clarification of RFC2560 will require the use of OCTET STRINGs and 4874 some implementations cannot handle the current raw format. Since OpenSSL 4875 copies and compares OCSP nonces as opaque blobs without any attempt at 4876 parsing them this should not create any compatibility issues. 4877 [Steve Henson] 4878 4879 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 4880 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 4881 this HMAC (and other) operations are several times slower than OpenSSL 4882 < 0.9.7. 4883 [Steve Henson] 4884 4885 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 4886 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 4887 4888 *) Use the correct content when signing type "other". 4889 [Steve Henson] 4890 4891 Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 4892 4893 *) Fix various bugs revealed by running the NISCC test suite: 4894 4895 Stop out of bounds reads in the ASN1 code when presented with 4896 invalid tags (CVE-2003-0543 and CVE-2003-0544). 4897 4898 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 4899 4900 If verify callback ignores invalid public key errors don't try to check 4901 certificate signature with the NULL public key. 4902 4903 [Steve Henson] 4904 4905 *) New -ignore_err option in ocsp application to stop the server 4906 exiting on the first error in a request. 4907 [Steve Henson] 4908 4909 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 4910 if the server requested one: as stated in TLS 1.0 and SSL 3.0 4911 specifications. 4912 [Steve Henson] 4913 4914 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 4915 extra data after the compression methods not only for TLS 1.0 4916 but also for SSL 3.0 (as required by the specification). 4917 [Bodo Moeller; problem pointed out by Matthias Loepfe] 4918 4919 *) Change X509_certificate_type() to mark the key as exported/exportable 4920 when it's 512 *bits* long, not 512 bytes. 4921 [Richard Levitte] 4922 4923 *) Change AES_cbc_encrypt() so it outputs exact multiple of 4924 blocks during encryption. 4925 [Richard Levitte] 4926 4927 *) Various fixes to base64 BIO and non blocking I/O. On write 4928 flushes were not handled properly if the BIO retried. On read 4929 data was not being buffered properly and had various logic bugs. 4930 This also affects blocking I/O when the data being decoded is a 4931 certain size. 4932 [Steve Henson] 4933 4934 *) Various S/MIME bugfixes and compatibility changes: 4935 output correct application/pkcs7 MIME type if 4936 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 4937 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 4938 of files as .eml work). Correctly handle very long lines in MIME 4939 parser. 4940 [Steve Henson] 4941 4942 Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 4943 4944 *) Countermeasure against the Klima-Pokorny-Rosa extension of 4945 Bleichbacher's attack on PKCS #1 v1.5 padding: treat 4946 a protocol version number mismatch like a decryption error 4947 in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 4948 [Bodo Moeller] 4949 4950 *) Turn on RSA blinding by default in the default implementation 4951 to avoid a timing attack. Applications that don't want it can call 4952 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 4953 They would be ill-advised to do so in most cases. 4954 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 4955 4956 *) Change RSA blinding code so that it works when the PRNG is not 4957 seeded (in this case, the secret RSA exponent is abused as 4958 an unpredictable seed -- if it is not unpredictable, there 4959 is no point in blinding anyway). Make RSA blinding thread-safe 4960 by remembering the creator's thread ID in rsa->blinding and 4961 having all other threads use local one-time blinding factors 4962 (this requires more computation than sharing rsa->blinding, but 4963 avoids excessive locking; and if an RSA object is not shared 4964 between threads, blinding will still be very fast). 4965 [Bodo Moeller] 4966 4967 *) Fixed a typo bug that would cause ENGINE_set_default() to set an 4968 ENGINE as defaults for all supported algorithms irrespective of 4969 the 'flags' parameter. 'flags' is now honoured, so applications 4970 should make sure they are passing it correctly. 4971 [Geoff Thorpe] 4972 4973 *) Target "mingw" now allows native Windows code to be generated in 4974 the Cygwin environment as well as with the MinGW compiler. 4975 [Ulf Moeller] 4976 4977 Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 4978 4979 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 4980 via timing by performing a MAC computation even if incorrrect 4981 block cipher padding has been found. This is a countermeasure 4982 against active attacks where the attacker has to distinguish 4983 between bad padding and a MAC verification error. (CVE-2003-0078) 4984 4985 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 4986 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 4987 Martin Vuagnoux (EPFL, Ilion)] 4988 4989 *) Make the no-err option work as intended. The intention with no-err 4990 is not to have the whole error stack handling routines removed from 4991 libcrypto, it's only intended to remove all the function name and 4992 reason texts, thereby removing some of the footprint that may not 4993 be interesting if those errors aren't displayed anyway. 4994 4995 NOTE: it's still possible for any application or module to have it's 4996 own set of error texts inserted. The routines are there, just not 4997 used by default when no-err is given. 4998 [Richard Levitte] 4999 5000 *) Add support for FreeBSD on IA64. 5001 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 5002 5003 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 5004 Kerberos function mit_des_cbc_cksum(). Before this change, 5005 the value returned by DES_cbc_cksum() was like the one from 5006 mit_des_cbc_cksum(), except the bytes were swapped. 5007 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 5008 5009 *) Allow an application to disable the automatic SSL chain building. 5010 Before this a rather primitive chain build was always performed in 5011 ssl3_output_cert_chain(): an application had no way to send the 5012 correct chain if the automatic operation produced an incorrect result. 5013 5014 Now the chain builder is disabled if either: 5015 5016 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 5017 5018 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 5019 5020 The reasoning behind this is that an application would not want the 5021 auto chain building to take place if extra chain certificates are 5022 present and it might also want a means of sending no additional 5023 certificates (for example the chain has two certificates and the 5024 root is omitted). 5025 [Steve Henson] 5026 5027 *) Add the possibility to build without the ENGINE framework. 5028 [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 5029 5030 *) Under Win32 gmtime() can return NULL: check return value in 5031 OPENSSL_gmtime(). Add error code for case where gmtime() fails. 5032 [Steve Henson] 5033 5034 *) DSA routines: under certain error conditions uninitialized BN objects 5035 could be freed. Solution: make sure initialization is performed early 5036 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 5037 Nils Larsch <nla@trustcenter.de> via PR#459) 5038 [Lutz Jaenicke] 5039 5040 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 5041 checked on reconnect on the client side, therefore session resumption 5042 could still fail with a "ssl session id is different" error. This 5043 behaviour is masked when SSL_OP_ALL is used due to 5044 SSL_OP_MICROSOFT_SESS_ID_BUG being set. 5045 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 5046 followup to PR #377. 5047 [Lutz Jaenicke] 5048 5049 *) IA-32 assembler support enhancements: unified ELF targets, support 5050 for SCO/Caldera platforms, fix for Cygwin shared build. 5051 [Andy Polyakov] 5052 5053 *) Add support for FreeBSD on sparc64. As a consequence, support for 5054 FreeBSD on non-x86 processors is separate from x86 processors on 5055 the config script, much like the NetBSD support. 5056 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 5057 5058 Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 5059 5060 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 5061 OpenSSL 0.9.7.] 5062 5063 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 5064 code (06) was taken as the first octet of the session ID and the last 5065 octet was ignored consequently. As a result SSLv2 client side session 5066 caching could not have worked due to the session ID mismatch between 5067 client and server. 5068 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 5069 PR #377. 5070 [Lutz Jaenicke] 5071 5072 *) Change the declaration of needed Kerberos libraries to use EX_LIBS 5073 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 5074 removed entirely. 5075 [Richard Levitte] 5076 5077 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 5078 seems that in spite of existing for more than a year, many application 5079 author have done nothing to provide the necessary callbacks, which 5080 means that this particular engine will not work properly anywhere. 5081 This is a very unfortunate situation which forces us, in the name 5082 of usability, to give the hw_ncipher.c a static lock, which is part 5083 of libcrypto. 5084 NOTE: This is for the 0.9.7 series ONLY. This hack will never 5085 appear in 0.9.8 or later. We EXPECT application authors to have 5086 dealt properly with this when 0.9.8 is released (unless we actually 5087 make such changes in the libcrypto locking code that changes will 5088 have to be made anyway). 5089 [Richard Levitte] 5090 5091 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 5092 octets have been read, EOF or an error occurs. Without this change 5093 some truncated ASN1 structures will not produce an error. 5094 [Steve Henson] 5095 5096 *) Disable Heimdal support, since it hasn't been fully implemented. 5097 Still give the possibility to force the use of Heimdal, but with 5098 warnings and a request that patches get sent to openssl-dev. 5099 [Richard Levitte] 5100 5101 *) Add the VC-CE target, introduce the WINCE sysname, and add 5102 INSTALL.WCE and appropriate conditionals to make it build. 5103 [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 5104 5105 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 5106 cygssl-x.y.z.dll, where x, y and z are the major, minor and 5107 edit numbers of the version. 5108 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 5109 5110 *) Introduce safe string copy and catenation functions 5111 (BUF_strlcpy() and BUF_strlcat()). 5112 [Ben Laurie (CHATS) and Richard Levitte] 5113 5114 *) Avoid using fixed-size buffers for one-line DNs. 5115 [Ben Laurie (CHATS)] 5116 5117 *) Add BUF_MEM_grow_clean() to avoid information leakage when 5118 resizing buffers containing secrets, and use where appropriate. 5119 [Ben Laurie (CHATS)] 5120 5121 *) Avoid using fixed size buffers for configuration file location. 5122 [Ben Laurie (CHATS)] 5123 5124 *) Avoid filename truncation for various CA files. 5125 [Ben Laurie (CHATS)] 5126 5127 *) Use sizeof in preference to magic numbers. 5128 [Ben Laurie (CHATS)] 5129 5130 *) Avoid filename truncation in cert requests. 5131 [Ben Laurie (CHATS)] 5132 5133 *) Add assertions to check for (supposedly impossible) buffer 5134 overflows. 5135 [Ben Laurie (CHATS)] 5136 5137 *) Don't cache truncated DNS entries in the local cache (this could 5138 potentially lead to a spoofing attack). 5139 [Ben Laurie (CHATS)] 5140 5141 *) Fix various buffers to be large enough for hex/decimal 5142 representations in a platform independent manner. 5143 [Ben Laurie (CHATS)] 5144 5145 *) Add CRYPTO_realloc_clean() to avoid information leakage when 5146 resizing buffers containing secrets, and use where appropriate. 5147 [Ben Laurie (CHATS)] 5148 5149 *) Add BIO_indent() to avoid much slightly worrying code to do 5150 indents. 5151 [Ben Laurie (CHATS)] 5152 5153 *) Convert sprintf()/BIO_puts() to BIO_printf(). 5154 [Ben Laurie (CHATS)] 5155 5156 *) buffer_gets() could terminate with the buffer only half 5157 full. Fixed. 5158 [Ben Laurie (CHATS)] 5159 5160 *) Add assertions to prevent user-supplied crypto functions from 5161 overflowing internal buffers by having large block sizes, etc. 5162 [Ben Laurie (CHATS)] 5163 5164 *) New OPENSSL_assert() macro (similar to assert(), but enabled 5165 unconditionally). 5166 [Ben Laurie (CHATS)] 5167 5168 *) Eliminate unused copy of key in RC4. 5169 [Ben Laurie (CHATS)] 5170 5171 *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 5172 [Ben Laurie (CHATS)] 5173 5174 *) Fix off-by-one error in EGD path. 5175 [Ben Laurie (CHATS)] 5176 5177 *) If RANDFILE path is too long, ignore instead of truncating. 5178 [Ben Laurie (CHATS)] 5179 5180 *) Eliminate unused and incorrectly sized X.509 structure 5181 CBCParameter. 5182 [Ben Laurie (CHATS)] 5183 5184 *) Eliminate unused and dangerous function knumber(). 5185 [Ben Laurie (CHATS)] 5186 5187 *) Eliminate unused and dangerous structure, KSSL_ERR. 5188 [Ben Laurie (CHATS)] 5189 5190 *) Protect against overlong session ID context length in an encoded 5191 session object. Since these are local, this does not appear to be 5192 exploitable. 5193 [Ben Laurie (CHATS)] 5194 5195 *) Change from security patch (see 0.9.6e below) that did not affect 5196 the 0.9.6 release series: 5197 5198 Remote buffer overflow in SSL3 protocol - an attacker could 5199 supply an oversized master key in Kerberos-enabled versions. 5200 (CVE-2002-0657) 5201 [Ben Laurie (CHATS)] 5202 5203 *) Change the SSL kerb5 codes to match RFC 2712. 5204 [Richard Levitte] 5205 5206 *) Make -nameopt work fully for req and add -reqopt switch. 5207 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 5208 5209 *) The "block size" for block ciphers in CFB and OFB mode should be 1. 5210 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 5211 5212 *) Make sure tests can be performed even if the corresponding algorithms 5213 have been removed entirely. This was also the last step to make 5214 OpenSSL compilable with DJGPP under all reasonable conditions. 5215 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 5216 5217 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 5218 to allow version independent disabling of normally unselected ciphers, 5219 which may be activated as a side-effect of selecting a single cipher. 5220 5221 (E.g., cipher list string "RSA" enables ciphersuites that are left 5222 out of "ALL" because they do not provide symmetric encryption. 5223 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 5224 [Lutz Jaenicke, Bodo Moeller] 5225 5226 *) Add appropriate support for separate platform-dependent build 5227 directories. The recommended way to make a platform-dependent 5228 build directory is the following (tested on Linux), maybe with 5229 some local tweaks: 5230 5231 # Place yourself outside of the OpenSSL source tree. In 5232 # this example, the environment variable OPENSSL_SOURCE 5233 # is assumed to contain the absolute OpenSSL source directory. 5234 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 5235 cd objtree/"`uname -s`-`uname -r`-`uname -m`" 5236 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 5237 mkdir -p `dirname $F` 5238 ln -s $OPENSSL_SOURCE/$F $F 5239 done 5240 5241 To be absolutely sure not to disturb the source tree, a "make clean" 5242 is a good thing. If it isn't successfull, don't worry about it, 5243 it probably means the source directory is very clean. 5244 [Richard Levitte] 5245 5246 *) Make sure any ENGINE control commands make local copies of string 5247 pointers passed to them whenever necessary. Otherwise it is possible 5248 the caller may have overwritten (or deallocated) the original string 5249 data when a later ENGINE operation tries to use the stored values. 5250 [G��tz Babin-Ebell <babinebell@trustcenter.de>] 5251 5252 *) Improve diagnostics in file reading and command-line digests. 5253 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 5254 5255 *) Add AES modes CFB and OFB to the object database. Correct an 5256 error in AES-CFB decryption. 5257 [Richard Levitte] 5258 5259 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 5260 allows existing EVP_CIPHER_CTX structures to be reused after 5261 calling EVP_*Final(). This behaviour is used by encryption 5262 BIOs and some applications. This has the side effect that 5263 applications must explicitly clean up cipher contexts with 5264 EVP_CIPHER_CTX_cleanup() or they will leak memory. 5265 [Steve Henson] 5266 5267 *) Check the values of dna and dnb in bn_mul_recursive before calling 5268 bn_mul_comba (a non zero value means the a or b arrays do not contain 5269 n2 elements) and fallback to bn_mul_normal if either is not zero. 5270 [Steve Henson] 5271 5272 *) Fix escaping of non-ASCII characters when using the -subj option 5273 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 5274 [Lutz Jaenicke] 5275 5276 *) Make object definitions compliant to LDAP (RFC2256): SN is the short 5277 form for "surname", serialNumber has no short form. 5278 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 5279 therefore remove "mail" short name for "internet 7". 5280 The OID for unique identifiers in X509 certificates is 5281 x500UniqueIdentifier, not uniqueIdentifier. 5282 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 5283 [Lutz Jaenicke] 5284 5285 *) Add an "init" command to the ENGINE config module and auto initialize 5286 ENGINEs. Without any "init" command the ENGINE will be initialized 5287 after all ctrl commands have been executed on it. If init=1 the 5288 ENGINE is initailized at that point (ctrls before that point are run 5289 on the uninitialized ENGINE and after on the initialized one). If 5290 init=0 then the ENGINE will not be iniatialized at all. 5291 [Steve Henson] 5292 5293 *) Fix the 'app_verify_callback' interface so that the user-defined 5294 argument is actually passed to the callback: In the 5295 SSL_CTX_set_cert_verify_callback() prototype, the callback 5296 declaration has been changed from 5297 int (*cb)() 5298 into 5299 int (*cb)(X509_STORE_CTX *,void *); 5300 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 5301 i=s->ctx->app_verify_callback(&ctx) 5302 has been changed into 5303 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 5304 5305 To update applications using SSL_CTX_set_cert_verify_callback(), 5306 a dummy argument can be added to their callback functions. 5307 [D. K. Smetters <smetters@parc.xerox.com>] 5308 5309 *) Added the '4758cca' ENGINE to support IBM 4758 cards. 5310 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 5311 5312 *) Add and OPENSSL_LOAD_CONF define which will cause 5313 OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 5314 This allows older applications to transparently support certain 5315 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 5316 Two new functions OPENSSL_add_all_algorithms_noconf() which will never 5317 load the config file and OPENSSL_add_all_algorithms_conf() which will 5318 always load it have also been added. 5319 [Steve Henson] 5320 5321 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 5322 Adjust NIDs and EVP layer. 5323 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 5324 5325 *) Config modules support in openssl utility. 5326 5327 Most commands now load modules from the config file, 5328 though in a few (such as version) this isn't done 5329 because it couldn't be used for anything. 5330 5331 In the case of ca and req the config file used is 5332 the same as the utility itself: that is the -config 5333 command line option can be used to specify an 5334 alternative file. 5335 [Steve Henson] 5336 5337 *) Move default behaviour from OPENSSL_config(). If appname is NULL 5338 use "openssl_conf" if filename is NULL use default openssl config file. 5339 [Steve Henson] 5340 5341 *) Add an argument to OPENSSL_config() to allow the use of an alternative 5342 config section name. Add a new flag to tolerate a missing config file 5343 and move code to CONF_modules_load_file(). 5344 [Steve Henson] 5345 5346 *) Support for crypto accelerator cards from Accelerated Encryption 5347 Processing, www.aep.ie. (Use engine 'aep') 5348 The support was copied from 0.9.6c [engine] and adapted/corrected 5349 to work with the new engine framework. 5350 [AEP Inc. and Richard Levitte] 5351 5352 *) Support for SureWare crypto accelerator cards from Baltimore 5353 Technologies. (Use engine 'sureware') 5354 The support was copied from 0.9.6c [engine] and adapted 5355 to work with the new engine framework. 5356 [Richard Levitte] 5357 5358 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 5359 make the newer ENGINE framework commands for the CHIL engine work. 5360 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 5361 5362 *) Make it possible to produce shared libraries on ReliantUNIX. 5363 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 5364 5365 *) Add the configuration target debug-linux-ppro. 5366 Make 'openssl rsa' use the general key loading routines 5367 implemented in apps.c, and make those routines able to 5368 handle the key format FORMAT_NETSCAPE and the variant 5369 FORMAT_IISSGC. 5370 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 5371 5372 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 5373 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 5374 5375 *) Add -keyform to rsautl, and document -engine. 5376 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 5377 5378 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 5379 BIO_R_NO_SUCH_FILE error code rather than the generic 5380 ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 5381 [Ben Laurie] 5382 5383 *) Add new functions 5384 ERR_peek_last_error 5385 ERR_peek_last_error_line 5386 ERR_peek_last_error_line_data. 5387 These are similar to 5388 ERR_peek_error 5389 ERR_peek_error_line 5390 ERR_peek_error_line_data, 5391 but report on the latest error recorded rather than the first one 5392 still in the error queue. 5393 [Ben Laurie, Bodo Moeller] 5394 5395 *) default_algorithms option in ENGINE config module. This allows things 5396 like: 5397 default_algorithms = ALL 5398 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 5399 [Steve Henson] 5400 5401 *) Prelminary ENGINE config module. 5402 [Steve Henson] 5403 5404 *) New experimental application configuration code. 5405 [Steve Henson] 5406 5407 *) Change the AES code to follow the same name structure as all other 5408 symmetric ciphers, and behave the same way. Move everything to 5409 the directory crypto/aes, thereby obsoleting crypto/rijndael. 5410 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 5411 5412 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 5413 [Ben Laurie and Theo de Raadt] 5414 5415 *) Add option to output public keys in req command. 5416 [Massimiliano Pala madwolf@openca.org] 5417 5418 *) Use wNAFs in EC_POINTs_mul() for improved efficiency 5419 (up to about 10% better than before for P-192 and P-224). 5420 [Bodo Moeller] 5421 5422 *) New functions/macros 5423 5424 SSL_CTX_set_msg_callback(ctx, cb) 5425 SSL_CTX_set_msg_callback_arg(ctx, arg) 5426 SSL_set_msg_callback(ssl, cb) 5427 SSL_set_msg_callback_arg(ssl, arg) 5428 5429 to request calling a callback function 5430 5431 void cb(int write_p, int version, int content_type, 5432 const void *buf, size_t len, SSL *ssl, void *arg) 5433 5434 whenever a protocol message has been completely received 5435 (write_p == 0) or sent (write_p == 1). Here 'version' is the 5436 protocol version according to which the SSL library interprets 5437 the current protocol message (SSL2_VERSION, SSL3_VERSION, or 5438 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 5439 the content type as defined in the SSL 3.0/TLS 1.0 protocol 5440 specification (change_cipher_spec(20), alert(21), handshake(22)). 5441 'buf' and 'len' point to the actual message, 'ssl' to the 5442 SSL object, and 'arg' is the application-defined value set by 5443 SSL[_CTX]_set_msg_callback_arg(). 5444 5445 'openssl s_client' and 'openssl s_server' have new '-msg' options 5446 to enable a callback that displays all protocol messages. 5447 [Bodo Moeller] 5448 5449 *) Change the shared library support so shared libraries are built as 5450 soon as the corresponding static library is finished, and thereby get 5451 openssl and the test programs linked against the shared library. 5452 This still only happens when the keyword "shard" has been given to 5453 the configuration scripts. 5454 5455 NOTE: shared library support is still an experimental thing, and 5456 backward binary compatibility is still not guaranteed. 5457 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 5458 5459 *) Add support for Subject Information Access extension. 5460 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 5461 5462 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 5463 additional bytes when new memory had to be allocated, not just 5464 when reusing an existing buffer. 5465 [Bodo Moeller] 5466 5467 *) New command line and configuration option 'utf8' for the req command. 5468 This allows field values to be specified as UTF8 strings. 5469 [Steve Henson] 5470 5471 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 5472 runs for the former and machine-readable output for the latter. 5473 [Ben Laurie] 5474 5475 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 5476 of the e-mail address in the DN (i.e., it will go into a certificate 5477 extension only). The new configuration file option 'email_in_dn = no' 5478 has the same effect. 5479 [Massimiliano Pala madwolf@openca.org] 5480 5481 *) Change all functions with names starting with des_ to be starting 5482 with DES_ instead. Add wrappers that are compatible with libdes, 5483 but are named _ossl_old_des_*. Finally, add macros that map the 5484 des_* symbols to the corresponding _ossl_old_des_* if libdes 5485 compatibility is desired. If OpenSSL 0.9.6c compatibility is 5486 desired, the des_* symbols will be mapped to DES_*, with one 5487 exception. 5488 5489 Since we provide two compatibility mappings, the user needs to 5490 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 5491 compatibility is desired. The default (i.e., when that macro 5492 isn't defined) is OpenSSL 0.9.6c compatibility. 5493 5494 There are also macros that enable and disable the support of old 5495 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 5496 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 5497 are defined, the default will apply: to support the old des routines. 5498 5499 In either case, one must include openssl/des.h to get the correct 5500 definitions. Do not try to just include openssl/des_old.h, that 5501 won't work. 5502 5503 NOTE: This is a major break of an old API into a new one. Software 5504 authors are encouraged to switch to the DES_ style functions. Some 5505 time in the future, des_old.h and the libdes compatibility functions 5506 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 5507 default), and then completely removed. 5508 [Richard Levitte] 5509 5510 *) Test for certificates which contain unsupported critical extensions. 5511 If such a certificate is found during a verify operation it is 5512 rejected by default: this behaviour can be overridden by either 5513 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 5514 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 5515 X509_supported_extension() has also been added which returns 1 if a 5516 particular extension is supported. 5517 [Steve Henson] 5518 5519 *) Modify the behaviour of EVP cipher functions in similar way to digests 5520 to retain compatibility with existing code. 5521 [Steve Henson] 5522 5523 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 5524 compatibility with existing code. In particular the 'ctx' parameter does 5525 not have to be to be initialized before the call to EVP_DigestInit() and 5526 it is tidied up after a call to EVP_DigestFinal(). New function 5527 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 5528 EVP_MD_CTX_copy() changed to not require the destination to be 5529 initialized valid and new function EVP_MD_CTX_copy_ex() added which 5530 requires the destination to be valid. 5531 5532 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 5533 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 5534 [Steve Henson] 5535 5536 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 5537 so that complete 'Handshake' protocol structures are kept in memory 5538 instead of overwriting 'msg_type' and 'length' with 'body' data. 5539 [Bodo Moeller] 5540 5541 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 5542 [Massimo Santin via Richard Levitte] 5543 5544 *) Major restructuring to the underlying ENGINE code. This includes 5545 reduction of linker bloat, separation of pure "ENGINE" manipulation 5546 (initialisation, etc) from functionality dealing with implementations 5547 of specific crypto iterfaces. This change also introduces integrated 5548 support for symmetric ciphers and digest implementations - so ENGINEs 5549 can now accelerate these by providing EVP_CIPHER and EVP_MD 5550 implementations of their own. This is detailed in crypto/engine/README 5551 as it couldn't be adequately described here. However, there are a few 5552 API changes worth noting - some RSA, DSA, DH, and RAND functions that 5553 were changed in the original introduction of ENGINE code have now 5554 reverted back - the hooking from this code to ENGINE is now a good 5555 deal more passive and at run-time, operations deal directly with 5556 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 5557 dereferencing through an ENGINE pointer any more. Also, the ENGINE 5558 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 5559 they were not being used by the framework as there is no concept of a 5560 BIGNUM_METHOD and they could not be generalised to the new 5561 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 5562 ENGINE_cpy() has been removed as it cannot be consistently defined in 5563 the new code. 5564 [Geoff Thorpe] 5565 5566 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 5567 [Steve Henson] 5568 5569 *) Change mkdef.pl to sort symbols that get the same entry number, 5570 and make sure the automatically generated functions ERR_load_* 5571 become part of libeay.num as well. 5572 [Richard Levitte] 5573 5574 *) New function SSL_renegotiate_pending(). This returns true once 5575 renegotiation has been requested (either SSL_renegotiate() call 5576 or HelloRequest/ClientHello receveived from the peer) and becomes 5577 false once a handshake has been completed. 5578 (For servers, SSL_renegotiate() followed by SSL_do_handshake() 5579 sends a HelloRequest, but does not ensure that a handshake takes 5580 place. SSL_renegotiate_pending() is useful for checking if the 5581 client has followed the request.) 5582 [Bodo Moeller] 5583 5584 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 5585 By default, clients may request session resumption even during 5586 renegotiation (if session ID contexts permit); with this option, 5587 session resumption is possible only in the first handshake. 5588 5589 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 5590 more bits available for options that should not be part of 5591 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 5592 [Bodo Moeller] 5593 5594 *) Add some demos for certificate and certificate request creation. 5595 [Steve Henson] 5596 5597 *) Make maximum certificate chain size accepted from the peer application 5598 settable (SSL*_get/set_max_cert_list()), as proposed by 5599 "Douglas E. Engert" <deengert@anl.gov>. 5600 [Lutz Jaenicke] 5601 5602 *) Add support for shared libraries for Unixware-7 5603 (Boyd Lynn Gerber <gerberb@zenez.com>). 5604 [Lutz Jaenicke] 5605 5606 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 5607 be done prior to destruction. Use this to unload error strings from 5608 ENGINEs that load their own error strings. NB: This adds two new API 5609 functions to "get" and "set" this destroy handler in an ENGINE. 5610 [Geoff Thorpe] 5611 5612 *) Alter all existing ENGINE implementations (except "openssl" and 5613 "openbsd") to dynamically instantiate their own error strings. This 5614 makes them more flexible to be built both as statically-linked ENGINEs 5615 and self-contained shared-libraries loadable via the "dynamic" ENGINE. 5616 Also, add stub code to each that makes building them as self-contained 5617 shared-libraries easier (see README.ENGINE). 5618 [Geoff Thorpe] 5619 5620 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 5621 implementations into applications that are completely implemented in 5622 self-contained shared-libraries. The "dynamic" ENGINE exposes control 5623 commands that can be used to configure what shared-library to load and 5624 to control aspects of the way it is handled. Also, made an update to 5625 the README.ENGINE file that brings its information up-to-date and 5626 provides some information and instructions on the "dynamic" ENGINE 5627 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 5628 [Geoff Thorpe] 5629 5630 *) Make it possible to unload ranges of ERR strings with a new 5631 "ERR_unload_strings" function. 5632 [Geoff Thorpe] 5633 5634 *) Add a copy() function to EVP_MD. 5635 [Ben Laurie] 5636 5637 *) Make EVP_MD routines take a context pointer instead of just the 5638 md_data void pointer. 5639 [Ben Laurie] 5640 5641 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 5642 that the digest can only process a single chunk of data 5643 (typically because it is provided by a piece of 5644 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 5645 is only going to provide a single chunk of data, and hence the 5646 framework needn't accumulate the data for oneshot drivers. 5647 [Ben Laurie] 5648 5649 *) As with "ERR", make it possible to replace the underlying "ex_data" 5650 functions. This change also alters the storage and management of global 5651 ex_data state - it's now all inside ex_data.c and all "class" code (eg. 5652 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 5653 index counters. The API functions that use this state have been changed 5654 to take a "class_index" rather than pointers to the class's local STACK 5655 and counter, and there is now an API function to dynamically create new 5656 classes. This centralisation allows us to (a) plug a lot of the 5657 thread-safety problems that existed, and (b) makes it possible to clean 5658 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 5659 such data would previously have always leaked in application code and 5660 workarounds were in place to make the memory debugging turn a blind eye 5661 to it. Application code that doesn't use this new function will still 5662 leak as before, but their memory debugging output will announce it now 5663 rather than letting it slide. 5664 5665 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 5666 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 5667 has a return value to indicate success or failure. 5668 [Geoff Thorpe] 5669 5670 *) Make it possible to replace the underlying "ERR" functions such that the 5671 global state (2 LHASH tables and 2 locks) is only used by the "default" 5672 implementation. This change also adds two functions to "get" and "set" 5673 the implementation prior to it being automatically set the first time 5674 any other ERR function takes place. Ie. an application can call "get", 5675 pass the return value to a module it has just loaded, and that module 5676 can call its own "set" function using that value. This means the 5677 module's "ERR" operations will use (and modify) the error state in the 5678 application and not in its own statically linked copy of OpenSSL code. 5679 [Geoff Thorpe] 5680 5681 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 5682 reference counts. This performs normal REF_PRINT/REF_CHECK macros on 5683 the operation, and provides a more encapsulated way for external code 5684 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 5685 to use these functions rather than manually incrementing the counts. 5686 5687 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 5688 [Geoff Thorpe] 5689 5690 *) Add EVP test program. 5691 [Ben Laurie] 5692 5693 *) Add symmetric cipher support to ENGINE. Expect the API to change! 5694 [Ben Laurie] 5695 5696 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 5697 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 5698 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 5699 These allow a CRL to be built without having to access X509_CRL fields 5700 directly. Modify 'ca' application to use new functions. 5701 [Steve Henson] 5702 5703 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 5704 bug workarounds. Rollback attack detection is a security feature. 5705 The problem will only arise on OpenSSL servers when TLSv1 is not 5706 available (sslv3_server_method() or SSL_OP_NO_TLSv1). 5707 Software authors not wanting to support TLSv1 will have special reasons 5708 for their choice and can explicitly enable this option. 5709 [Bodo Moeller, Lutz Jaenicke] 5710 5711 *) Rationalise EVP so it can be extended: don't include a union of 5712 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 5713 (similar to those existing for EVP_CIPHER_CTX). 5714 Usage example: 5715 5716 EVP_MD_CTX md; 5717 5718 EVP_MD_CTX_init(&md); /* new function call */ 5719 EVP_DigestInit(&md, EVP_sha1()); 5720 EVP_DigestUpdate(&md, in, len); 5721 EVP_DigestFinal(&md, out, NULL); 5722 EVP_MD_CTX_cleanup(&md); /* new function call */ 5723 5724 [Ben Laurie] 5725 5726 *) Make DES key schedule conform to the usual scheme, as well as 5727 correcting its structure. This means that calls to DES functions 5728 now have to pass a pointer to a des_key_schedule instead of a 5729 plain des_key_schedule (which was actually always a pointer 5730 anyway): E.g., 5731 5732 des_key_schedule ks; 5733 5734 des_set_key_checked(..., &ks); 5735 des_ncbc_encrypt(..., &ks, ...); 5736 5737 (Note that a later change renames 'des_...' into 'DES_...'.) 5738 [Ben Laurie] 5739 5740 *) Initial reduction of linker bloat: the use of some functions, such as 5741 PEM causes large amounts of unused functions to be linked in due to 5742 poor organisation. For example pem_all.c contains every PEM function 5743 which has a knock on effect of linking in large amounts of (unused) 5744 ASN1 code. Grouping together similar functions and splitting unrelated 5745 functions prevents this. 5746 [Steve Henson] 5747 5748 *) Cleanup of EVP macros. 5749 [Ben Laurie] 5750 5751 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 5752 correct _ecb suffix. 5753 [Ben Laurie] 5754 5755 *) Add initial OCSP responder support to ocsp application. The 5756 revocation information is handled using the text based index 5757 use by the ca application. The responder can either handle 5758 requests generated internally, supplied in files (for example 5759 via a CGI script) or using an internal minimal server. 5760 [Steve Henson] 5761 5762 *) Add configuration choices to get zlib compression for TLS. 5763 [Richard Levitte] 5764 5765 *) Changes to Kerberos SSL for RFC 2712 compliance: 5766 1. Implemented real KerberosWrapper, instead of just using 5767 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 5768 2. Implemented optional authenticator field of KerberosWrapper. 5769 5770 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 5771 and authenticator structs; see crypto/krb5/. 5772 5773 Generalized Kerberos calls to support multiple Kerberos libraries. 5774 [Vern Staats <staatsvr@asc.hpc.mil>, 5775 Jeffrey Altman <jaltman@columbia.edu> 5776 via Richard Levitte] 5777 5778 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 5779 already does with RSA. testdsa.h now has 'priv_key/pub_key' 5780 values for each of the key sizes rather than having just 5781 parameters (and 'speed' generating keys each time). 5782 [Geoff Thorpe] 5783 5784 *) Speed up EVP routines. 5785 Before: 5786encrypt 5787type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 5788des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 5789des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 5790des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 5791decrypt 5792des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 5793des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 5794des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 5795 After: 5796encrypt 5797des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 5798decrypt 5799des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 5800 [Ben Laurie] 5801 5802 *) Added the OS2-EMX target. 5803 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 5804 5805 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 5806 to support NCONF routines in extension code. New function CONF_set_nconf() 5807 to allow functions which take an NCONF to also handle the old LHASH 5808 structure: this means that the old CONF compatible routines can be 5809 retained (in particular wrt extensions) without having to duplicate the 5810 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 5811 [Steve Henson] 5812 5813 *) Enhance the general user interface with mechanisms for inner control 5814 and with possibilities to have yes/no kind of prompts. 5815 [Richard Levitte] 5816 5817 *) Change all calls to low level digest routines in the library and 5818 applications to use EVP. Add missing calls to HMAC_cleanup() and 5819 don't assume HMAC_CTX can be copied using memcpy(). 5820 [Verdon Walker <VWalker@novell.com>, Steve Henson] 5821 5822 *) Add the possibility to control engines through control names but with 5823 arbitrary arguments instead of just a string. 5824 Change the key loaders to take a UI_METHOD instead of a callback 5825 function pointer. NOTE: this breaks binary compatibility with earlier 5826 versions of OpenSSL [engine]. 5827 Adapt the nCipher code for these new conditions and add a card insertion 5828 callback. 5829 [Richard Levitte] 5830 5831 *) Enhance the general user interface with mechanisms to better support 5832 dialog box interfaces, application-defined prompts, the possibility 5833 to use defaults (for example default passwords from somewhere else) 5834 and interrupts/cancellations. 5835 [Richard Levitte] 5836 5837 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 5838 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 5839 [Steve Henson] 5840 5841 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 5842 tidy up some unnecessarily weird code in 'sk_new()'). 5843 [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 5844 5845 *) Change the key loading routines for ENGINEs to use the same kind 5846 callback (pem_password_cb) as all other routines that need this 5847 kind of callback. 5848 [Richard Levitte] 5849 5850 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 5851 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 5852 than this minimum value is recommended. 5853 [Lutz Jaenicke] 5854 5855 *) New random seeder for OpenVMS, using the system process statistics 5856 that are easily reachable. 5857 [Richard Levitte] 5858 5859 *) Windows apparently can't transparently handle global 5860 variables defined in DLLs. Initialisations such as: 5861 5862 const ASN1_ITEM *it = &ASN1_INTEGER_it; 5863 5864 wont compile. This is used by the any applications that need to 5865 declare their own ASN1 modules. This was fixed by adding the option 5866 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 5867 needed for static libraries under Win32. 5868 [Steve Henson] 5869 5870 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 5871 setting of purpose and trust fields. New X509_STORE trust and 5872 purpose functions and tidy up setting in other SSL functions. 5873 [Steve Henson] 5874 5875 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 5876 structure. These are inherited by X509_STORE_CTX when it is 5877 initialised. This allows various defaults to be set in the 5878 X509_STORE structure (such as flags for CRL checking and custom 5879 purpose or trust settings) for functions which only use X509_STORE_CTX 5880 internally such as S/MIME. 5881 5882 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 5883 trust settings if they are not set in X509_STORE. This allows X509_STORE 5884 purposes and trust (in S/MIME for example) to override any set by default. 5885 5886 Add command line options for CRL checking to smime, s_client and s_server 5887 applications. 5888 [Steve Henson] 5889 5890 *) Initial CRL based revocation checking. If the CRL checking flag(s) 5891 are set then the CRL is looked up in the X509_STORE structure and 5892 its validity and signature checked, then if the certificate is found 5893 in the CRL the verify fails with a revoked error. 5894 5895 Various new CRL related callbacks added to X509_STORE_CTX structure. 5896 5897 Command line options added to 'verify' application to support this. 5898 5899 This needs some additional work, such as being able to handle multiple 5900 CRLs with different times, extension based lookup (rather than just 5901 by subject name) and ultimately more complete V2 CRL extension 5902 handling. 5903 [Steve Henson] 5904 5905 *) Add a general user interface API (crypto/ui/). This is designed 5906 to replace things like des_read_password and friends (backward 5907 compatibility functions using this new API are provided). 5908 The purpose is to remove prompting functions from the DES code 5909 section as well as provide for prompting through dialog boxes in 5910 a window system and the like. 5911 [Richard Levitte] 5912 5913 *) Add "ex_data" support to ENGINE so implementations can add state at a 5914 per-structure level rather than having to store it globally. 5915 [Geoff] 5916 5917 *) Make it possible for ENGINE structures to be copied when retrieved by 5918 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 5919 This causes the "original" ENGINE structure to act like a template, 5920 analogous to the RSA vs. RSA_METHOD type of separation. Because of this 5921 operational state can be localised to each ENGINE structure, despite the 5922 fact they all share the same "methods". New ENGINE structures returned in 5923 this case have no functional references and the return value is the single 5924 structural reference. This matches the single structural reference returned 5925 by ENGINE_by_id() normally, when it is incremented on the pre-existing 5926 ENGINE structure. 5927 [Geoff] 5928 5929 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 5930 needs to match any other type at all we need to manually clear the 5931 tag cache. 5932 [Steve Henson] 5933 5934 *) Changes to the "openssl engine" utility to include; 5935 - verbosity levels ('-v', '-vv', and '-vvv') that provide information 5936 about an ENGINE's available control commands. 5937 - executing control commands from command line arguments using the 5938 '-pre' and '-post' switches. '-post' is only used if '-t' is 5939 specified and the ENGINE is successfully initialised. The syntax for 5940 the individual commands are colon-separated, for example; 5941 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 5942 [Geoff] 5943 5944 *) New dynamic control command support for ENGINEs. ENGINEs can now 5945 declare their own commands (numbers), names (strings), descriptions, 5946 and input types for run-time discovery by calling applications. A 5947 subset of these commands are implicitly classed as "executable" 5948 depending on their input type, and only these can be invoked through 5949 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 5950 can be based on user input, config files, etc). The distinction is 5951 that "executable" commands cannot return anything other than a boolean 5952 result and can only support numeric or string input, whereas some 5953 discoverable commands may only be for direct use through 5954 ENGINE_ctrl(), eg. supporting the exchange of binary data, function 5955 pointers, or other custom uses. The "executable" commands are to 5956 support parameterisations of ENGINE behaviour that can be 5957 unambiguously defined by ENGINEs and used consistently across any 5958 OpenSSL-based application. Commands have been added to all the 5959 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 5960 control over shared-library paths without source code alterations. 5961 [Geoff] 5962 5963 *) Changed all ENGINE implementations to dynamically allocate their 5964 ENGINEs rather than declaring them statically. Apart from this being 5965 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 5966 this also allows the implementations to compile without using the 5967 internal engine_int.h header. 5968 [Geoff] 5969 5970 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 5971 'const' value. Any code that should be able to modify a RAND_METHOD 5972 should already have non-const pointers to it (ie. they should only 5973 modify their own ones). 5974 [Geoff] 5975 5976 *) Made a variety of little tweaks to the ENGINE code. 5977 - "atalla" and "ubsec" string definitions were moved from header files 5978 to C code. "nuron" string definitions were placed in variables 5979 rather than hard-coded - allowing parameterisation of these values 5980 later on via ctrl() commands. 5981 - Removed unused "#if 0"'d code. 5982 - Fixed engine list iteration code so it uses ENGINE_free() to release 5983 structural references. 5984 - Constified the RAND_METHOD element of ENGINE structures. 5985 - Constified various get/set functions as appropriate and added 5986 missing functions (including a catch-all ENGINE_cpy that duplicates 5987 all ENGINE values onto a new ENGINE except reference counts/state). 5988 - Removed NULL parameter checks in get/set functions. Setting a method 5989 or function to NULL is a way of cancelling out a previously set 5990 value. Passing a NULL ENGINE parameter is just plain stupid anyway 5991 and doesn't justify the extra error symbols and code. 5992 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 5993 flags from engine_int.h to engine.h. 5994 - Changed prototypes for ENGINE handler functions (init(), finish(), 5995 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 5996 [Geoff] 5997 5998 *) Implement binary inversion algorithm for BN_mod_inverse in addition 5999 to the algorithm using long division. The binary algorithm can be 6000 used only if the modulus is odd. On 32-bit systems, it is faster 6001 only for relatively small moduli (roughly 20-30% for 128-bit moduli, 6002 roughly 5-15% for 256-bit moduli), so we use it only for moduli 6003 up to 450 bits. In 64-bit environments, the binary algorithm 6004 appears to be advantageous for much longer moduli; here we use it 6005 for moduli up to 2048 bits. 6006 [Bodo Moeller] 6007 6008 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 6009 could not support the combine flag in choice fields. 6010 [Steve Henson] 6011 6012 *) Add a 'copy_extensions' option to the 'ca' utility. This copies 6013 extensions from a certificate request to the certificate. 6014 [Steve Henson] 6015 6016 *) Allow multiple 'certopt' and 'nameopt' options to be separated 6017 by commas. Add 'namopt' and 'certopt' options to the 'ca' config 6018 file: this allows the display of the certificate about to be 6019 signed to be customised, to allow certain fields to be included 6020 or excluded and extension details. The old system didn't display 6021 multicharacter strings properly, omitted fields not in the policy 6022 and couldn't display additional details such as extensions. 6023 [Steve Henson] 6024 6025 *) Function EC_POINTs_mul for multiple scalar multiplication 6026 of an arbitrary number of elliptic curve points 6027 \sum scalars[i]*points[i], 6028 optionally including the generator defined for the EC_GROUP: 6029 scalar*generator + \sum scalars[i]*points[i]. 6030 6031 EC_POINT_mul is a simple wrapper function for the typical case 6032 that the point list has just one item (besides the optional 6033 generator). 6034 [Bodo Moeller] 6035 6036 *) First EC_METHODs for curves over GF(p): 6037 6038 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 6039 operations and provides various method functions that can also 6040 operate with faster implementations of modular arithmetic. 6041 6042 EC_GFp_mont_method() reuses most functions that are part of 6043 EC_GFp_simple_method, but uses Montgomery arithmetic. 6044 6045 [Bodo Moeller; point addition and point doubling 6046 implementation directly derived from source code provided by 6047 Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 6048 6049 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 6050 crypto/ec/ec_lib.c): 6051 6052 Curves are EC_GROUP objects (with an optional group generator) 6053 based on EC_METHODs that are built into the library. 6054 6055 Points are EC_POINT objects based on EC_GROUP objects. 6056 6057 Most of the framework would be able to handle curves over arbitrary 6058 finite fields, but as there are no obvious types for fields other 6059 than GF(p), some functions are limited to that for now. 6060 [Bodo Moeller] 6061 6062 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 6063 that the file contains a complete HTTP response. 6064 [Richard Levitte] 6065 6066 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 6067 change the def and num file printf format specifier from "%-40sXXX" 6068 to "%-39s XXX". The latter will always guarantee a space after the 6069 field while the former will cause them to run together if the field 6070 is 40 of more characters long. 6071 [Steve Henson] 6072 6073 *) Constify the cipher and digest 'method' functions and structures 6074 and modify related functions to take constant EVP_MD and EVP_CIPHER 6075 pointers. 6076 [Steve Henson] 6077 6078 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 6079 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 6080 [Bodo Moeller] 6081 6082 *) Modify EVP_Digest*() routines so they now return values. Although the 6083 internal software routines can never fail additional hardware versions 6084 might. 6085 [Steve Henson] 6086 6087 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 6088 6089 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 6090 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 6091 6092 ASN1 error codes 6093 ERR_R_NESTED_ASN1_ERROR 6094 ... 6095 ERR_R_MISSING_ASN1_EOS 6096 were 4 .. 9, conflicting with 6097 ERR_LIB_RSA (= ERR_R_RSA_LIB) 6098 ... 6099 ERR_LIB_PEM (= ERR_R_PEM_LIB). 6100 They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 6101 6102 Add new error code 'ERR_R_INTERNAL_ERROR'. 6103 [Bodo Moeller] 6104 6105 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 6106 suffices. 6107 [Bodo Moeller] 6108 6109 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 6110 sets the subject name for a new request or supersedes the 6111 subject name in a given request. Formats that can be parsed are 6112 'CN=Some Name, OU=myOU, C=IT' 6113 and 6114 'CN=Some Name/OU=myOU/C=IT'. 6115 6116 Add options '-batch' and '-verbose' to 'openssl req'. 6117 [Massimiliano Pala <madwolf@hackmasters.net>] 6118 6119 *) Introduce the possibility to access global variables through 6120 functions on platform were that's the best way to handle exporting 6121 global variables in shared libraries. To enable this functionality, 6122 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 6123 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 6124 is normally done by Configure or something similar). 6125 6126 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 6127 in the source file (foo.c) like this: 6128 6129 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 6130 OPENSSL_IMPLEMENT_GLOBAL(double,bar); 6131 6132 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 6133 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 6134 6135 OPENSSL_DECLARE_GLOBAL(int,foo); 6136 #define foo OPENSSL_GLOBAL_REF(foo) 6137 OPENSSL_DECLARE_GLOBAL(double,bar); 6138 #define bar OPENSSL_GLOBAL_REF(bar) 6139 6140 The #defines are very important, and therefore so is including the 6141 header file everywhere where the defined globals are used. 6142 6143 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 6144 of ASN.1 items, but that structure is a bit different. 6145 6146 The largest change is in util/mkdef.pl which has been enhanced with 6147 better and easier to understand logic to choose which symbols should 6148 go into the Windows .def files as well as a number of fixes and code 6149 cleanup (among others, algorithm keywords are now sorted 6150 lexicographically to avoid constant rewrites). 6151 [Richard Levitte] 6152 6153 *) In BN_div() keep a copy of the sign of 'num' before writing the 6154 result to 'rm' because if rm==num the value will be overwritten 6155 and produce the wrong result if 'num' is negative: this caused 6156 problems with BN_mod() and BN_nnmod(). 6157 [Steve Henson] 6158 6159 *) Function OCSP_request_verify(). This checks the signature on an 6160 OCSP request and verifies the signer certificate. The signer 6161 certificate is just checked for a generic purpose and OCSP request 6162 trust settings. 6163 [Steve Henson] 6164 6165 *) Add OCSP_check_validity() function to check the validity of OCSP 6166 responses. OCSP responses are prepared in real time and may only 6167 be a few seconds old. Simply checking that the current time lies 6168 between thisUpdate and nextUpdate max reject otherwise valid responses 6169 caused by either OCSP responder or client clock inaccuracy. Instead 6170 we allow thisUpdate and nextUpdate to fall within a certain period of 6171 the current time. The age of the response can also optionally be 6172 checked. Two new options -validity_period and -status_age added to 6173 ocsp utility. 6174 [Steve Henson] 6175 6176 *) If signature or public key algorithm is unrecognized print out its 6177 OID rather that just UNKNOWN. 6178 [Steve Henson] 6179 6180 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 6181 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 6182 ID to be generated from the issuer certificate alone which can then be 6183 passed to OCSP_id_issuer_cmp(). 6184 [Steve Henson] 6185 6186 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 6187 ASN1 modules to export functions returning ASN1_ITEM pointers 6188 instead of the ASN1_ITEM structures themselves. This adds several 6189 new macros which allow the underlying ASN1 function/structure to 6190 be accessed transparently. As a result code should not use ASN1_ITEM 6191 references directly (such as &X509_it) but instead use the relevant 6192 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 6193 use of the new ASN1 code on platforms where exporting structures 6194 is problematical (for example in shared libraries) but exporting 6195 functions returning pointers to structures is not. 6196 [Steve Henson] 6197 6198 *) Add support for overriding the generation of SSL/TLS session IDs. 6199 These callbacks can be registered either in an SSL_CTX or per SSL. 6200 The purpose of this is to allow applications to control, if they wish, 6201 the arbitrary values chosen for use as session IDs, particularly as it 6202 can be useful for session caching in multiple-server environments. A 6203 command-line switch for testing this (and any client code that wishes 6204 to use such a feature) has been added to "s_server". 6205 [Geoff Thorpe, Lutz Jaenicke] 6206 6207 *) Modify mkdef.pl to recognise and parse preprocessor conditionals 6208 of the form '#if defined(...) || defined(...) || ...' and 6209 '#if !defined(...) && !defined(...) && ...'. This also avoids 6210 the growing number of special cases it was previously handling. 6211 [Richard Levitte] 6212 6213 *) Make all configuration macros available for application by making 6214 sure they are available in opensslconf.h, by giving them names starting 6215 with "OPENSSL_" to avoid conflicts with other packages and by making 6216 sure e_os2.h will cover all platform-specific cases together with 6217 opensslconf.h. 6218 Additionally, it is now possible to define configuration/platform- 6219 specific names (called "system identities"). In the C code, these 6220 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 6221 macro with the name beginning with "OPENSSL_SYS_", which is determined 6222 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 6223 what is available. 6224 [Richard Levitte] 6225 6226 *) New option -set_serial to 'req' and 'x509' this allows the serial 6227 number to use to be specified on the command line. Previously self 6228 signed certificates were hard coded with serial number 0 and the 6229 CA options of 'x509' had to use a serial number in a file which was 6230 auto incremented. 6231 [Steve Henson] 6232 6233 *) New options to 'ca' utility to support V2 CRL entry extensions. 6234 Currently CRL reason, invalidity date and hold instruction are 6235 supported. Add new CRL extensions to V3 code and some new objects. 6236 [Steve Henson] 6237 6238 *) New function EVP_CIPHER_CTX_set_padding() this is used to 6239 disable standard block padding (aka PKCS#5 padding) in the EVP 6240 API, which was previously mandatory. This means that the data is 6241 not padded in any way and so the total length much be a multiple 6242 of the block size, otherwise an error occurs. 6243 [Steve Henson] 6244 6245 *) Initial (incomplete) OCSP SSL support. 6246 [Steve Henson] 6247 6248 *) New function OCSP_parse_url(). This splits up a URL into its host, 6249 port and path components: primarily to parse OCSP URLs. New -url 6250 option to ocsp utility. 6251 [Steve Henson] 6252 6253 *) New nonce behavior. The return value of OCSP_check_nonce() now 6254 reflects the various checks performed. Applications can decide 6255 whether to tolerate certain situations such as an absent nonce 6256 in a response when one was present in a request: the ocsp application 6257 just prints out a warning. New function OCSP_add1_basic_nonce() 6258 this is to allow responders to include a nonce in a response even if 6259 the request is nonce-less. 6260 [Steve Henson] 6261 6262 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 6263 skipped when using openssl x509 multiple times on a single input file, 6264 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 6265 [Bodo Moeller] 6266 6267 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 6268 set string type: to handle setting ASN1_TIME structures. Fix ca 6269 utility to correctly initialize revocation date of CRLs. 6270 [Steve Henson] 6271 6272 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 6273 the clients preferred ciphersuites and rather use its own preferences. 6274 Should help to work around M$ SGC (Server Gated Cryptography) bug in 6275 Internet Explorer by ensuring unchanged hash method during stepup. 6276 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 6277 [Lutz Jaenicke] 6278 6279 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 6280 to aes and add a new 'exist' option to print out symbols that don't 6281 appear to exist. 6282 [Steve Henson] 6283 6284 *) Additional options to ocsp utility to allow flags to be set and 6285 additional certificates supplied. 6286 [Steve Henson] 6287 6288 *) Add the option -VAfile to 'openssl ocsp', so the user can give the 6289 OCSP client a number of certificate to only verify the response 6290 signature against. 6291 [Richard Levitte] 6292 6293 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 6294 handle the new API. Currently only ECB, CBC modes supported. Add new 6295 AES OIDs. 6296 6297 Add TLS AES ciphersuites as described in RFC3268, "Advanced 6298 Encryption Standard (AES) Ciphersuites for Transport Layer 6299 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 6300 not enabled by default and were not part of the "ALL" ciphersuite 6301 alias because they were not yet official; they could be 6302 explicitly requested by specifying the "AESdraft" ciphersuite 6303 group alias. In the final release of OpenSSL 0.9.7, the group 6304 alias is called "AES" and is part of "ALL".) 6305 [Ben Laurie, Steve Henson, Bodo Moeller] 6306 6307 *) New function OCSP_copy_nonce() to copy nonce value (if present) from 6308 request to response. 6309 [Steve Henson] 6310 6311 *) Functions for OCSP responders. OCSP_request_onereq_count(), 6312 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 6313 extract information from a certificate request. OCSP_response_create() 6314 creates a response and optionally adds a basic response structure. 6315 OCSP_basic_add1_status() adds a complete single response to a basic 6316 response and returns the OCSP_SINGLERESP structure just added (to allow 6317 extensions to be included for example). OCSP_basic_add1_cert() adds a 6318 certificate to a basic response and OCSP_basic_sign() signs a basic 6319 response with various flags. New helper functions ASN1_TIME_check() 6320 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 6321 (converts ASN1_TIME to GeneralizedTime). 6322 [Steve Henson] 6323 6324 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 6325 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 6326 structure from a certificate. X509_pubkey_digest() digests the public_key 6327 contents: this is used in various key identifiers. 6328 [Steve Henson] 6329 6330 *) Make sk_sort() tolerate a NULL argument. 6331 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 6332 6333 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 6334 passed by the function are trusted implicitly. If any of them signed the 6335 response then it is assumed to be valid and is not verified. 6336 [Steve Henson] 6337 6338 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 6339 to data. This was previously part of the PKCS7 ASN1 code. This 6340 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 6341 [Steve Henson, reported by Kenneth R. Robinette 6342 <support@securenetterm.com>] 6343 6344 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 6345 routines: without these tracing memory leaks is very painful. 6346 Fix leaks in PKCS12 and PKCS7 routines. 6347 [Steve Henson] 6348 6349 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 6350 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 6351 effectively meant GeneralizedTime would never be used. Now it 6352 is initialised to -1 but X509_time_adj() now has to check the value 6353 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 6354 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 6355 [Steve Henson, reported by Kenneth R. Robinette 6356 <support@securenetterm.com>] 6357 6358 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 6359 result in a zero length in the ASN1_INTEGER structure which was 6360 not consistent with the structure when d2i_ASN1_INTEGER() was used 6361 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 6362 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 6363 where it did not print out a minus for negative ASN1_INTEGER. 6364 [Steve Henson] 6365 6366 *) Add summary printout to ocsp utility. The various functions which 6367 convert status values to strings have been renamed to: 6368 OCSP_response_status_str(), OCSP_cert_status_str() and 6369 OCSP_crl_reason_str() and are no longer static. New options 6370 to verify nonce values and to disable verification. OCSP response 6371 printout format cleaned up. 6372 [Steve Henson] 6373 6374 *) Add additional OCSP certificate checks. These are those specified 6375 in RFC2560. This consists of two separate checks: the CA of the 6376 certificate being checked must either be the OCSP signer certificate 6377 or the issuer of the OCSP signer certificate. In the latter case the 6378 OCSP signer certificate must contain the OCSP signing extended key 6379 usage. This check is performed by attempting to match the OCSP 6380 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 6381 in the OCSP_CERTID structures of the response. 6382 [Steve Henson] 6383 6384 *) Initial OCSP certificate verification added to OCSP_basic_verify() 6385 and related routines. This uses the standard OpenSSL certificate 6386 verify routines to perform initial checks (just CA validity) and 6387 to obtain the certificate chain. Then additional checks will be 6388 performed on the chain. Currently the root CA is checked to see 6389 if it is explicitly trusted for OCSP signing. This is used to set 6390 a root CA as a global signing root: that is any certificate that 6391 chains to that CA is an acceptable OCSP signing certificate. 6392 [Steve Henson] 6393 6394 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 6395 extensions from a separate configuration file. 6396 As when reading extensions from the main configuration file, 6397 the '-extensions ...' option may be used for specifying the 6398 section to use. 6399 [Massimiliano Pala <madwolf@comune.modena.it>] 6400 6401 *) New OCSP utility. Allows OCSP requests to be generated or 6402 read. The request can be sent to a responder and the output 6403 parsed, outputed or printed in text form. Not complete yet: 6404 still needs to check the OCSP response validity. 6405 [Steve Henson] 6406 6407 *) New subcommands for 'openssl ca': 6408 'openssl ca -status <serial>' prints the status of the cert with 6409 the given serial number (according to the index file). 6410 'openssl ca -updatedb' updates the expiry status of certificates 6411 in the index file. 6412 [Massimiliano Pala <madwolf@comune.modena.it>] 6413 6414 *) New '-newreq-nodes' command option to CA.pl. This is like 6415 '-newreq', but calls 'openssl req' with the '-nodes' option 6416 so that the resulting key is not encrypted. 6417 [Damien Miller <djm@mindrot.org>] 6418 6419 *) New configuration for the GNU Hurd. 6420 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 6421 6422 *) Initial code to implement OCSP basic response verify. This 6423 is currently incomplete. Currently just finds the signer's 6424 certificate and verifies the signature on the response. 6425 [Steve Henson] 6426 6427 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 6428 value of OPENSSLDIR. This is available via the new '-d' option 6429 to 'openssl version', and is also included in 'openssl version -a'. 6430 [Bodo Moeller] 6431 6432 *) Allowing defining memory allocation callbacks that will be given 6433 file name and line number information in additional arguments 6434 (a const char* and an int). The basic functionality remains, as 6435 well as the original possibility to just replace malloc(), 6436 realloc() and free() by functions that do not know about these 6437 additional arguments. To register and find out the current 6438 settings for extended allocation functions, the following 6439 functions are provided: 6440 6441 CRYPTO_set_mem_ex_functions 6442 CRYPTO_set_locked_mem_ex_functions 6443 CRYPTO_get_mem_ex_functions 6444 CRYPTO_get_locked_mem_ex_functions 6445 6446 These work the same way as CRYPTO_set_mem_functions and friends. 6447 CRYPTO_get_[locked_]mem_functions now writes 0 where such an 6448 extended allocation function is enabled. 6449 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 6450 a conventional allocation function is enabled. 6451 [Richard Levitte, Bodo Moeller] 6452 6453 *) Finish off removing the remaining LHASH function pointer casts. 6454 There should no longer be any prototype-casting required when using 6455 the LHASH abstraction, and any casts that remain are "bugs". See 6456 the callback types and macros at the head of lhash.h for details 6457 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 6458 [Geoff Thorpe] 6459 6460 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 6461 If /dev/[u]random devices are not available or do not return enough 6462 entropy, EGD style sockets (served by EGD or PRNGD) will automatically 6463 be queried. 6464 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 6465 /etc/entropy will be queried once each in this sequence, quering stops 6466 when enough entropy was collected without querying more sockets. 6467 [Lutz Jaenicke] 6468 6469 *) Change the Unix RAND_poll() variant to be able to poll several 6470 random devices, as specified by DEVRANDOM, until a sufficient amount 6471 of data has been collected. We spend at most 10 ms on each file 6472 (select timeout) and read in non-blocking mode. DEVRANDOM now 6473 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 6474 (previously it was just the string "/dev/urandom"), so on typical 6475 platforms the 10 ms delay will never occur. 6476 Also separate out the Unix variant to its own file, rand_unix.c. 6477 For VMS, there's a currently-empty rand_vms.c. 6478 [Richard Levitte] 6479 6480 *) Move OCSP client related routines to ocsp_cl.c. These 6481 provide utility functions which an application needing 6482 to issue a request to an OCSP responder and analyse the 6483 response will typically need: as opposed to those which an 6484 OCSP responder itself would need which will be added later. 6485 6486 OCSP_request_sign() signs an OCSP request with an API similar 6487 to PKCS7_sign(). OCSP_response_status() returns status of OCSP 6488 response. OCSP_response_get1_basic() extracts basic response 6489 from response. OCSP_resp_find_status(): finds and extracts status 6490 information from an OCSP_CERTID structure (which will be created 6491 when the request structure is built). These are built from lower 6492 level functions which work on OCSP_SINGLERESP structures but 6493 wont normally be used unless the application wishes to examine 6494 extensions in the OCSP response for example. 6495 6496 Replace nonce routines with a pair of functions. 6497 OCSP_request_add1_nonce() adds a nonce value and optionally 6498 generates a random value. OCSP_check_nonce() checks the 6499 validity of the nonce in an OCSP response. 6500 [Steve Henson] 6501 6502 *) Change function OCSP_request_add() to OCSP_request_add0_id(). 6503 This doesn't copy the supplied OCSP_CERTID and avoids the 6504 need to free up the newly created id. Change return type 6505 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 6506 This can then be used to add extensions to the request. 6507 Deleted OCSP_request_new(), since most of its functionality 6508 is now in OCSP_REQUEST_new() (and the case insensitive name 6509 clash) apart from the ability to set the request name which 6510 will be added elsewhere. 6511 [Steve Henson] 6512 6513 *) Update OCSP API. Remove obsolete extensions argument from 6514 various functions. Extensions are now handled using the new 6515 OCSP extension code. New simple OCSP HTTP function which 6516 can be used to send requests and parse the response. 6517 [Steve Henson] 6518 6519 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 6520 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 6521 uses the special reorder version of SET OF to sort the attributes 6522 and reorder them to match the encoded order. This resolves a long 6523 standing problem: a verify on a PKCS7 structure just after signing 6524 it used to fail because the attribute order did not match the 6525 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 6526 it uses the received order. This is necessary to tolerate some broken 6527 software that does not order SET OF. This is handled by encoding 6528 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 6529 to produce the required SET OF. 6530 [Steve Henson] 6531 6532 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 6533 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 6534 files to get correct declarations of the ASN.1 item variables. 6535 [Richard Levitte] 6536 6537 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 6538 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 6539 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 6540 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 6541 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 6542 ASN1_ITEM and no wrapper functions. 6543 [Steve Henson] 6544 6545 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 6546 replace the old function pointer based I/O routines. Change most of 6547 the *_d2i_bio() and *_d2i_fp() functions to use these. 6548 [Steve Henson] 6549 6550 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 6551 lines, recognice more "algorithms" that can be deselected, and make 6552 it complain about algorithm deselection that isn't recognised. 6553 [Richard Levitte] 6554 6555 *) New ASN1 functions to handle dup, sign, verify, digest, pack and 6556 unpack operations in terms of ASN1_ITEM. Modify existing wrappers 6557 to use new functions. Add NO_ASN1_OLD which can be set to remove 6558 some old style ASN1 functions: this can be used to determine if old 6559 code will still work when these eventually go away. 6560 [Steve Henson] 6561 6562 *) New extension functions for OCSP structures, these follow the 6563 same conventions as certificates and CRLs. 6564 [Steve Henson] 6565 6566 *) New function X509V3_add1_i2d(). This automatically encodes and 6567 adds an extension. Its behaviour can be customised with various 6568 flags to append, replace or delete. Various wrappers added for 6569 certifcates and CRLs. 6570 [Steve Henson] 6571 6572 *) Fix to avoid calling the underlying ASN1 print routine when 6573 an extension cannot be parsed. Correct a typo in the 6574 OCSP_SERVICELOC extension. Tidy up print OCSP format. 6575 [Steve Henson] 6576 6577 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 6578 entries for variables. 6579 [Steve Henson] 6580 6581 *) Add functionality to apps/openssl.c for detecting locking 6582 problems: As the program is single-threaded, all we have 6583 to do is register a locking callback using an array for 6584 storing which locks are currently held by the program. 6585 [Bodo Moeller] 6586 6587 *) Use a lock around the call to CRYPTO_get_ex_new_index() in 6588 SSL_get_ex_data_X509_STORE_idx(), which is used in 6589 ssl_verify_cert_chain() and thus can be called at any time 6590 during TLS/SSL handshakes so that thread-safety is essential. 6591 Unfortunately, the ex_data design is not at all suited 6592 for multi-threaded use, so it probably should be abolished. 6593 [Bodo Moeller] 6594 6595 *) Added Broadcom "ubsec" ENGINE to OpenSSL. 6596 [Broadcom, tweaked and integrated by Geoff Thorpe] 6597 6598 *) Move common extension printing code to new function 6599 X509V3_print_extensions(). Reorganise OCSP print routines and 6600 implement some needed OCSP ASN1 functions. Add OCSP extensions. 6601 [Steve Henson] 6602 6603 *) New function X509_signature_print() to remove duplication in some 6604 print routines. 6605 [Steve Henson] 6606 6607 *) Add a special meaning when SET OF and SEQUENCE OF flags are both 6608 set (this was treated exactly the same as SET OF previously). This 6609 is used to reorder the STACK representing the structure to match the 6610 encoding. This will be used to get round a problem where a PKCS7 6611 structure which was signed could not be verified because the STACK 6612 order did not reflect the encoded order. 6613 [Steve Henson] 6614 6615 *) Reimplement the OCSP ASN1 module using the new code. 6616 [Steve Henson] 6617 6618 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 6619 for its ASN1 operations. The old style function pointers still exist 6620 for now but they will eventually go away. 6621 [Steve Henson] 6622 6623 *) Merge in replacement ASN1 code from the ASN1 branch. This almost 6624 completely replaces the old ASN1 functionality with a table driven 6625 encoder and decoder which interprets an ASN1_ITEM structure describing 6626 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 6627 largely maintained. Almost all of the old asn1_mac.h macro based ASN1 6628 has also been converted to the new form. 6629 [Steve Henson] 6630 6631 *) Change BN_mod_exp_recp so that negative moduli are tolerated 6632 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 6633 so that BN_mod_exp_mont and BN_mod_exp_mont_word work 6634 for negative moduli. 6635 [Bodo Moeller] 6636 6637 *) Fix BN_uadd and BN_usub: Always return non-negative results instead 6638 of not touching the result's sign bit. 6639 [Bodo Moeller] 6640 6641 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 6642 set. 6643 [Bodo Moeller] 6644 6645 *) Changed the LHASH code to use prototypes for callbacks, and created 6646 macros to declare and implement thin (optionally static) functions 6647 that provide type-safety and avoid function pointer casting for the 6648 type-specific callbacks. 6649 [Geoff Thorpe] 6650 6651 *) Added Kerberos Cipher Suites to be used with TLS, as written in 6652 RFC 2712. 6653 [Veers Staats <staatsvr@asc.hpc.mil>, 6654 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 6655 6656 *) Reformat the FAQ so the different questions and answers can be divided 6657 in sections depending on the subject. 6658 [Richard Levitte] 6659 6660 *) Have the zlib compression code load ZLIB.DLL dynamically under 6661 Windows. 6662 [Richard Levitte] 6663 6664 *) New function BN_mod_sqrt for computing square roots modulo a prime 6665 (using the probabilistic Tonelli-Shanks algorithm unless 6666 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 6667 be handled deterministically). 6668 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 6669 6670 *) Make BN_mod_inverse faster by explicitly handling small quotients 6671 in the Euclid loop. (Speed gain about 20% for small moduli [256 or 6672 512 bits], about 30% for larger ones [1024 or 2048 bits].) 6673 [Bodo Moeller] 6674 6675 *) New function BN_kronecker. 6676 [Bodo Moeller] 6677 6678 *) Fix BN_gcd so that it works on negative inputs; the result is 6679 positive unless both parameters are zero. 6680 Previously something reasonably close to an infinite loop was 6681 possible because numbers could be growing instead of shrinking 6682 in the implementation of Euclid's algorithm. 6683 [Bodo Moeller] 6684 6685 *) Fix BN_is_word() and BN_is_one() macros to take into account the 6686 sign of the number in question. 6687 6688 Fix BN_is_word(a,w) to work correctly for w == 0. 6689 6690 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 6691 because its test if the absolute value of 'a' equals 'w'. 6692 Note that BN_abs_is_word does *not* handle w == 0 reliably; 6693 it exists mostly for use in the implementations of BN_is_zero(), 6694 BN_is_one(), and BN_is_word(). 6695 [Bodo Moeller] 6696 6697 *) New function BN_swap. 6698 [Bodo Moeller] 6699 6700 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 6701 the exponentiation functions are more likely to produce reasonable 6702 results on negative inputs. 6703 [Bodo Moeller] 6704 6705 *) Change BN_mod_mul so that the result is always non-negative. 6706 Previously, it could be negative if one of the factors was negative; 6707 I don't think anyone really wanted that behaviour. 6708 [Bodo Moeller] 6709 6710 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 6711 (except for exponentiation, which stays in crypto/bn/bn_exp.c, 6712 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 6713 and add new functions: 6714 6715 BN_nnmod 6716 BN_mod_sqr 6717 BN_mod_add 6718 BN_mod_add_quick 6719 BN_mod_sub 6720 BN_mod_sub_quick 6721 BN_mod_lshift1 6722 BN_mod_lshift1_quick 6723 BN_mod_lshift 6724 BN_mod_lshift_quick 6725 6726 These functions always generate non-negative results. 6727 6728 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 6729 such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 6730 6731 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 6732 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 6733 be reduced modulo m. 6734 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 6735 6736#if 0 6737 The following entry accidentily appeared in the CHANGES file 6738 distributed with OpenSSL 0.9.7. The modifications described in 6739 it do *not* apply to OpenSSL 0.9.7. 6740 6741 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 6742 was actually never needed) and in BN_mul(). The removal in BN_mul() 6743 required a small change in bn_mul_part_recursive() and the addition 6744 of the functions bn_cmp_part_words(), bn_sub_part_words() and 6745 bn_add_part_words(), which do the same thing as bn_cmp_words(), 6746 bn_sub_words() and bn_add_words() except they take arrays with 6747 differing sizes. 6748 [Richard Levitte] 6749#endif 6750 6751 *) In 'openssl passwd', verify passwords read from the terminal 6752 unless the '-salt' option is used (which usually means that 6753 verification would just waste user's time since the resulting 6754 hash is going to be compared with some given password hash) 6755 or the new '-noverify' option is used. 6756 6757 This is an incompatible change, but it does not affect 6758 non-interactive use of 'openssl passwd' (passwords on the command 6759 line, '-stdin' option, '-in ...' option) and thus should not 6760 cause any problems. 6761 [Bodo Moeller] 6762 6763 *) Remove all references to RSAref, since there's no more need for it. 6764 [Richard Levitte] 6765 6766 *) Make DSO load along a path given through an environment variable 6767 (SHLIB_PATH) with shl_load(). 6768 [Richard Levitte] 6769 6770 *) Constify the ENGINE code as a result of BIGNUM constification. 6771 Also constify the RSA code and most things related to it. In a 6772 few places, most notable in the depth of the ASN.1 code, ugly 6773 casts back to non-const were required (to be solved at a later 6774 time) 6775 [Richard Levitte] 6776 6777 *) Make it so the openssl application has all engines loaded by default. 6778 [Richard Levitte] 6779 6780 *) Constify the BIGNUM routines a little more. 6781 [Richard Levitte] 6782 6783 *) Add the following functions: 6784 6785 ENGINE_load_cswift() 6786 ENGINE_load_chil() 6787 ENGINE_load_atalla() 6788 ENGINE_load_nuron() 6789 ENGINE_load_builtin_engines() 6790 6791 That way, an application can itself choose if external engines that 6792 are built-in in OpenSSL shall ever be used or not. The benefit is 6793 that applications won't have to be linked with libdl or other dso 6794 libraries unless it's really needed. 6795 6796 Changed 'openssl engine' to load all engines on demand. 6797 Changed the engine header files to avoid the duplication of some 6798 declarations (they differed!). 6799 [Richard Levitte] 6800 6801 *) 'openssl engine' can now list capabilities. 6802 [Richard Levitte] 6803 6804 *) Better error reporting in 'openssl engine'. 6805 [Richard Levitte] 6806 6807 *) Never call load_dh_param(NULL) in s_server. 6808 [Bodo Moeller] 6809 6810 *) Add engine application. It can currently list engines by name and 6811 identity, and test if they are actually available. 6812 [Richard Levitte] 6813 6814 *) Improve RPM specification file by forcing symbolic linking and making 6815 sure the installed documentation is also owned by root.root. 6816 [Damien Miller <djm@mindrot.org>] 6817 6818 *) Give the OpenSSL applications more possibilities to make use of 6819 keys (public as well as private) handled by engines. 6820 [Richard Levitte] 6821 6822 *) Add OCSP code that comes from CertCo. 6823 [Richard Levitte] 6824 6825 *) Add VMS support for the Rijndael code. 6826 [Richard Levitte] 6827 6828 *) Added untested support for Nuron crypto accelerator. 6829 [Ben Laurie] 6830 6831 *) Add support for external cryptographic devices. This code was 6832 previously distributed separately as the "engine" branch. 6833 [Geoff Thorpe, Richard Levitte] 6834 6835 *) Rework the filename-translation in the DSO code. It is now possible to 6836 have far greater control over how a "name" is turned into a filename 6837 depending on the operating environment and any oddities about the 6838 different shared library filenames on each system. 6839 [Geoff Thorpe] 6840 6841 *) Support threads on FreeBSD-elf in Configure. 6842 [Richard Levitte] 6843 6844 *) Fix for SHA1 assembly problem with MASM: it produces 6845 warnings about corrupt line number information when assembling 6846 with debugging information. This is caused by the overlapping 6847 of two sections. 6848 [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 6849 6850 *) NCONF changes. 6851 NCONF_get_number() has no error checking at all. As a replacement, 6852 NCONF_get_number_e() is defined (_e for "error checking") and is 6853 promoted strongly. The old NCONF_get_number is kept around for 6854 binary backward compatibility. 6855 Make it possible for methods to load from something other than a BIO, 6856 by providing a function pointer that is given a name instead of a BIO. 6857 For example, this could be used to load configuration data from an 6858 LDAP server. 6859 [Richard Levitte] 6860 6861 *) Fix for non blocking accept BIOs. Added new I/O special reason 6862 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 6863 with non blocking I/O was not possible because no retry code was 6864 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 6865 this case. 6866 [Steve Henson] 6867 6868 *) Added the beginnings of Rijndael support. 6869 [Ben Laurie] 6870 6871 *) Fix for bug in DirectoryString mask setting. Add support for 6872 X509_NAME_print_ex() in 'req' and X509_print_ex() function 6873 to allow certificate printing to more controllable, additional 6874 'certopt' option to 'x509' to allow new printing options to be 6875 set. 6876 [Steve Henson] 6877 6878 *) Clean old EAY MD5 hack from e_os.h. 6879 [Richard Levitte] 6880 6881 Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 6882 6883 *) Fix null-pointer assignment in do_change_cipher_spec() revealed 6884 by using the Codenomicon TLS Test Tool (CVE-2004-0079) 6885 [Joe Orton, Steve Henson] 6886 6887 Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 6888 6889 *) Fix additional bug revealed by the NISCC test suite: 6890 6891 Stop bug triggering large recursion when presented with 6892 certain ASN.1 tags (CVE-2003-0851) 6893 [Steve Henson] 6894 6895 Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 6896 6897 *) Fix various bugs revealed by running the NISCC test suite: 6898 6899 Stop out of bounds reads in the ASN1 code when presented with 6900 invalid tags (CVE-2003-0543 and CVE-2003-0544). 6901 6902 If verify callback ignores invalid public key errors don't try to check 6903 certificate signature with the NULL public key. 6904 6905 [Steve Henson] 6906 6907 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 6908 if the server requested one: as stated in TLS 1.0 and SSL 3.0 6909 specifications. 6910 [Steve Henson] 6911 6912 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 6913 extra data after the compression methods not only for TLS 1.0 6914 but also for SSL 3.0 (as required by the specification). 6915 [Bodo Moeller; problem pointed out by Matthias Loepfe] 6916 6917 *) Change X509_certificate_type() to mark the key as exported/exportable 6918 when it's 512 *bits* long, not 512 bytes. 6919 [Richard Levitte] 6920 6921 Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 6922 6923 *) Countermeasure against the Klima-Pokorny-Rosa extension of 6924 Bleichbacher's attack on PKCS #1 v1.5 padding: treat 6925 a protocol version number mismatch like a decryption error 6926 in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 6927 [Bodo Moeller] 6928 6929 *) Turn on RSA blinding by default in the default implementation 6930 to avoid a timing attack. Applications that don't want it can call 6931 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 6932 They would be ill-advised to do so in most cases. 6933 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 6934 6935 *) Change RSA blinding code so that it works when the PRNG is not 6936 seeded (in this case, the secret RSA exponent is abused as 6937 an unpredictable seed -- if it is not unpredictable, there 6938 is no point in blinding anyway). Make RSA blinding thread-safe 6939 by remembering the creator's thread ID in rsa->blinding and 6940 having all other threads use local one-time blinding factors 6941 (this requires more computation than sharing rsa->blinding, but 6942 avoids excessive locking; and if an RSA object is not shared 6943 between threads, blinding will still be very fast). 6944 [Bodo Moeller] 6945 6946 Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 6947 6948 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 6949 via timing by performing a MAC computation even if incorrrect 6950 block cipher padding has been found. This is a countermeasure 6951 against active attacks where the attacker has to distinguish 6952 between bad padding and a MAC verification error. (CVE-2003-0078) 6953 6954 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 6955 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 6956 Martin Vuagnoux (EPFL, Ilion)] 6957 6958 Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 6959 6960 *) New function OPENSSL_cleanse(), which is used to cleanse a section of 6961 memory from it's contents. This is done with a counter that will 6962 place alternating values in each byte. This can be used to solve 6963 two issues: 1) the removal of calls to memset() by highly optimizing 6964 compilers, and 2) cleansing with other values than 0, since those can 6965 be read through on certain media, for example a swap space on disk. 6966 [Geoff Thorpe] 6967 6968 *) Bugfix: client side session caching did not work with external caching, 6969 because the session->cipher setting was not restored when reloading 6970 from the external cache. This problem was masked, when 6971 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 6972 (Found by Steve Haslam <steve@araqnid.ddts.net>.) 6973 [Lutz Jaenicke] 6974 6975 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 6976 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 6977 [Zeev Lieber <zeev-l@yahoo.com>] 6978 6979 *) Undo an undocumented change introduced in 0.9.6e which caused 6980 repeated calls to OpenSSL_add_all_ciphers() and 6981 OpenSSL_add_all_digests() to be ignored, even after calling 6982 EVP_cleanup(). 6983 [Richard Levitte] 6984 6985 *) Change the default configuration reader to deal with last line not 6986 being properly terminated. 6987 [Richard Levitte] 6988 6989 *) Change X509_NAME_cmp() so it applies the special rules on handling 6990 DN values that are of type PrintableString, as well as RDNs of type 6991 emailAddress where the value has the type ia5String. 6992 [stefank@valicert.com via Richard Levitte] 6993 6994 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 6995 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 6996 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 6997 the bitwise-OR of the two for use by the majority of applications 6998 wanting this behaviour, and update the docs. The documented 6999 behaviour and actual behaviour were inconsistent and had been 7000 changing anyway, so this is more a bug-fix than a behavioural 7001 change. 7002 [Geoff Thorpe, diagnosed by Nadav Har'El] 7003 7004 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 7005 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 7006 [Bodo Moeller] 7007 7008 *) Fix initialization code race conditions in 7009 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 7010 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 7011 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 7012 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 7013 ssl2_get_cipher_by_char(), 7014 ssl3_get_cipher_by_char(). 7015 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 7016 7017 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 7018 the cached sessions are flushed, as the remove_cb() might use ex_data 7019 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 7020 (see [openssl.org #212]). 7021 [Geoff Thorpe, Lutz Jaenicke] 7022 7023 *) Fix typo in OBJ_txt2obj which incorrectly passed the content 7024 length, instead of the encoding length to d2i_ASN1_OBJECT. 7025 [Steve Henson] 7026 7027 Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 7028 7029 *) [In 0.9.6g-engine release:] 7030 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 7031 [Lynn Gazis <lgazis@rainbow.com>] 7032 7033 Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 7034 7035 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 7036 and get fix the header length calculation. 7037 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 7038 Alon Kantor <alonk@checkpoint.com> (and others), 7039 Steve Henson] 7040 7041 *) Use proper error handling instead of 'assertions' in buffer 7042 overflow checks added in 0.9.6e. This prevents DoS (the 7043 assertions could call abort()). 7044 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 7045 7046 Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 7047 7048 *) Add various sanity checks to asn1_get_length() to reject 7049 the ASN1 length bytes if they exceed sizeof(long), will appear 7050 negative or the content length exceeds the length of the 7051 supplied buffer. 7052 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 7053 7054 *) Fix cipher selection routines: ciphers without encryption had no flags 7055 for the cipher strength set and where therefore not handled correctly 7056 by the selection routines (PR #130). 7057 [Lutz Jaenicke] 7058 7059 *) Fix EVP_dsa_sha macro. 7060 [Nils Larsch] 7061 7062 *) New option 7063 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 7064 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 7065 that was added in OpenSSL 0.9.6d. 7066 7067 As the countermeasure turned out to be incompatible with some 7068 broken SSL implementations, the new option is part of SSL_OP_ALL. 7069 SSL_OP_ALL is usually employed when compatibility with weird SSL 7070 implementations is desired (e.g. '-bugs' option to 's_client' and 7071 's_server'), so the new option is automatically set in many 7072 applications. 7073 [Bodo Moeller] 7074 7075 *) Changes in security patch: 7076 7077 Changes marked "(CHATS)" were sponsored by the Defense Advanced 7078 Research Projects Agency (DARPA) and Air Force Research Laboratory, 7079 Air Force Materiel Command, USAF, under agreement number 7080 F30602-01-2-0537. 7081 7082 *) Add various sanity checks to asn1_get_length() to reject 7083 the ASN1 length bytes if they exceed sizeof(long), will appear 7084 negative or the content length exceeds the length of the 7085 supplied buffer. (CVE-2002-0659) 7086 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 7087 7088 *) Assertions for various potential buffer overflows, not known to 7089 happen in practice. 7090 [Ben Laurie (CHATS)] 7091 7092 *) Various temporary buffers to hold ASCII versions of integers were 7093 too small for 64 bit platforms. (CVE-2002-0655) 7094 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 7095 7096 *) Remote buffer overflow in SSL3 protocol - an attacker could 7097 supply an oversized session ID to a client. (CVE-2002-0656) 7098 [Ben Laurie (CHATS)] 7099 7100 *) Remote buffer overflow in SSL2 protocol - an attacker could 7101 supply an oversized client master key. (CVE-2002-0656) 7102 [Ben Laurie (CHATS)] 7103 7104 Changes between 0.9.6c and 0.9.6d [9 May 2002] 7105 7106 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 7107 encoded as NULL) with id-dsa-with-sha1. 7108 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 7109 7110 *) Check various X509_...() return values in apps/req.c. 7111 [Nils Larsch <nla@trustcenter.de>] 7112 7113 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 7114 an end-of-file condition would erronously be flagged, when the CRLF 7115 was just at the end of a processed block. The bug was discovered when 7116 processing data through a buffering memory BIO handing the data to a 7117 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 7118 <ptsekov@syntrex.com> and Nedelcho Stanev. 7119 [Lutz Jaenicke] 7120 7121 *) Implement a countermeasure against a vulnerability recently found 7122 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 7123 before application data chunks to avoid the use of known IVs 7124 with data potentially chosen by the attacker. 7125 [Bodo Moeller] 7126 7127 *) Fix length checks in ssl3_get_client_hello(). 7128 [Bodo Moeller] 7129 7130 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 7131 to prevent ssl3_read_internal() from incorrectly assuming that 7132 ssl3_read_bytes() found application data while handshake 7133 processing was enabled when in fact s->s3->in_read_app_data was 7134 merely automatically cleared during the initial handshake. 7135 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 7136 7137 *) Fix object definitions for Private and Enterprise: they were not 7138 recognized in their shortname (=lowercase) representation. Extend 7139 obj_dat.pl to issue an error when using undefined keywords instead 7140 of silently ignoring the problem (Svenning Sorensen 7141 <sss@sss.dnsalias.net>). 7142 [Lutz Jaenicke] 7143 7144 *) Fix DH_generate_parameters() so that it works for 'non-standard' 7145 generators, i.e. generators other than 2 and 5. (Previously, the 7146 code did not properly initialise the 'add' and 'rem' values to 7147 BN_generate_prime().) 7148 7149 In the new general case, we do not insist that 'generator' is 7150 actually a primitive root: This requirement is rather pointless; 7151 a generator of the order-q subgroup is just as good, if not 7152 better. 7153 [Bodo Moeller] 7154 7155 *) Map new X509 verification errors to alerts. Discovered and submitted by 7156 Tom Wu <tom@arcot.com>. 7157 [Lutz Jaenicke] 7158 7159 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 7160 returning non-zero before the data has been completely received 7161 when using non-blocking I/O. 7162 [Bodo Moeller; problem pointed out by John Hughes] 7163 7164 *) Some of the ciphers missed the strength entry (SSL_LOW etc). 7165 [Ben Laurie, Lutz Jaenicke] 7166 7167 *) Fix bug in SSL_clear(): bad sessions were not removed (found by 7168 Yoram Zahavi <YoramZ@gilian.com>). 7169 [Lutz Jaenicke] 7170 7171 *) Add information about CygWin 1.3 and on, and preserve proper 7172 configuration for the versions before that. 7173 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 7174 7175 *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 7176 check whether we deal with a copy of a session and do not delete from 7177 the cache in this case. Problem reported by "Izhar Shoshani Levi" 7178 <izhar@checkpoint.com>. 7179 [Lutz Jaenicke] 7180 7181 *) Do not store session data into the internal session cache, if it 7182 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 7183 flag is set). Proposed by Aslam <aslam@funk.com>. 7184 [Lutz Jaenicke] 7185 7186 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 7187 value is 0. 7188 [Richard Levitte] 7189 7190 *) [In 0.9.6d-engine release:] 7191 Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 7192 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 7193 7194 *) Add the configuration target linux-s390x. 7195 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 7196 7197 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 7198 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 7199 variable as an indication that a ClientHello message has been 7200 received. As the flag value will be lost between multiple 7201 invocations of ssl3_accept when using non-blocking I/O, the 7202 function may not be aware that a handshake has actually taken 7203 place, thus preventing a new session from being added to the 7204 session cache. 7205 7206 To avoid this problem, we now set s->new_session to 2 instead of 7207 using a local variable. 7208 [Lutz Jaenicke, Bodo Moeller] 7209 7210 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 7211 if the SSL_R_LENGTH_MISMATCH error is detected. 7212 [Geoff Thorpe, Bodo Moeller] 7213 7214 *) New 'shared_ldflag' column in Configure platform table. 7215 [Richard Levitte] 7216 7217 *) Fix EVP_CIPHER_mode macro. 7218 ["Dan S. Camper" <dan@bti.net>] 7219 7220 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 7221 type, we must throw them away by setting rr->length to 0. 7222 [D P Chang <dpc@qualys.com>] 7223 7224 Changes between 0.9.6b and 0.9.6c [21 dec 2001] 7225 7226 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 7227 <Dominikus.Scherkl@biodata.com>. (The previous implementation 7228 worked incorrectly for those cases where range = 10..._2 and 7229 3*range is two bits longer than range.) 7230 [Bodo Moeller] 7231 7232 *) Only add signing time to PKCS7 structures if it is not already 7233 present. 7234 [Steve Henson] 7235 7236 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 7237 OBJ_ld_ce should be OBJ_id_ce. 7238 Also some ip-pda OIDs in crypto/objects/objects.txt were 7239 incorrect (cf. RFC 3039). 7240 [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 7241 7242 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 7243 returns early because it has nothing to do. 7244 [Andy Schneider <andy.schneider@bjss.co.uk>] 7245 7246 *) [In 0.9.6c-engine release:] 7247 Fix mutex callback return values in crypto/engine/hw_ncipher.c. 7248 [Andy Schneider <andy.schneider@bjss.co.uk>] 7249 7250 *) [In 0.9.6c-engine release:] 7251 Add support for Cryptographic Appliance's keyserver technology. 7252 (Use engine 'keyclient') 7253 [Cryptographic Appliances and Geoff Thorpe] 7254 7255 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 7256 is called via tools/c89.sh because arguments have to be 7257 rearranged (all '-L' options must appear before the first object 7258 modules). 7259 [Richard Shapiro <rshapiro@abinitio.com>] 7260 7261 *) [In 0.9.6c-engine release:] 7262 Add support for Broadcom crypto accelerator cards, backported 7263 from 0.9.7. 7264 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 7265 7266 *) [In 0.9.6c-engine release:] 7267 Add support for SureWare crypto accelerator cards from 7268 Baltimore Technologies. (Use engine 'sureware') 7269 [Baltimore Technologies and Mark Cox] 7270 7271 *) [In 0.9.6c-engine release:] 7272 Add support for crypto accelerator cards from Accelerated 7273 Encryption Processing, www.aep.ie. (Use engine 'aep') 7274 [AEP Inc. and Mark Cox] 7275 7276 *) Add a configuration entry for gcc on UnixWare. 7277 [Gary Benson <gbenson@redhat.com>] 7278 7279 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 7280 messages are stored in a single piece (fixed-length part and 7281 variable-length part combined) and fix various bugs found on the way. 7282 [Bodo Moeller] 7283 7284 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 7285 instead. BIO_gethostbyname() does not know what timeouts are 7286 appropriate, so entries would stay in cache even when they have 7287 become invalid. 7288 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 7289 7290 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 7291 faced with a pathologically small ClientHello fragment that does 7292 not contain client_version: Instead of aborting with an error, 7293 simply choose the highest available protocol version (i.e., 7294 TLS 1.0 unless it is disabled). In practice, ClientHello 7295 messages are never sent like this, but this change gives us 7296 strictly correct behaviour at least for TLS. 7297 [Bodo Moeller] 7298 7299 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 7300 never resets s->method to s->ctx->method when called from within 7301 one of the SSL handshake functions. 7302 [Bodo Moeller; problem pointed out by Niko Baric] 7303 7304 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 7305 (sent using the client's version number) if client_version is 7306 smaller than the protocol version in use. Also change 7307 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 7308 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 7309 the client will at least see that alert. 7310 [Bodo Moeller] 7311 7312 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 7313 correctly. 7314 [Bodo Moeller] 7315 7316 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 7317 client receives HelloRequest while in a handshake. 7318 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 7319 7320 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 7321 should end in 'break', not 'goto end' which circuments various 7322 cleanups done in state SSL_ST_OK. But session related stuff 7323 must be disabled for SSL_ST_OK in the case that we just sent a 7324 HelloRequest. 7325 7326 Also avoid some overhead by not calling ssl_init_wbio_buffer() 7327 before just sending a HelloRequest. 7328 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 7329 7330 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 7331 reveal whether illegal block cipher padding was found or a MAC 7332 verification error occured. (Neither SSLerr() codes nor alerts 7333 are directly visible to potential attackers, but the information 7334 may leak via logfiles.) 7335 7336 Similar changes are not required for the SSL 2.0 implementation 7337 because the number of padding bytes is sent in clear for SSL 2.0, 7338 and the extra bytes are just ignored. However ssl/s2_pkt.c 7339 failed to verify that the purported number of padding bytes is in 7340 the legal range. 7341 [Bodo Moeller] 7342 7343 *) Add OpenUNIX-8 support including shared libraries 7344 (Boyd Lynn Gerber <gerberb@zenez.com>). 7345 [Lutz Jaenicke] 7346 7347 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 7348 'wristwatch attack' using huge encoding parameters (cf. 7349 James H. Manger's CRYPTO 2001 paper). Note that the 7350 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 7351 encoding parameters and hence was not vulnerable. 7352 [Bodo Moeller] 7353 7354 *) BN_sqr() bug fix. 7355 [Ulf M��ller, reported by Jim Ellis <jim.ellis@cavium.com>] 7356 7357 *) Rabin-Miller test analyses assume uniformly distributed witnesses, 7358 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 7359 followed by modular reduction. 7360 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 7361 7362 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 7363 equivalent based on BN_pseudo_rand() instead of BN_rand(). 7364 [Bodo Moeller] 7365 7366 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 7367 This function was broken, as the check for a new client hello message 7368 to handle SGC did not allow these large messages. 7369 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 7370 [Lutz Jaenicke] 7371 7372 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 7373 [Lutz Jaenicke] 7374 7375 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 7376 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 7377 [Lutz Jaenicke] 7378 7379 *) Rework the configuration and shared library support for Tru64 Unix. 7380 The configuration part makes use of modern compiler features and 7381 still retains old compiler behavior for those that run older versions 7382 of the OS. The shared library support part includes a variant that 7383 uses the RPATH feature, and is available through the special 7384 configuration target "alpha-cc-rpath", which will never be selected 7385 automatically. 7386 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 7387 7388 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 7389 with the same message size as in ssl3_get_certificate_request(). 7390 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 7391 messages might inadvertently be reject as too long. 7392 [Petr Lampa <lampa@fee.vutbr.cz>] 7393 7394 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 7395 [Andy Polyakov] 7396 7397 *) Modified SSL library such that the verify_callback that has been set 7398 specificly for an SSL object with SSL_set_verify() is actually being 7399 used. Before the change, a verify_callback set with this function was 7400 ignored and the verify_callback() set in the SSL_CTX at the time of 7401 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 7402 to allow the necessary settings. 7403 [Lutz Jaenicke] 7404 7405 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 7406 explicitly to NULL, as at least on Solaris 8 this seems not always to be 7407 done automatically (in contradiction to the requirements of the C 7408 standard). This made problems when used from OpenSSH. 7409 [Lutz Jaenicke] 7410 7411 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 7412 dh->length and always used 7413 7414 BN_rand_range(priv_key, dh->p). 7415 7416 BN_rand_range() is not necessary for Diffie-Hellman, and this 7417 specific range makes Diffie-Hellman unnecessarily inefficient if 7418 dh->length (recommended exponent length) is much smaller than the 7419 length of dh->p. We could use BN_rand_range() if the order of 7420 the subgroup was stored in the DH structure, but we only have 7421 dh->length. 7422 7423 So switch back to 7424 7425 BN_rand(priv_key, l, ...) 7426 7427 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 7428 otherwise. 7429 [Bodo Moeller] 7430 7431 *) In 7432 7433 RSA_eay_public_encrypt 7434 RSA_eay_private_decrypt 7435 RSA_eay_private_encrypt (signing) 7436 RSA_eay_public_decrypt (signature verification) 7437 7438 (default implementations for RSA_public_encrypt, 7439 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 7440 always reject numbers >= n. 7441 [Bodo Moeller] 7442 7443 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 7444 to synchronize access to 'locking_thread'. This is necessary on 7445 systems where access to 'locking_thread' (an 'unsigned long' 7446 variable) is not atomic. 7447 [Bodo Moeller] 7448 7449 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 7450 *before* setting the 'crypto_lock_rand' flag. The previous code had 7451 a race condition if 0 is a valid thread ID. 7452 [Travis Vitek <vitek@roguewave.com>] 7453 7454 *) Add support for shared libraries under Irix. 7455 [Albert Chin-A-Young <china@thewrittenword.com>] 7456 7457 *) Add configuration option to build on Linux on both big-endian and 7458 little-endian MIPS. 7459 [Ralf Baechle <ralf@uni-koblenz.de>] 7460 7461 *) Add the possibility to create shared libraries on HP-UX. 7462 [Richard Levitte] 7463 7464 Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 7465 7466 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 7467 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 7468 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 7469 PRNG state recovery was possible based on the output of 7470 one PRNG request appropriately sized to gain knowledge on 7471 'md' followed by enough consecutive 1-byte PRNG requests 7472 to traverse all of 'state'. 7473 7474 1. When updating 'md_local' (the current thread's copy of 'md') 7475 during PRNG output generation, hash all of the previous 7476 'md_local' value, not just the half used for PRNG output. 7477 7478 2. Make the number of bytes from 'state' included into the hash 7479 independent from the number of PRNG bytes requested. 7480 7481 The first measure alone would be sufficient to avoid 7482 Markku-Juhani's attack. (Actually it had never occurred 7483 to me that the half of 'md_local' used for chaining was the 7484 half from which PRNG output bytes were taken -- I had always 7485 assumed that the secret half would be used.) The second 7486 measure makes sure that additional data from 'state' is never 7487 mixed into 'md_local' in small portions; this heuristically 7488 further strengthens the PRNG. 7489 [Bodo Moeller] 7490 7491 *) Fix crypto/bn/asm/mips3.s. 7492 [Andy Polyakov] 7493 7494 *) When only the key is given to "enc", the IV is undefined. Print out 7495 an error message in this case. 7496 [Lutz Jaenicke] 7497 7498 *) Handle special case when X509_NAME is empty in X509 printing routines. 7499 [Steve Henson] 7500 7501 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 7502 positive and less than q. 7503 [Bodo Moeller] 7504 7505 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 7506 used: it isn't thread safe and the add_lock_callback should handle 7507 that itself. 7508 [Paul Rose <Paul.Rose@bridge.com>] 7509 7510 *) Verify that incoming data obeys the block size in 7511 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 7512 [Bodo Moeller] 7513 7514 *) Fix OAEP check. 7515 [Ulf M��ller, Bodo M��ller] 7516 7517 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 7518 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 7519 when fixing the server behaviour for backwards-compatible 'client 7520 hello' messages. (Note that the attack is impractical against 7521 SSL 3.0 and TLS 1.0 anyway because length and version checking 7522 means that the probability of guessing a valid ciphertext is 7523 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 7524 paper.) 7525 7526 Before 0.9.5, the countermeasure (hide the error by generating a 7527 random 'decryption result') did not work properly because 7528 ERR_clear_error() was missing, meaning that SSL_get_error() would 7529 detect the supposedly ignored error. 7530 7531 Both problems are now fixed. 7532 [Bodo Moeller] 7533 7534 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 7535 (previously it was 1024). 7536 [Bodo Moeller] 7537 7538 *) Fix for compatibility mode trust settings: ignore trust settings 7539 unless some valid trust or reject settings are present. 7540 [Steve Henson] 7541 7542 *) Fix for blowfish EVP: its a variable length cipher. 7543 [Steve Henson] 7544 7545 *) Fix various bugs related to DSA S/MIME verification. Handle missing 7546 parameters in DSA public key structures and return an error in the 7547 DSA routines if parameters are absent. 7548 [Steve Henson] 7549 7550 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 7551 in the current directory if neither $RANDFILE nor $HOME was set. 7552 RAND_file_name() in 0.9.6a returned NULL in this case. This has 7553 caused some confusion to Windows users who haven't defined $HOME. 7554 Thus RAND_file_name() is changed again: e_os.h can define a 7555 DEFAULT_HOME, which will be used if $HOME is not set. 7556 For Windows, we use "C:"; on other platforms, we still require 7557 environment variables. 7558 7559 *) Move 'if (!initialized) RAND_poll()' into regions protected by 7560 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 7561 having multiple threads call RAND_poll() concurrently. 7562 [Bodo Moeller] 7563 7564 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 7565 combination of a flag and a thread ID variable. 7566 Otherwise while one thread is in ssleay_rand_bytes (which sets the 7567 flag), *other* threads can enter ssleay_add_bytes without obeying 7568 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 7569 that they do not hold after the first thread unsets add_do_not_lock). 7570 [Bodo Moeller] 7571 7572 *) Change bctest again: '-x' expressions are not available in all 7573 versions of 'test'. 7574 [Bodo Moeller] 7575 7576 Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 7577 7578 *) Fix a couple of memory leaks in PKCS7_dataDecode() 7579 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 7580 7581 *) Change Configure and Makefiles to provide EXE_EXT, which will contain 7582 the default extension for executables, if any. Also, make the perl 7583 scripts that use symlink() to test if it really exists and use "cp" 7584 if it doesn't. All this made OpenSSL compilable and installable in 7585 CygWin. 7586 [Richard Levitte] 7587 7588 *) Fix for asn1_GetSequence() for indefinite length constructed data. 7589 If SEQUENCE is length is indefinite just set c->slen to the total 7590 amount of data available. 7591 [Steve Henson, reported by shige@FreeBSD.org] 7592 [This change does not apply to 0.9.7.] 7593 7594 *) Change bctest to avoid here-documents inside command substitution 7595 (workaround for FreeBSD /bin/sh bug). 7596 For compatibility with Ultrix, avoid shell functions (introduced 7597 in the bctest version that searches along $PATH). 7598 [Bodo Moeller] 7599 7600 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 7601 with des_encrypt() defined on some operating systems, like Solaris 7602 and UnixWare. 7603 [Richard Levitte] 7604 7605 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 7606 On the Importance of Eliminating Errors in Cryptographic 7607 Computations, J. Cryptology 14 (2001) 2, 101-119, 7608 http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 7609 [Ulf Moeller] 7610 7611 *) MIPS assembler BIGNUM division bug fix. 7612 [Andy Polyakov] 7613 7614 *) Disabled incorrect Alpha assembler code. 7615 [Richard Levitte] 7616 7617 *) Fix PKCS#7 decode routines so they correctly update the length 7618 after reading an EOC for the EXPLICIT tag. 7619 [Steve Henson] 7620 [This change does not apply to 0.9.7.] 7621 7622 *) Fix bug in PKCS#12 key generation routines. This was triggered 7623 if a 3DES key was generated with a 0 initial byte. Include 7624 PKCS12_BROKEN_KEYGEN compilation option to retain the old 7625 (but broken) behaviour. 7626 [Steve Henson] 7627 7628 *) Enhance bctest to search for a working bc along $PATH and print 7629 it when found. 7630 [Tim Rice <tim@multitalents.net> via Richard Levitte] 7631 7632 *) Fix memory leaks in err.c: free err_data string if necessary; 7633 don't write to the wrong index in ERR_set_error_data. 7634 [Bodo Moeller] 7635 7636 *) Implement ssl23_peek (analogous to ssl23_read), which previously 7637 did not exist. 7638 [Bodo Moeller] 7639 7640 *) Replace rdtsc with _emit statements for VC++ version 5. 7641 [Jeremy Cooper <jeremy@baymoo.org>] 7642 7643 *) Make it possible to reuse SSLv2 sessions. 7644 [Richard Levitte] 7645 7646 *) In copy_email() check for >= 0 as a return value for 7647 X509_NAME_get_index_by_NID() since 0 is a valid index. 7648 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 7649 7650 *) Avoid coredump with unsupported or invalid public keys by checking if 7651 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 7652 PKCS7_verify() fails with non detached data. 7653 [Steve Henson] 7654 7655 *) Don't use getenv in library functions when run as setuid/setgid. 7656 New function OPENSSL_issetugid(). 7657 [Ulf Moeller] 7658 7659 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 7660 due to incorrect handling of multi-threading: 7661 7662 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 7663 7664 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 7665 7666 3. Count how many times MemCheck_off() has been called so that 7667 nested use can be treated correctly. This also avoids 7668 inband-signalling in the previous code (which relied on the 7669 assumption that thread ID 0 is impossible). 7670 [Bodo Moeller] 7671 7672 *) Add "-rand" option also to s_client and s_server. 7673 [Lutz Jaenicke] 7674 7675 *) Fix CPU detection on Irix 6.x. 7676 [Kurt Hockenbury <khockenb@stevens-tech.edu> and 7677 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 7678 7679 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 7680 was empty. 7681 [Steve Henson] 7682 [This change does not apply to 0.9.7.] 7683 7684 *) Use the cached encoding of an X509_NAME structure rather than 7685 copying it. This is apparently the reason for the libsafe "errors" 7686 but the code is actually correct. 7687 [Steve Henson] 7688 7689 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 7690 Bleichenbacher's DSA attack. 7691 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 7692 to be set and top=0 forces the highest bit to be set; top=-1 is new 7693 and leaves the highest bit random. 7694 [Ulf Moeller, Bodo Moeller] 7695 7696 *) In the NCONF_...-based implementations for CONF_... queries 7697 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 7698 a temporary CONF structure with the data component set to NULL 7699 (which gives segmentation faults in lh_retrieve). 7700 Instead, use NULL for the CONF pointer in CONF_get_string and 7701 CONF_get_number (which may use environment variables) and directly 7702 return NULL from CONF_get_section. 7703 [Bodo Moeller] 7704 7705 *) Fix potential buffer overrun for EBCDIC. 7706 [Ulf Moeller] 7707 7708 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 7709 keyUsage if basicConstraints absent for a CA. 7710 [Steve Henson] 7711 7712 *) Make SMIME_write_PKCS7() write mail header values with a format that 7713 is more generally accepted (no spaces before the semicolon), since 7714 some programs can't parse those values properly otherwise. Also make 7715 sure BIO's that break lines after each write do not create invalid 7716 headers. 7717 [Richard Levitte] 7718 7719 *) Make the CRL encoding routines work with empty SEQUENCE OF. The 7720 macros previously used would not encode an empty SEQUENCE OF 7721 and break the signature. 7722 [Steve Henson] 7723 [This change does not apply to 0.9.7.] 7724 7725 *) Zero the premaster secret after deriving the master secret in 7726 DH ciphersuites. 7727 [Steve Henson] 7728 7729 *) Add some EVP_add_digest_alias registrations (as found in 7730 OpenSSL_add_all_digests()) to SSL_library_init() 7731 aka OpenSSL_add_ssl_algorithms(). This provides improved 7732 compatibility with peers using X.509 certificates 7733 with unconventional AlgorithmIdentifier OIDs. 7734 [Bodo Moeller] 7735 7736 *) Fix for Irix with NO_ASM. 7737 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 7738 7739 *) ./config script fixes. 7740 [Ulf Moeller, Richard Levitte] 7741 7742 *) Fix 'openssl passwd -1'. 7743 [Bodo Moeller] 7744 7745 *) Change PKCS12_key_gen_asc() so it can cope with non null 7746 terminated strings whose length is passed in the passlen 7747 parameter, for example from PEM callbacks. This was done 7748 by adding an extra length parameter to asc2uni(). 7749 [Steve Henson, reported by <oddissey@samsung.co.kr>] 7750 7751 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 7752 call failed, free the DSA structure. 7753 [Bodo Moeller] 7754 7755 *) Fix to uni2asc() to cope with zero length Unicode strings. 7756 These are present in some PKCS#12 files. 7757 [Steve Henson] 7758 7759 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 7760 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 7761 when writing a 32767 byte record. 7762 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 7763 7764 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 7765 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 7766 7767 (RSA objects have a reference count access to which is protected 7768 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 7769 so they are meant to be shared between threads.) 7770 [Bodo Moeller, Geoff Thorpe; original patch submitted by 7771 "Reddie, Steven" <Steven.Reddie@ca.com>] 7772 7773 *) Fix a deadlock in CRYPTO_mem_leaks(). 7774 [Bodo Moeller] 7775 7776 *) Use better test patterns in bntest. 7777 [Ulf M��ller] 7778 7779 *) rand_win.c fix for Borland C. 7780 [Ulf M��ller] 7781 7782 *) BN_rshift bugfix for n == 0. 7783 [Bodo Moeller] 7784 7785 *) Add a 'bctest' script that checks for some known 'bc' bugs 7786 so that 'make test' does not abort just because 'bc' is broken. 7787 [Bodo Moeller] 7788 7789 *) Store verify_result within SSL_SESSION also for client side to 7790 avoid potential security hole. (Re-used sessions on the client side 7791 always resulted in verify_result==X509_V_OK, not using the original 7792 result of the server certificate verification.) 7793 [Lutz Jaenicke] 7794 7795 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 7796 SSL3_RT_APPLICATION_DATA, return 0. 7797 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 7798 [Bodo Moeller] 7799 7800 *) Fix SSL_peek: 7801 Both ssl2_peek and ssl3_peek, which were totally broken in earlier 7802 releases, have been re-implemented by renaming the previous 7803 implementations of ssl2_read and ssl3_read to ssl2_read_internal 7804 and ssl3_read_internal, respectively, and adding 'peek' parameters 7805 to them. The new ssl[23]_{read,peek} functions are calls to 7806 ssl[23]_read_internal with the 'peek' flag set appropriately. 7807 A 'peek' parameter has also been added to ssl3_read_bytes, which 7808 does the actual work for ssl3_read_internal. 7809 [Bodo Moeller] 7810 7811 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 7812 the method-specific "init()" handler. Also clean up ex_data after 7813 calling the method-specific "finish()" handler. Previously, this was 7814 happening the other way round. 7815 [Geoff Thorpe] 7816 7817 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 7818 The previous value, 12, was not always sufficient for BN_mod_exp(). 7819 [Bodo Moeller] 7820 7821 *) Make sure that shared libraries get the internal name engine with 7822 the full version number and not just 0. This should mark the 7823 shared libraries as not backward compatible. Of course, this should 7824 be changed again when we can guarantee backward binary compatibility. 7825 [Richard Levitte] 7826 7827 *) Fix typo in get_cert_by_subject() in by_dir.c 7828 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 7829 7830 *) Rework the system to generate shared libraries: 7831 7832 - Make note of the expected extension for the shared libraries and 7833 if there is a need for symbolic links from for example libcrypto.so.0 7834 to libcrypto.so.0.9.7. There is extended info in Configure for 7835 that. 7836 7837 - Make as few rebuilds of the shared libraries as possible. 7838 7839 - Still avoid linking the OpenSSL programs with the shared libraries. 7840 7841 - When installing, install the shared libraries separately from the 7842 static ones. 7843 [Richard Levitte] 7844 7845 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 7846 7847 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 7848 and not in SSL_clear because the latter is also used by the 7849 accept/connect functions; previously, the settings made by 7850 SSL_set_read_ahead would be lost during the handshake. 7851 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 7852 7853 *) Correct util/mkdef.pl to be selective about disabled algorithms. 7854 Previously, it would create entries for disableed algorithms no 7855 matter what. 7856 [Richard Levitte] 7857 7858 *) Added several new manual pages for SSL_* function. 7859 [Lutz Jaenicke] 7860 7861 Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 7862 7863 *) In ssl23_get_client_hello, generate an error message when faced 7864 with an initial SSL 3.0/TLS record that is too small to contain the 7865 first two bytes of the ClientHello message, i.e. client_version. 7866 (Note that this is a pathologic case that probably has never happened 7867 in real life.) The previous approach was to use the version number 7868 from the record header as a substitute; but our protocol choice 7869 should not depend on that one because it is not authenticated 7870 by the Finished messages. 7871 [Bodo Moeller] 7872 7873 *) More robust randomness gathering functions for Windows. 7874 [Jeffrey Altman <jaltman@columbia.edu>] 7875 7876 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 7877 not set then we don't setup the error code for issuer check errors 7878 to avoid possibly overwriting other errors which the callback does 7879 handle. If an application does set the flag then we assume it knows 7880 what it is doing and can handle the new informational codes 7881 appropriately. 7882 [Steve Henson] 7883 7884 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 7885 a general "ANY" type, as such it should be able to decode anything 7886 including tagged types. However it didn't check the class so it would 7887 wrongly interpret tagged types in the same way as their universal 7888 counterpart and unknown types were just rejected. Changed so that the 7889 tagged and unknown types are handled in the same way as a SEQUENCE: 7890 that is the encoding is stored intact. There is also a new type 7891 "V_ASN1_OTHER" which is used when the class is not universal, in this 7892 case we have no idea what the actual type is so we just lump them all 7893 together. 7894 [Steve Henson] 7895 7896 *) On VMS, stdout may very well lead to a file that is written to 7897 in a record-oriented fashion. That means that every write() will 7898 write a separate record, which will be read separately by the 7899 programs trying to read from it. This can be very confusing. 7900 7901 The solution is to put a BIO filter in the way that will buffer 7902 text until a linefeed is reached, and then write everything a 7903 line at a time, so every record written will be an actual line, 7904 not chunks of lines and not (usually doesn't happen, but I've 7905 seen it once) several lines in one record. BIO_f_linebuffer() is 7906 the answer. 7907 7908 Currently, it's a VMS-only method, because that's where it has 7909 been tested well enough. 7910 [Richard Levitte] 7911 7912 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 7913 it can return incorrect results. 7914 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 7915 but it was in 0.9.6-beta[12].) 7916 [Bodo Moeller] 7917 7918 *) Disable the check for content being present when verifying detached 7919 signatures in pk7_smime.c. Some versions of Netscape (wrongly) 7920 include zero length content when signing messages. 7921 [Steve Henson] 7922 7923 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 7924 BIO_ctrl (for BIO pairs). 7925 [Bodo M��ller] 7926 7927 *) Add DSO method for VMS. 7928 [Richard Levitte] 7929 7930 *) Bug fix: Montgomery multiplication could produce results with the 7931 wrong sign. 7932 [Ulf M��ller] 7933 7934 *) Add RPM specification openssl.spec and modify it to build three 7935 packages. The default package contains applications, application 7936 documentation and run-time libraries. The devel package contains 7937 include files, static libraries and function documentation. The 7938 doc package contains the contents of the doc directory. The original 7939 openssl.spec was provided by Damien Miller <djm@mindrot.org>. 7940 [Richard Levitte] 7941 7942 *) Add a large number of documentation files for many SSL routines. 7943 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 7944 7945 *) Add a configuration entry for Sony News 4. 7946 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 7947 7948 *) Don't set the two most significant bits to one when generating a 7949 random number < q in the DSA library. 7950 [Ulf M��ller] 7951 7952 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 7953 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 7954 the underlying transport is blocking) if a handshake took place. 7955 (The default behaviour is needed by applications such as s_client 7956 and s_server that use select() to determine when to use SSL_read; 7957 but for applications that know in advance when to expect data, it 7958 just makes things more complicated.) 7959 [Bodo Moeller] 7960 7961 *) Add RAND_egd_bytes(), which gives control over the number of bytes read 7962 from EGD. 7963 [Ben Laurie] 7964 7965 *) Add a few more EBCDIC conditionals that make `req' and `x509' 7966 work better on such systems. 7967 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 7968 7969 *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 7970 Update PKCS12_parse() so it copies the friendlyName and the 7971 keyid to the certificates aux info. 7972 [Steve Henson] 7973 7974 *) Fix bug in PKCS7_verify() which caused an infinite loop 7975 if there was more than one signature. 7976 [Sven Uszpelkat <su@celocom.de>] 7977 7978 *) Major change in util/mkdef.pl to include extra information 7979 about each symbol, as well as presentig variables as well 7980 as functions. This change means that there's n more need 7981 to rebuild the .num files when some algorithms are excluded. 7982 [Richard Levitte] 7983 7984 *) Allow the verify time to be set by an application, 7985 rather than always using the current time. 7986 [Steve Henson] 7987 7988 *) Phase 2 verify code reorganisation. The certificate 7989 verify code now looks up an issuer certificate by a 7990 number of criteria: subject name, authority key id 7991 and key usage. It also verifies self signed certificates 7992 by the same criteria. The main comparison function is 7993 X509_check_issued() which performs these checks. 7994 7995 Lot of changes were necessary in order to support this 7996 without completely rewriting the lookup code. 7997 7998 Authority and subject key identifier are now cached. 7999 8000 The LHASH 'certs' is X509_STORE has now been replaced 8001 by a STACK_OF(X509_OBJECT). This is mainly because an 8002 LHASH can't store or retrieve multiple objects with 8003 the same hash value. 8004 8005 As a result various functions (which were all internal 8006 use only) have changed to handle the new X509_STORE 8007 structure. This will break anything that messed round 8008 with X509_STORE internally. 8009 8010 The functions X509_STORE_add_cert() now checks for an 8011 exact match, rather than just subject name. 8012 8013 The X509_STORE API doesn't directly support the retrieval 8014 of multiple certificates matching a given criteria, however 8015 this can be worked round by performing a lookup first 8016 (which will fill the cache with candidate certificates) 8017 and then examining the cache for matches. This is probably 8018 the best we can do without throwing out X509_LOOKUP 8019 entirely (maybe later...). 8020 8021 The X509_VERIFY_CTX structure has been enhanced considerably. 8022 8023 All certificate lookup operations now go via a get_issuer() 8024 callback. Although this currently uses an X509_STORE it 8025 can be replaced by custom lookups. This is a simple way 8026 to bypass the X509_STORE hackery necessary to make this 8027 work and makes it possible to use more efficient techniques 8028 in future. A very simple version which uses a simple 8029 STACK for its trusted certificate store is also provided 8030 using X509_STORE_CTX_trusted_stack(). 8031 8032 The verify_cb() and verify() callbacks now have equivalents 8033 in the X509_STORE_CTX structure. 8034 8035 X509_STORE_CTX also has a 'flags' field which can be used 8036 to customise the verify behaviour. 8037 [Steve Henson] 8038 8039 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 8040 excludes S/MIME capabilities. 8041 [Steve Henson] 8042 8043 *) When a certificate request is read in keep a copy of the 8044 original encoding of the signed data and use it when outputing 8045 again. Signatures then use the original encoding rather than 8046 a decoded, encoded version which may cause problems if the 8047 request is improperly encoded. 8048 [Steve Henson] 8049 8050 *) For consistency with other BIO_puts implementations, call 8051 buffer_write(b, ...) directly in buffer_puts instead of calling 8052 BIO_write(b, ...). 8053 8054 In BIO_puts, increment b->num_write as in BIO_write. 8055 [Peter.Sylvester@EdelWeb.fr] 8056 8057 *) Fix BN_mul_word for the case where the word is 0. (We have to use 8058 BN_zero, we may not return a BIGNUM with an array consisting of 8059 words set to zero.) 8060 [Bodo Moeller] 8061 8062 *) Avoid calling abort() from within the library when problems are 8063 detected, except if preprocessor symbols have been defined 8064 (such as REF_CHECK, BN_DEBUG etc.). 8065 [Bodo Moeller] 8066 8067 *) New openssl application 'rsautl'. This utility can be 8068 used for low level RSA operations. DER public key 8069 BIO/fp routines also added. 8070 [Steve Henson] 8071 8072 *) New Configure entry and patches for compiling on QNX 4. 8073 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 8074 8075 *) A demo state-machine implementation was sponsored by 8076 Nuron (http://www.nuron.com/) and is now available in 8077 demos/state_machine. 8078 [Ben Laurie] 8079 8080 *) New options added to the 'dgst' utility for signature 8081 generation and verification. 8082 [Steve Henson] 8083 8084 *) Unrecognized PKCS#7 content types are now handled via a 8085 catch all ASN1_TYPE structure. This allows unsupported 8086 types to be stored as a "blob" and an application can 8087 encode and decode it manually. 8088 [Steve Henson] 8089 8090 *) Fix various signed/unsigned issues to make a_strex.c 8091 compile under VC++. 8092 [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 8093 8094 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 8095 length if passed a buffer. ASN1_INTEGER_to_BN failed 8096 if passed a NULL BN and its argument was negative. 8097 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 8098 8099 *) Modification to PKCS#7 encoding routines to output definite 8100 length encoding. Since currently the whole structures are in 8101 memory there's not real point in using indefinite length 8102 constructed encoding. However if OpenSSL is compiled with 8103 the flag PKCS7_INDEFINITE_ENCODING the old form is used. 8104 [Steve Henson] 8105 8106 *) Added BIO_vprintf() and BIO_vsnprintf(). 8107 [Richard Levitte] 8108 8109 *) Added more prefixes to parse for in the the strings written 8110 through a logging bio, to cover all the levels that are available 8111 through syslog. The prefixes are now: 8112 8113 PANIC, EMERG, EMR => LOG_EMERG 8114 ALERT, ALR => LOG_ALERT 8115 CRIT, CRI => LOG_CRIT 8116 ERROR, ERR => LOG_ERR 8117 WARNING, WARN, WAR => LOG_WARNING 8118 NOTICE, NOTE, NOT => LOG_NOTICE 8119 INFO, INF => LOG_INFO 8120 DEBUG, DBG => LOG_DEBUG 8121 8122 and as before, if none of those prefixes are present at the 8123 beginning of the string, LOG_ERR is chosen. 8124 8125 On Win32, the LOG_* levels are mapped according to this: 8126 8127 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 8128 LOG_WARNING => EVENTLOG_WARNING_TYPE 8129 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 8130 8131 [Richard Levitte] 8132 8133 *) Made it possible to reconfigure with just the configuration 8134 argument "reconf" or "reconfigure". The command line arguments 8135 are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 8136 and are retrieved from there when reconfiguring. 8137 [Richard Levitte] 8138 8139 *) MD4 implemented. 8140 [Assar Westerlund <assar@sics.se>, Richard Levitte] 8141 8142 *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 8143 [Richard Levitte] 8144 8145 *) The obj_dat.pl script was messing up the sorting of object 8146 names. The reason was that it compared the quoted version 8147 of strings as a result "OCSP" > "OCSP Signing" because 8148 " > SPACE. Changed script to store unquoted versions of 8149 names and add quotes on output. It was also omitting some 8150 names from the lookup table if they were given a default 8151 value (that is if SN is missing it is given the same 8152 value as LN and vice versa), these are now added on the 8153 grounds that if an object has a name we should be able to 8154 look it up. Finally added warning output when duplicate 8155 short or long names are found. 8156 [Steve Henson] 8157 8158 *) Changes needed for Tandem NSK. 8159 [Scott Uroff <scott@xypro.com>] 8160 8161 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 8162 RSA_padding_check_SSLv23(), special padding was never detected 8163 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 8164 version rollback attacks was not effective. 8165 8166 In s23_clnt.c, don't use special rollback-attack detection padding 8167 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 8168 client; similarly, in s23_srvr.c, don't do the rollback check if 8169 SSL 2.0 is the only protocol enabled in the server. 8170 [Bodo Moeller] 8171 8172 *) Make it possible to get hexdumps of unprintable data with 'openssl 8173 asn1parse'. By implication, the functions ASN1_parse_dump() and 8174 BIO_dump_indent() are added. 8175 [Richard Levitte] 8176 8177 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 8178 these print out strings and name structures based on various 8179 flags including RFC2253 support and proper handling of 8180 multibyte characters. Added options to the 'x509' utility 8181 to allow the various flags to be set. 8182 [Steve Henson] 8183 8184 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 8185 Also change the functions X509_cmp_current_time() and 8186 X509_gmtime_adj() work with an ASN1_TIME structure, 8187 this will enable certificates using GeneralizedTime in validity 8188 dates to be checked. 8189 [Steve Henson] 8190 8191 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 8192 negative public key encodings) on by default, 8193 NO_NEG_PUBKEY_BUG can be set to disable it. 8194 [Steve Henson] 8195 8196 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 8197 content octets. An i2c_ASN1_OBJECT is unnecessary because 8198 the encoding can be trivially obtained from the structure. 8199 [Steve Henson] 8200 8201 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 8202 not read locks (CRYPTO_r_[un]lock). 8203 [Bodo Moeller] 8204 8205 *) A first attempt at creating official support for shared 8206 libraries through configuration. I've kept it so the 8207 default is static libraries only, and the OpenSSL programs 8208 are always statically linked for now, but there are 8209 preparations for dynamic linking in place. 8210 This has been tested on Linux and Tru64. 8211 [Richard Levitte] 8212 8213 *) Randomness polling function for Win9x, as described in: 8214 Peter Gutmann, Software Generation of Practically Strong 8215 Random Numbers. 8216 [Ulf M��ller] 8217 8218 *) Fix so PRNG is seeded in req if using an already existing 8219 DSA key. 8220 [Steve Henson] 8221 8222 *) New options to smime application. -inform and -outform 8223 allow alternative formats for the S/MIME message including 8224 PEM and DER. The -content option allows the content to be 8225 specified separately. This should allow things like Netscape 8226 form signing output easier to verify. 8227 [Steve Henson] 8228 8229 *) Fix the ASN1 encoding of tags using the 'long form'. 8230 [Steve Henson] 8231 8232 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 8233 STRING types. These convert content octets to and from the 8234 underlying type. The actual tag and length octets are 8235 already assumed to have been read in and checked. These 8236 are needed because all other string types have virtually 8237 identical handling apart from the tag. By having versions 8238 of the ASN1 functions that just operate on content octets 8239 IMPLICIT tagging can be handled properly. It also allows 8240 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 8241 and ASN1_INTEGER are identical apart from the tag. 8242 [Steve Henson] 8243 8244 *) Change the handling of OID objects as follows: 8245 8246 - New object identifiers are inserted in objects.txt, following 8247 the syntax given in objects.README. 8248 - objects.pl is used to process obj_mac.num and create a new 8249 obj_mac.h. 8250 - obj_dat.pl is used to create a new obj_dat.h, using the data in 8251 obj_mac.h. 8252 8253 This is currently kind of a hack, and the perl code in objects.pl 8254 isn't very elegant, but it works as I intended. The simplest way 8255 to check that it worked correctly is to look in obj_dat.h and 8256 check the array nid_objs and make sure the objects haven't moved 8257 around (this is important!). Additions are OK, as well as 8258 consistent name changes. 8259 [Richard Levitte] 8260 8261 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 8262 [Bodo Moeller] 8263 8264 *) Addition of the command line parameter '-rand file' to 'openssl req'. 8265 The given file adds to whatever has already been seeded into the 8266 random pool through the RANDFILE configuration file option or 8267 environment variable, or the default random state file. 8268 [Richard Levitte] 8269 8270 *) mkstack.pl now sorts each macro group into lexical order. 8271 Previously the output order depended on the order the files 8272 appeared in the directory, resulting in needless rewriting 8273 of safestack.h . 8274 [Steve Henson] 8275 8276 *) Patches to make OpenSSL compile under Win32 again. Mostly 8277 work arounds for the VC++ problem that it treats func() as 8278 func(void). Also stripped out the parts of mkdef.pl that 8279 added extra typesafe functions: these no longer exist. 8280 [Steve Henson] 8281 8282 *) Reorganisation of the stack code. The macros are now all 8283 collected in safestack.h . Each macro is defined in terms of 8284 a "stack macro" of the form SKM_<name>(type, a, b). The 8285 DEBUG_SAFESTACK is now handled in terms of function casts, 8286 this has the advantage of retaining type safety without the 8287 use of additional functions. If DEBUG_SAFESTACK is not defined 8288 then the non typesafe macros are used instead. Also modified the 8289 mkstack.pl script to handle the new form. Needs testing to see 8290 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 8291 the default if no major problems. Similar behaviour for ASN1_SET_OF 8292 and PKCS12_STACK_OF. 8293 [Steve Henson] 8294 8295 *) When some versions of IIS use the 'NET' form of private key the 8296 key derivation algorithm is different. Normally MD5(password) is 8297 used as a 128 bit RC4 key. In the modified case 8298 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 8299 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 8300 as the old Netscape_RSA functions except they have an additional 8301 'sgckey' parameter which uses the modified algorithm. Also added 8302 an -sgckey command line option to the rsa utility. Thanks to 8303 Adrian Peck <bertie@ncipher.com> for posting details of the modified 8304 algorithm to openssl-dev. 8305 [Steve Henson] 8306 8307 *) The evp_local.h macros were using 'c.##kname' which resulted in 8308 invalid expansion on some systems (SCO 5.0.5 for example). 8309 Corrected to 'c.kname'. 8310 [Phillip Porch <root@theporch.com>] 8311 8312 *) New X509_get1_email() and X509_REQ_get1_email() functions that return 8313 a STACK of email addresses from a certificate or request, these look 8314 in the subject name and the subject alternative name extensions and 8315 omit any duplicate addresses. 8316 [Steve Henson] 8317 8318 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 8319 This makes DSA verification about 2 % faster. 8320 [Bodo Moeller] 8321 8322 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 8323 (meaning that now 2^5 values will be precomputed, which is only 4 KB 8324 plus overhead for 1024 bit moduli). 8325 This makes exponentiations about 0.5 % faster for 1024 bit 8326 exponents (as measured by "openssl speed rsa2048"). 8327 [Bodo Moeller] 8328 8329 *) Rename memory handling macros to avoid conflicts with other 8330 software: 8331 Malloc => OPENSSL_malloc 8332 Malloc_locked => OPENSSL_malloc_locked 8333 Realloc => OPENSSL_realloc 8334 Free => OPENSSL_free 8335 [Richard Levitte] 8336 8337 *) New function BN_mod_exp_mont_word for small bases (roughly 15% 8338 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 8339 [Bodo Moeller] 8340 8341 *) CygWin32 support. 8342 [John Jarvie <jjarvie@newsguy.com>] 8343 8344 *) The type-safe stack code has been rejigged. It is now only compiled 8345 in when OpenSSL is configured with the DEBUG_SAFESTACK option and 8346 by default all type-specific stack functions are "#define"d back to 8347 standard stack functions. This results in more streamlined output 8348 but retains the type-safety checking possibilities of the original 8349 approach. 8350 [Geoff Thorpe] 8351 8352 *) The STACK code has been cleaned up, and certain type declarations 8353 that didn't make a lot of sense have been brought in line. This has 8354 also involved a cleanup of sorts in safestack.h to more correctly 8355 map type-safe stack functions onto their plain stack counterparts. 8356 This work has also resulted in a variety of "const"ifications of 8357 lots of the code, especially "_cmp" operations which should normally 8358 be prototyped with "const" parameters anyway. 8359 [Geoff Thorpe] 8360 8361 *) When generating bytes for the first time in md_rand.c, 'stir the pool' 8362 by seeding with STATE_SIZE dummy bytes (with zero entropy count). 8363 (The PRNG state consists of two parts, the large pool 'state' and 'md', 8364 where all of 'md' is used each time the PRNG is used, but 'state' 8365 is used only indexed by a cyclic counter. As entropy may not be 8366 well distributed from the beginning, 'md' is important as a 8367 chaining variable. However, the output function chains only half 8368 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 8369 all of 'md', and seeding with STATE_SIZE dummy bytes will result 8370 in all of 'state' being rewritten, with the new values depending 8371 on virtually all of 'md'. This overcomes the 80 bit limitation.) 8372 [Bodo Moeller] 8373 8374 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 8375 the handshake is continued after ssl_verify_cert_chain(); 8376 otherwise, if SSL_VERIFY_NONE is set, remaining error codes 8377 can lead to 'unexplainable' connection aborts later. 8378 [Bodo Moeller; problem tracked down by Lutz Jaenicke] 8379 8380 *) Major EVP API cipher revision. 8381 Add hooks for extra EVP features. This allows various cipher 8382 parameters to be set in the EVP interface. Support added for variable 8383 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 8384 setting of RC2 and RC5 parameters. 8385 8386 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 8387 ciphers. 8388 8389 Remove lots of duplicated code from the EVP library. For example *every* 8390 cipher init() function handles the 'iv' in the same way according to the 8391 cipher mode. They also all do nothing if the 'key' parameter is NULL and 8392 for CFB and OFB modes they zero ctx->num. 8393 8394 New functionality allows removal of S/MIME code RC2 hack. 8395 8396 Most of the routines have the same form and so can be declared in terms 8397 of macros. 8398 8399 By shifting this to the top level EVP_CipherInit() it can be removed from 8400 all individual ciphers. If the cipher wants to handle IVs or keys 8401 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 8402 flags. 8403 8404 Change lots of functions like EVP_EncryptUpdate() to now return a 8405 value: although software versions of the algorithms cannot fail 8406 any installed hardware versions can. 8407 [Steve Henson] 8408 8409 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 8410 this option is set, tolerate broken clients that send the negotiated 8411 protocol version number instead of the requested protocol version 8412 number. 8413 [Bodo Moeller] 8414 8415 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 8416 i.e. non-zero for export ciphersuites, zero otherwise. 8417 Previous versions had this flag inverted, inconsistent with 8418 rsa_tmp_cb (..._TMP_RSA_CB). 8419 [Bodo Moeller; problem reported by Amit Chopra] 8420 8421 *) Add missing DSA library text string. Work around for some IIS 8422 key files with invalid SEQUENCE encoding. 8423 [Steve Henson] 8424 8425 *) Add a document (doc/standards.txt) that list all kinds of standards 8426 and so on that are implemented in OpenSSL. 8427 [Richard Levitte] 8428 8429 *) Enhance c_rehash script. Old version would mishandle certificates 8430 with the same subject name hash and wouldn't handle CRLs at all. 8431 Added -fingerprint option to crl utility, to support new c_rehash 8432 features. 8433 [Steve Henson] 8434 8435 *) Eliminate non-ANSI declarations in crypto.h and stack.h. 8436 [Ulf M��ller] 8437 8438 *) Fix for SSL server purpose checking. Server checking was 8439 rejecting certificates which had extended key usage present 8440 but no ssl client purpose. 8441 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 8442 8443 *) Make PKCS#12 code work with no password. The PKCS#12 spec 8444 is a little unclear about how a blank password is handled. 8445 Since the password in encoded as a BMPString with terminating 8446 double NULL a zero length password would end up as just the 8447 double NULL. However no password at all is different and is 8448 handled differently in the PKCS#12 key generation code. NS 8449 treats a blank password as zero length. MSIE treats it as no 8450 password on export: but it will try both on import. We now do 8451 the same: PKCS12_parse() tries zero length and no password if 8452 the password is set to "" or NULL (NULL is now a valid password: 8453 it wasn't before) as does the pkcs12 application. 8454 [Steve Henson] 8455 8456 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 8457 perror when PEM_read_bio_X509_REQ fails, the error message must 8458 be obtained from the error queue. 8459 [Bodo Moeller] 8460 8461 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 8462 it in ERR_remove_state if appropriate, and change ERR_get_state 8463 accordingly to avoid race conditions (this is necessary because 8464 thread_hash is no longer constant once set). 8465 [Bodo Moeller] 8466 8467 *) Bugfix for linux-elf makefile.one. 8468 [Ulf M��ller] 8469 8470 *) RSA_get_default_method() will now cause a default 8471 RSA_METHOD to be chosen if one doesn't exist already. 8472 Previously this was only set during a call to RSA_new() 8473 or RSA_new_method(NULL) meaning it was possible for 8474 RSA_get_default_method() to return NULL. 8475 [Geoff Thorpe] 8476 8477 *) Added native name translation to the existing DSO code 8478 that will convert (if the flag to do so is set) filenames 8479 that are sufficiently small and have no path information 8480 into a canonical native form. Eg. "blah" converted to 8481 "libblah.so" or "blah.dll" etc. 8482 [Geoff Thorpe] 8483 8484 *) New function ERR_error_string_n(e, buf, len) which is like 8485 ERR_error_string(e, buf), but writes at most 'len' bytes 8486 including the 0 terminator. For ERR_error_string_n, 'buf' 8487 may not be NULL. 8488 [Damien Miller <djm@mindrot.org>, Bodo Moeller] 8489 8490 *) CONF library reworked to become more general. A new CONF 8491 configuration file reader "class" is implemented as well as a 8492 new functions (NCONF_*, for "New CONF") to handle it. The now 8493 old CONF_* functions are still there, but are reimplemented to 8494 work in terms of the new functions. Also, a set of functions 8495 to handle the internal storage of the configuration data is 8496 provided to make it easier to write new configuration file 8497 reader "classes" (I can definitely see something reading a 8498 configuration file in XML format, for example), called _CONF_*, 8499 or "the configuration storage API"... 8500 8501 The new configuration file reading functions are: 8502 8503 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 8504 NCONF_get_section, NCONF_get_string, NCONF_get_numbre 8505 8506 NCONF_default, NCONF_WIN32 8507 8508 NCONF_dump_fp, NCONF_dump_bio 8509 8510 NCONF_default and NCONF_WIN32 are method (or "class") choosers, 8511 NCONF_new creates a new CONF object. This works in the same way 8512 as other interfaces in OpenSSL, like the BIO interface. 8513 NCONF_dump_* dump the internal storage of the configuration file, 8514 which is useful for debugging. All other functions take the same 8515 arguments as the old CONF_* functions wth the exception of the 8516 first that must be a `CONF *' instead of a `LHASH *'. 8517 8518 To make it easer to use the new classes with the old CONF_* functions, 8519 the function CONF_set_default_method is provided. 8520 [Richard Levitte] 8521 8522 *) Add '-tls1' option to 'openssl ciphers', which was already 8523 mentioned in the documentation but had not been implemented. 8524 (This option is not yet really useful because even the additional 8525 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 8526 [Bodo Moeller] 8527 8528 *) Initial DSO code added into libcrypto for letting OpenSSL (and 8529 OpenSSL-based applications) load shared libraries and bind to 8530 them in a portable way. 8531 [Geoff Thorpe, with contributions from Richard Levitte] 8532 8533 Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 8534 8535 *) Make sure _lrotl and _lrotr are only used with MSVC. 8536 8537 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 8538 (the default implementation of RAND_status). 8539 8540 *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 8541 to '-clrext' (= clear extensions), as intended and documented. 8542 [Bodo Moeller; inconsistency pointed out by Michael Attili 8543 <attili@amaxo.com>] 8544 8545 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 8546 was larger than the MD block size. 8547 [Steve Henson, pointed out by Yost William <YostW@tce.com>] 8548 8549 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 8550 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 8551 using the passed key: if the passed key was a private key the result 8552 of X509_print(), for example, would be to print out all the private key 8553 components. 8554 [Steve Henson] 8555 8556 *) des_quad_cksum() byte order bug fix. 8557 [Ulf M��ller, using the problem description in krb4-0.9.7, where 8558 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 8559 8560 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 8561 discouraged. 8562 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 8563 8564 *) For easily testing in shell scripts whether some command 8565 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 8566 returns with exit code 0 iff no command of the given name is available. 8567 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 8568 the output goes to stdout and nothing is printed to stderr. 8569 Additional arguments are always ignored. 8570 8571 Since for each cipher there is a command of the same name, 8572 the 'no-cipher' compilation switches can be tested this way. 8573 8574 ('openssl no-XXX' is not able to detect pseudo-commands such 8575 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 8576 [Bodo Moeller] 8577 8578 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 8579 [Bodo Moeller] 8580 8581 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 8582 is set; it will be thrown away anyway because each handshake creates 8583 its own key. 8584 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 8585 to parameters -- in previous versions (since OpenSSL 0.9.3) the 8586 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 8587 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 8588 [Bodo Moeller] 8589 8590 *) New s_client option -ign_eof: EOF at stdin is ignored, and 8591 'Q' and 'R' lose their special meanings (quit/renegotiate). 8592 This is part of what -quiet does; unlike -quiet, -ign_eof 8593 does not suppress any output. 8594 [Richard Levitte] 8595 8596 *) Add compatibility options to the purpose and trust code. The 8597 purpose X509_PURPOSE_ANY is "any purpose" which automatically 8598 accepts a certificate or CA, this was the previous behaviour, 8599 with all the associated security issues. 8600 8601 X509_TRUST_COMPAT is the old trust behaviour: only and 8602 automatically trust self signed roots in certificate store. A 8603 new trust setting X509_TRUST_DEFAULT is used to specify that 8604 a purpose has no associated trust setting and it should instead 8605 use the value in the default purpose. 8606 [Steve Henson] 8607 8608 *) Fix the PKCS#8 DSA private key code so it decodes keys again 8609 and fix a memory leak. 8610 [Steve Henson] 8611 8612 *) In util/mkerr.pl (which implements 'make errors'), preserve 8613 reason strings from the previous version of the .c file, as 8614 the default to have only downcase letters (and digits) in 8615 automatically generated reasons codes is not always appropriate. 8616 [Bodo Moeller] 8617 8618 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 8619 using strerror. Previously, ERR_reason_error_string() returned 8620 library names as reason strings for SYSerr; but SYSerr is a special 8621 case where small numbers are errno values, not library numbers. 8622 [Bodo Moeller] 8623 8624 *) Add '-dsaparam' option to 'openssl dhparam' application. This 8625 converts DSA parameters into DH parameters. (When creating parameters, 8626 DSA_generate_parameters is used.) 8627 [Bodo Moeller] 8628 8629 *) Include 'length' (recommended exponent length) in C code generated 8630 by 'openssl dhparam -C'. 8631 [Bodo Moeller] 8632 8633 *) The second argument to set_label in perlasm was already being used 8634 so couldn't be used as a "file scope" flag. Moved to third argument 8635 which was free. 8636 [Steve Henson] 8637 8638 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 8639 instead of RAND_bytes for encryption IVs and salts. 8640 [Bodo Moeller] 8641 8642 *) Include RAND_status() into RAND_METHOD instead of implementing 8643 it only for md_rand.c Otherwise replacing the PRNG by calling 8644 RAND_set_rand_method would be impossible. 8645 [Bodo Moeller] 8646 8647 *) Don't let DSA_generate_key() enter an infinite loop if the random 8648 number generation fails. 8649 [Bodo Moeller] 8650 8651 *) New 'rand' application for creating pseudo-random output. 8652 [Bodo Moeller] 8653 8654 *) Added configuration support for Linux/IA64 8655 [Rolf Haberrecker <rolf@suse.de>] 8656 8657 *) Assembler module support for Mingw32. 8658 [Ulf M��ller] 8659 8660 *) Shared library support for HPUX (in shlib/). 8661 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 8662 8663 *) Shared library support for Solaris gcc. 8664 [Lutz Behnke <behnke@trustcenter.de>] 8665 8666 Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 8667 8668 *) PKCS7_encrypt() was adding text MIME headers twice because they 8669 were added manually and by SMIME_crlf_copy(). 8670 [Steve Henson] 8671 8672 *) In bntest.c don't call BN_rand with zero bits argument. 8673 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 8674 8675 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 8676 case was implemented. This caused BN_div_recp() to fail occasionally. 8677 [Ulf M��ller] 8678 8679 *) Add an optional second argument to the set_label() in the perl 8680 assembly language builder. If this argument exists and is set 8681 to 1 it signals that the assembler should use a symbol whose 8682 scope is the entire file, not just the current function. This 8683 is needed with MASM which uses the format label:: for this scope. 8684 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 8685 8686 *) Change the ASN1 types so they are typedefs by default. Before 8687 almost all types were #define'd to ASN1_STRING which was causing 8688 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 8689 for example. 8690 [Steve Henson] 8691 8692 *) Change names of new functions to the new get1/get0 naming 8693 convention: After 'get1', the caller owns a reference count 8694 and has to call ..._free; 'get0' returns a pointer to some 8695 data structure without incrementing reference counters. 8696 (Some of the existing 'get' functions increment a reference 8697 counter, some don't.) 8698 Similarly, 'set1' and 'add1' functions increase reference 8699 counters or duplicate objects. 8700 [Steve Henson] 8701 8702 *) Allow for the possibility of temp RSA key generation failure: 8703 the code used to assume it always worked and crashed on failure. 8704 [Steve Henson] 8705 8706 *) Fix potential buffer overrun problem in BIO_printf(). 8707 [Ulf M��ller, using public domain code by Patrick Powell; problem 8708 pointed out by David Sacerdote <das33@cornell.edu>] 8709 8710 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 8711 RAND_egd() and RAND_status(). In the command line application, 8712 the EGD socket can be specified like a seed file using RANDFILE 8713 or -rand. 8714 [Ulf M��ller] 8715 8716 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 8717 Some CAs (e.g. Verisign) distribute certificates in this form. 8718 [Steve Henson] 8719 8720 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 8721 list to exclude them. This means that no special compilation option 8722 is needed to use anonymous DH: it just needs to be included in the 8723 cipher list. 8724 [Steve Henson] 8725 8726 *) Change the EVP_MD_CTX_type macro so its meaning consistent with 8727 EVP_MD_type. The old functionality is available in a new macro called 8728 EVP_MD_md(). Change code that uses it and update docs. 8729 [Steve Henson] 8730 8731 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 8732 where the 'void *' argument is replaced by a function pointer argument. 8733 Previously 'void *' was abused to point to functions, which works on 8734 many platforms, but is not correct. As these functions are usually 8735 called by macros defined in OpenSSL header files, most source code 8736 should work without changes. 8737 [Richard Levitte] 8738 8739 *) <openssl/opensslconf.h> (which is created by Configure) now contains 8740 sections with information on -D... compiler switches used for 8741 compiling the library so that applications can see them. To enable 8742 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 8743 must be defined. E.g., 8744 #define OPENSSL_ALGORITHM_DEFINES 8745 #include <openssl/opensslconf.h> 8746 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 8747 [Richard Levitte, Ulf and Bodo M��ller] 8748 8749 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 8750 record layer. 8751 [Bodo Moeller] 8752 8753 *) Change the 'other' type in certificate aux info to a STACK_OF 8754 X509_ALGOR. Although not an AlgorithmIdentifier as such it has 8755 the required ASN1 format: arbitrary types determined by an OID. 8756 [Steve Henson] 8757 8758 *) Add some PEM_write_X509_REQ_NEW() functions and a command line 8759 argument to 'req'. This is not because the function is newer or 8760 better than others it just uses the work 'NEW' in the certificate 8761 request header lines. Some software needs this. 8762 [Steve Henson] 8763 8764 *) Reorganise password command line arguments: now passwords can be 8765 obtained from various sources. Delete the PEM_cb function and make 8766 it the default behaviour: i.e. if the callback is NULL and the 8767 usrdata argument is not NULL interpret it as a null terminated pass 8768 phrase. If usrdata and the callback are NULL then the pass phrase 8769 is prompted for as usual. 8770 [Steve Henson] 8771 8772 *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 8773 the support is automatically enabled. The resulting binaries will 8774 autodetect the card and use it if present. 8775 [Ben Laurie and Compaq Inc.] 8776 8777 *) Work around for Netscape hang bug. This sends certificate request 8778 and server done in one record. Since this is perfectly legal in the 8779 SSL/TLS protocol it isn't a "bug" option and is on by default. See 8780 the bugs/SSLv3 entry for more info. 8781 [Steve Henson] 8782 8783 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 8784 [Andy Polyakov] 8785 8786 *) Add -rand argument to smime and pkcs12 applications and read/write 8787 of seed file. 8788 [Steve Henson] 8789 8790 *) New 'passwd' tool for crypt(3) and apr1 password hashes. 8791 [Bodo Moeller] 8792 8793 *) Add command line password options to the remaining applications. 8794 [Steve Henson] 8795 8796 *) Bug fix for BN_div_recp() for numerators with an even number of 8797 bits. 8798 [Ulf M��ller] 8799 8800 *) More tests in bntest.c, and changed test_bn output. 8801 [Ulf M��ller] 8802 8803 *) ./config recognizes MacOS X now. 8804 [Andy Polyakov] 8805 8806 *) Bug fix for BN_div() when the first words of num and divsor are 8807 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 8808 [Ulf M��ller] 8809 8810 *) Add support for various broken PKCS#8 formats, and command line 8811 options to produce them. 8812 [Steve Henson] 8813 8814 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 8815 get temporary BIGNUMs from a BN_CTX. 8816 [Ulf M��ller] 8817 8818 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 8819 for p == 0. 8820 [Ulf M��ller] 8821 8822 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 8823 include a #define from the old name to the new. The original intent 8824 was that statically linked binaries could for example just call 8825 SSLeay_add_all_ciphers() to just add ciphers to the table and not 8826 link with digests. This never worked becayse SSLeay_add_all_digests() 8827 and SSLeay_add_all_ciphers() were in the same source file so calling 8828 one would link with the other. They are now in separate source files. 8829 [Steve Henson] 8830 8831 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 8832 [Steve Henson] 8833 8834 *) Use a less unusual form of the Miller-Rabin primality test (it used 8835 a binary algorithm for exponentiation integrated into the Miller-Rabin 8836 loop, our standard modexp algorithms are faster). 8837 [Bodo Moeller] 8838 8839 *) Support for the EBCDIC character set completed. 8840 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 8841 8842 *) Source code cleanups: use const where appropriate, eliminate casts, 8843 use void * instead of char * in lhash. 8844 [Ulf M��ller] 8845 8846 *) Bugfix: ssl3_send_server_key_exchange was not restartable 8847 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 8848 this the server could overwrite ephemeral keys that the client 8849 has already seen). 8850 [Bodo Moeller] 8851 8852 *) Turn DSA_is_prime into a macro that calls BN_is_prime, 8853 using 50 iterations of the Rabin-Miller test. 8854 8855 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 8856 iterations of the Rabin-Miller test as required by the appendix 8857 to FIPS PUB 186[-1]) instead of DSA_is_prime. 8858 As BN_is_prime_fasttest includes trial division, DSA parameter 8859 generation becomes much faster. 8860 8861 This implies a change for the callback functions in DSA_is_prime 8862 and DSA_generate_parameters: The callback function is called once 8863 for each positive witness in the Rabin-Miller test, not just 8864 occasionally in the inner loop; and the parameters to the 8865 callback function now provide an iteration count for the outer 8866 loop rather than for the current invocation of the inner loop. 8867 DSA_generate_parameters additionally can call the callback 8868 function with an 'iteration count' of -1, meaning that a 8869 candidate has passed the trial division test (when q is generated 8870 from an application-provided seed, trial division is skipped). 8871 [Bodo Moeller] 8872 8873 *) New function BN_is_prime_fasttest that optionally does trial 8874 division before starting the Rabin-Miller test and has 8875 an additional BN_CTX * argument (whereas BN_is_prime always 8876 has to allocate at least one BN_CTX). 8877 'callback(1, -1, cb_arg)' is called when a number has passed the 8878 trial division stage. 8879 [Bodo Moeller] 8880 8881 *) Fix for bug in CRL encoding. The validity dates weren't being handled 8882 as ASN1_TIME. 8883 [Steve Henson] 8884 8885 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 8886 [Steve Henson] 8887 8888 *) New function BN_pseudo_rand(). 8889 [Ulf M��ller] 8890 8891 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 8892 bignum version of BN_from_montgomery() with the working code from 8893 SSLeay 0.9.0 (the word based version is faster anyway), and clean up 8894 the comments. 8895 [Ulf M��ller] 8896 8897 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 8898 made it impossible to use the same SSL_SESSION data structure in 8899 SSL2 clients in multiple threads. 8900 [Bodo Moeller] 8901 8902 *) The return value of RAND_load_file() no longer counts bytes obtained 8903 by stat(). RAND_load_file(..., -1) is new and uses the complete file 8904 to seed the PRNG (previously an explicit byte count was required). 8905 [Ulf M��ller, Bodo M��ller] 8906 8907 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 8908 used (char *) instead of (void *) and had casts all over the place. 8909 [Steve Henson] 8910 8911 *) Make BN_generate_prime() return NULL on error if ret!=NULL. 8912 [Ulf M��ller] 8913 8914 *) Retain source code compatibility for BN_prime_checks macro: 8915 BN_is_prime(..., BN_prime_checks, ...) now uses 8916 BN_prime_checks_for_size to determine the appropriate number of 8917 Rabin-Miller iterations. 8918 [Ulf M��ller] 8919 8920 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 8921 DH_CHECK_P_NOT_SAFE_PRIME. 8922 (Check if this is true? OpenPGP calls them "strong".) 8923 [Ulf M��ller] 8924 8925 *) Merge the functionality of "dh" and "gendh" programs into a new program 8926 "dhparam". The old programs are retained for now but will handle DH keys 8927 (instead of parameters) in future. 8928 [Steve Henson] 8929 8930 *) Make the ciphers, s_server and s_client programs check the return values 8931 when a new cipher list is set. 8932 [Steve Henson] 8933 8934 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 8935 ciphers. Before when the 56bit ciphers were enabled the sorting was 8936 wrong. 8937 8938 The syntax for the cipher sorting has been extended to support sorting by 8939 cipher-strength (using the strength_bits hard coded in the tables). 8940 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 8941 8942 Fix a bug in the cipher-command parser: when supplying a cipher command 8943 string with an "undefined" symbol (neither command nor alphanumeric 8944 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 8945 an error is flagged. 8946 8947 Due to the strength-sorting extension, the code of the 8948 ssl_create_cipher_list() function was completely rearranged. I hope that 8949 the readability was also increased :-) 8950 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 8951 8952 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 8953 for the first serial number and places 2 in the serial number file. This 8954 avoids problems when the root CA is created with serial number zero and 8955 the first user certificate has the same issuer name and serial number 8956 as the root CA. 8957 [Steve Henson] 8958 8959 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 8960 the new code. Add documentation for this stuff. 8961 [Steve Henson] 8962 8963 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 8964 X509_*() to X509at_*() on the grounds that they don't handle X509 8965 structures and behave in an analagous way to the X509v3 functions: 8966 they shouldn't be called directly but wrapper functions should be used 8967 instead. 8968 8969 So we also now have some wrapper functions that call the X509at functions 8970 when passed certificate requests. (TO DO: similar things can be done with 8971 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 8972 things. Some of these need some d2i or i2d and print functionality 8973 because they handle more complex structures.) 8974 [Steve Henson] 8975 8976 *) Add missing #ifndefs that caused missing symbols when building libssl 8977 as a shared library without RSA. Use #ifndef NO_SSL2 instead of 8978 NO_RSA in ssl/s2*.c. 8979 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M��ller] 8980 8981 *) Precautions against using the PRNG uninitialized: RAND_bytes() now 8982 has a return value which indicates the quality of the random data 8983 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 8984 error queue. New function RAND_pseudo_bytes() generates output that is 8985 guaranteed to be unique but not unpredictable. RAND_add is like 8986 RAND_seed, but takes an extra argument for an entropy estimate 8987 (RAND_seed always assumes full entropy). 8988 [Ulf M��ller] 8989 8990 *) Do more iterations of Rabin-Miller probable prime test (specifically, 8991 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 8992 instead of only 2 for all lengths; see BN_prime_checks_for_size definition 8993 in crypto/bn/bn_prime.c for the complete table). This guarantees a 8994 false-positive rate of at most 2^-80 for random input. 8995 [Bodo Moeller] 8996 8997 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 8998 [Bodo Moeller] 8999 9000 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 9001 in the 0.9.5 release), this returns the chain 9002 from an X509_CTX structure with a dup of the stack and all 9003 the X509 reference counts upped: so the stack will exist 9004 after X509_CTX_cleanup() has been called. Modify pkcs12.c 9005 to use this. 9006 9007 Also make SSL_SESSION_print() print out the verify return 9008 code. 9009 [Steve Henson] 9010 9011 *) Add manpage for the pkcs12 command. Also change the default 9012 behaviour so MAC iteration counts are used unless the new 9013 -nomaciter option is used. This improves file security and 9014 only older versions of MSIE (4.0 for example) need it. 9015 [Steve Henson] 9016 9017 *) Honor the no-xxx Configure options when creating .DEF files. 9018 [Ulf M��ller] 9019 9020 *) Add PKCS#10 attributes to field table: challengePassword, 9021 unstructuredName and unstructuredAddress. These are taken from 9022 draft PKCS#9 v2.0 but are compatible with v1.2 provided no 9023 international characters are used. 9024 9025 More changes to X509_ATTRIBUTE code: allow the setting of types 9026 based on strings. Remove the 'loc' parameter when adding 9027 attributes because these will be a SET OF encoding which is sorted 9028 in ASN1 order. 9029 [Steve Henson] 9030 9031 *) Initial changes to the 'req' utility to allow request generation 9032 automation. This will allow an application to just generate a template 9033 file containing all the field values and have req construct the 9034 request. 9035 9036 Initial support for X509_ATTRIBUTE handling. Stacks of these are 9037 used all over the place including certificate requests and PKCS#7 9038 structures. They are currently handled manually where necessary with 9039 some primitive wrappers for PKCS#7. The new functions behave in a 9040 manner analogous to the X509 extension functions: they allow 9041 attributes to be looked up by NID and added. 9042 9043 Later something similar to the X509V3 code would be desirable to 9044 automatically handle the encoding, decoding and printing of the 9045 more complex types. The string types like challengePassword can 9046 be handled by the string table functions. 9047 9048 Also modified the multi byte string table handling. Now there is 9049 a 'global mask' which masks out certain types. The table itself 9050 can use the flag STABLE_NO_MASK to ignore the mask setting: this 9051 is useful when for example there is only one permissible type 9052 (as in countryName) and using the mask might result in no valid 9053 types at all. 9054 [Steve Henson] 9055 9056 *) Clean up 'Finished' handling, and add functions SSL_get_finished and 9057 SSL_get_peer_finished to allow applications to obtain the latest 9058 Finished messages sent to the peer or expected from the peer, 9059 respectively. (SSL_get_peer_finished is usually the Finished message 9060 actually received from the peer, otherwise the protocol will be aborted.) 9061 9062 As the Finished message are message digests of the complete handshake 9063 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 9064 be used for external authentication procedures when the authentication 9065 provided by SSL/TLS is not desired or is not enough. 9066 [Bodo Moeller] 9067 9068 *) Enhanced support for Alpha Linux is added. Now ./config checks if 9069 the host supports BWX extension and if Compaq C is present on the 9070 $PATH. Just exploiting of the BWX extension results in 20-30% 9071 performance kick for some algorithms, e.g. DES and RC4 to mention 9072 a couple. Compaq C in turn generates ~20% faster code for MD5 and 9073 SHA1. 9074 [Andy Polyakov] 9075 9076 *) Add support for MS "fast SGC". This is arguably a violation of the 9077 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 9078 weak crypto and after checking the certificate is SGC a second one 9079 with strong crypto. MS SGC stops the first handshake after receiving 9080 the server certificate message and sends a second client hello. Since 9081 a server will typically do all the time consuming operations before 9082 expecting any further messages from the client (server key exchange 9083 is the most expensive) there is little difference between the two. 9084 9085 To get OpenSSL to support MS SGC we have to permit a second client 9086 hello message after we have sent server done. In addition we have to 9087 reset the MAC if we do get this second client hello. 9088 [Steve Henson] 9089 9090 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 9091 if a DER encoded private key is RSA or DSA traditional format. Changed 9092 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 9093 format DER encoded private key. Newer code should use PKCS#8 format which 9094 has the key type encoded in the ASN1 structure. Added DER private key 9095 support to pkcs8 application. 9096 [Steve Henson] 9097 9098 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 9099 ciphersuites has been selected (as required by the SSL 3/TLS 1 9100 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 9101 is set, we interpret this as a request to violate the specification 9102 (the worst that can happen is a handshake failure, and 'correct' 9103 behaviour would result in a handshake failure anyway). 9104 [Bodo Moeller] 9105 9106 *) In SSL_CTX_add_session, take into account that there might be multiple 9107 SSL_SESSION structures with the same session ID (e.g. when two threads 9108 concurrently obtain them from an external cache). 9109 The internal cache can handle only one SSL_SESSION with a given ID, 9110 so if there's a conflict, we now throw out the old one to achieve 9111 consistency. 9112 [Bodo Moeller] 9113 9114 *) Add OIDs for idea and blowfish in CBC mode. This will allow both 9115 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 9116 some routines that use cipher OIDs: some ciphers do not have OIDs 9117 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 9118 example. 9119 [Steve Henson] 9120 9121 *) Simplify the trust setting structure and code. Now we just have 9122 two sequences of OIDs for trusted and rejected settings. These will 9123 typically have values the same as the extended key usage extension 9124 and any application specific purposes. 9125 9126 The trust checking code now has a default behaviour: it will just 9127 check for an object with the same NID as the passed id. Functions can 9128 be provided to override either the default behaviour or the behaviour 9129 for a given id. SSL client, server and email already have functions 9130 in place for compatibility: they check the NID and also return "trusted" 9131 if the certificate is self signed. 9132 [Steve Henson] 9133 9134 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 9135 traditional format into an EVP_PKEY structure. 9136 [Steve Henson] 9137 9138 *) Add a password callback function PEM_cb() which either prompts for 9139 a password if usr_data is NULL or otherwise assumes it is a null 9140 terminated password. Allow passwords to be passed on command line 9141 environment or config files in a few more utilities. 9142 [Steve Henson] 9143 9144 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 9145 keys. Add some short names for PKCS#8 PBE algorithms and allow them 9146 to be specified on the command line for the pkcs8 and pkcs12 utilities. 9147 Update documentation. 9148 [Steve Henson] 9149 9150 *) Support for ASN1 "NULL" type. This could be handled before by using 9151 ASN1_TYPE but there wasn't any function that would try to read a NULL 9152 and produce an error if it couldn't. For compatibility we also have 9153 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 9154 don't allocate anything because they don't need to. 9155 [Steve Henson] 9156 9157 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 9158 for details. 9159 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 9160 9161 *) Rebuild of the memory allocation routines used by OpenSSL code and 9162 possibly others as well. The purpose is to make an interface that 9163 provide hooks so anyone can build a separate set of allocation and 9164 deallocation routines to be used by OpenSSL, for example memory 9165 pool implementations, or something else, which was previously hard 9166 since Malloc(), Realloc() and Free() were defined as macros having 9167 the values malloc, realloc and free, respectively (except for Win32 9168 compilations). The same is provided for memory debugging code. 9169 OpenSSL already comes with functionality to find memory leaks, but 9170 this gives people a chance to debug other memory problems. 9171 9172 With these changes, a new set of functions and macros have appeared: 9173 9174 CRYPTO_set_mem_debug_functions() [F] 9175 CRYPTO_get_mem_debug_functions() [F] 9176 CRYPTO_dbg_set_options() [F] 9177 CRYPTO_dbg_get_options() [F] 9178 CRYPTO_malloc_debug_init() [M] 9179 9180 The memory debug functions are NULL by default, unless the library 9181 is compiled with CRYPTO_MDEBUG or friends is defined. If someone 9182 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 9183 gives the standard debugging functions that come with OpenSSL) or 9184 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 9185 provided by the library user) must be used. When the standard 9186 debugging functions are used, CRYPTO_dbg_set_options can be used to 9187 request additional information: 9188 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 9189 the CRYPTO_MDEBUG_xxx macro when compiling the library. 9190 9191 Also, things like CRYPTO_set_mem_functions will always give the 9192 expected result (the new set of functions is used for allocation 9193 and deallocation) at all times, regardless of platform and compiler 9194 options. 9195 9196 To finish it up, some functions that were never use in any other 9197 way than through macros have a new API and new semantic: 9198 9199 CRYPTO_dbg_malloc() 9200 CRYPTO_dbg_realloc() 9201 CRYPTO_dbg_free() 9202 9203 All macros of value have retained their old syntax. 9204 [Richard Levitte and Bodo Moeller] 9205 9206 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 9207 ordering of SMIMECapabilities wasn't in "strength order" and there 9208 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 9209 algorithm. 9210 [Steve Henson] 9211 9212 *) Some ASN1 types with illegal zero length encoding (INTEGER, 9213 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 9214 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 9215 9216 *) Merge in my S/MIME library for OpenSSL. This provides a simple 9217 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 9218 functionality to handle multipart/signed properly) and a utility 9219 called 'smime' to call all this stuff. This is based on code I 9220 originally wrote for Celo who have kindly allowed it to be 9221 included in OpenSSL. 9222 [Steve Henson] 9223 9224 *) Add variants des_set_key_checked and des_set_key_unchecked of 9225 des_set_key (aka des_key_sched). Global variable des_check_key 9226 decides which of these is called by des_set_key; this way 9227 des_check_key behaves as it always did, but applications and 9228 the library itself, which was buggy for des_check_key == 1, 9229 have a cleaner way to pick the version they need. 9230 [Bodo Moeller] 9231 9232 *) New function PKCS12_newpass() which changes the password of a 9233 PKCS12 structure. 9234 [Steve Henson] 9235 9236 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 9237 dynamic mix. In both cases the ids can be used as an index into the 9238 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 9239 functions so they accept a list of the field values and the 9240 application doesn't need to directly manipulate the X509_TRUST 9241 structure. 9242 [Steve Henson] 9243 9244 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 9245 need initialising. 9246 [Steve Henson] 9247 9248 *) Modify the way the V3 extension code looks up extensions. This now 9249 works in a similar way to the object code: we have some "standard" 9250 extensions in a static table which is searched with OBJ_bsearch() 9251 and the application can add dynamic ones if needed. The file 9252 crypto/x509v3/ext_dat.h now has the info: this file needs to be 9253 updated whenever a new extension is added to the core code and kept 9254 in ext_nid order. There is a simple program 'tabtest.c' which checks 9255 this. New extensions are not added too often so this file can readily 9256 be maintained manually. 9257 9258 There are two big advantages in doing things this way. The extensions 9259 can be looked up immediately and no longer need to be "added" using 9260 X509V3_add_standard_extensions(): this function now does nothing. 9261 [Side note: I get *lots* of email saying the extension code doesn't 9262 work because people forget to call this function] 9263 Also no dynamic allocation is done unless new extensions are added: 9264 so if we don't add custom extensions there is no need to call 9265 X509V3_EXT_cleanup(). 9266 [Steve Henson] 9267 9268 *) Modify enc utility's salting as follows: make salting the default. Add a 9269 magic header, so unsalted files fail gracefully instead of just decrypting 9270 to garbage. This is because not salting is a big security hole, so people 9271 should be discouraged from doing it. 9272 [Ben Laurie] 9273 9274 *) Fixes and enhancements to the 'x509' utility. It allowed a message 9275 digest to be passed on the command line but it only used this 9276 parameter when signing a certificate. Modified so all relevant 9277 operations are affected by the digest parameter including the 9278 -fingerprint and -x509toreq options. Also -x509toreq choked if a 9279 DSA key was used because it didn't fix the digest. 9280 [Steve Henson] 9281 9282 *) Initial certificate chain verify code. Currently tests the untrusted 9283 certificates for consistency with the verify purpose (which is set 9284 when the X509_STORE_CTX structure is set up) and checks the pathlength. 9285 9286 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 9287 this is because it will reject chains with invalid extensions whereas 9288 every previous version of OpenSSL and SSLeay made no checks at all. 9289 9290 Trust code: checks the root CA for the relevant trust settings. Trust 9291 settings have an initial value consistent with the verify purpose: e.g. 9292 if the verify purpose is for SSL client use it expects the CA to be 9293 trusted for SSL client use. However the default value can be changed to 9294 permit custom trust settings: one example of this would be to only trust 9295 certificates from a specific "secure" set of CAs. 9296 9297 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 9298 which should be used for version portability: especially since the 9299 verify structure is likely to change more often now. 9300 9301 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 9302 to set them. If not set then assume SSL clients will verify SSL servers 9303 and vice versa. 9304 9305 Two new options to the verify program: -untrusted allows a set of 9306 untrusted certificates to be passed in and -purpose which sets the 9307 intended purpose of the certificate. If a purpose is set then the 9308 new chain verify code is used to check extension consistency. 9309 [Steve Henson] 9310 9311 *) Support for the authority information access extension. 9312 [Steve Henson] 9313 9314 *) Modify RSA and DSA PEM read routines to transparently handle 9315 PKCS#8 format private keys. New *_PUBKEY_* functions that handle 9316 public keys in a format compatible with certificate 9317 SubjectPublicKeyInfo structures. Unfortunately there were already 9318 functions called *_PublicKey_* which used various odd formats so 9319 these are retained for compatibility: however the DSA variants were 9320 never in a public release so they have been deleted. Changed dsa/rsa 9321 utilities to handle the new format: note no releases ever handled public 9322 keys so we should be OK. 9323 9324 The primary motivation for this change is to avoid the same fiasco 9325 that dogs private keys: there are several incompatible private key 9326 formats some of which are standard and some OpenSSL specific and 9327 require various evil hacks to allow partial transparent handling and 9328 even then it doesn't work with DER formats. Given the option anything 9329 other than PKCS#8 should be dumped: but the other formats have to 9330 stay in the name of compatibility. 9331 9332 With public keys and the benefit of hindsight one standard format 9333 is used which works with EVP_PKEY, RSA or DSA structures: though 9334 it clearly returns an error if you try to read the wrong kind of key. 9335 9336 Added a -pubkey option to the 'x509' utility to output the public key. 9337 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 9338 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 9339 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 9340 that do the same as the EVP_PKEY_assign_*() except they up the 9341 reference count of the added key (they don't "swallow" the 9342 supplied key). 9343 [Steve Henson] 9344 9345 *) Fixes to crypto/x509/by_file.c the code to read in certificates and 9346 CRLs would fail if the file contained no certificates or no CRLs: 9347 added a new function to read in both types and return the number 9348 read: this means that if none are read it will be an error. The 9349 DER versions of the certificate and CRL reader would always fail 9350 because it isn't possible to mix certificates and CRLs in DER format 9351 without choking one or the other routine. Changed this to just read 9352 a certificate: this is the best we can do. Also modified the code 9353 in apps/verify.c to take notice of return codes: it was previously 9354 attempting to read in certificates from NULL pointers and ignoring 9355 any errors: this is one reason why the cert and CRL reader seemed 9356 to work. It doesn't check return codes from the default certificate 9357 routines: these may well fail if the certificates aren't installed. 9358 [Steve Henson] 9359 9360 *) Code to support otherName option in GeneralName. 9361 [Steve Henson] 9362 9363 *) First update to verify code. Change the verify utility 9364 so it warns if it is passed a self signed certificate: 9365 for consistency with the normal behaviour. X509_verify 9366 has been modified to it will now verify a self signed 9367 certificate if *exactly* the same certificate appears 9368 in the store: it was previously impossible to trust a 9369 single self signed certificate. This means that: 9370 openssl verify ss.pem 9371 now gives a warning about a self signed certificate but 9372 openssl verify -CAfile ss.pem ss.pem 9373 is OK. 9374 [Steve Henson] 9375 9376 *) For servers, store verify_result in SSL_SESSION data structure 9377 (and add it to external session representation). 9378 This is needed when client certificate verifications fails, 9379 but an application-provided verification callback (set by 9380 SSL_CTX_set_cert_verify_callback) allows accepting the session 9381 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 9382 but returns 1): When the session is reused, we have to set 9383 ssl->verify_result to the appropriate error code to avoid 9384 security holes. 9385 [Bodo Moeller, problem pointed out by Lutz Jaenicke] 9386 9387 *) Fix a bug in the new PKCS#7 code: it didn't consider the 9388 case in PKCS7_dataInit() where the signed PKCS7 structure 9389 didn't contain any existing data because it was being created. 9390 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 9391 9392 *) Add a salt to the key derivation routines in enc.c. This 9393 forms the first 8 bytes of the encrypted file. Also add a 9394 -S option to allow a salt to be input on the command line. 9395 [Steve Henson] 9396 9397 *) New function X509_cmp(). Oddly enough there wasn't a function 9398 to compare two certificates. We do this by working out the SHA1 9399 hash and comparing that. X509_cmp() will be needed by the trust 9400 code. 9401 [Steve Henson] 9402 9403 *) SSL_get1_session() is like SSL_get_session(), but increments 9404 the reference count in the SSL_SESSION returned. 9405 [Geoff Thorpe <geoff@eu.c2.net>] 9406 9407 *) Fix for 'req': it was adding a null to request attributes. 9408 Also change the X509_LOOKUP and X509_INFO code to handle 9409 certificate auxiliary information. 9410 [Steve Henson] 9411 9412 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 9413 the 'enc' command. 9414 [Steve Henson] 9415 9416 *) Add the possibility to add extra information to the memory leak 9417 detecting output, to form tracebacks, showing from where each 9418 allocation was originated: CRYPTO_push_info("constant string") adds 9419 the string plus current file name and line number to a per-thread 9420 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 9421 is like calling CYRPTO_pop_info() until the stack is empty. 9422 Also updated memory leak detection code to be multi-thread-safe. 9423 [Richard Levitte] 9424 9425 *) Add options -text and -noout to pkcs7 utility and delete the 9426 encryption options which never did anything. Update docs. 9427 [Steve Henson] 9428 9429 *) Add options to some of the utilities to allow the pass phrase 9430 to be included on either the command line (not recommended on 9431 OSes like Unix) or read from the environment. Update the 9432 manpages and fix a few bugs. 9433 [Steve Henson] 9434 9435 *) Add a few manpages for some of the openssl commands. 9436 [Steve Henson] 9437 9438 *) Fix the -revoke option in ca. It was freeing up memory twice, 9439 leaking and not finding already revoked certificates. 9440 [Steve Henson] 9441 9442 *) Extensive changes to support certificate auxiliary information. 9443 This involves the use of X509_CERT_AUX structure and X509_AUX 9444 functions. An X509_AUX function such as PEM_read_X509_AUX() 9445 can still read in a certificate file in the usual way but it 9446 will also read in any additional "auxiliary information". By 9447 doing things this way a fair degree of compatibility can be 9448 retained: existing certificates can have this information added 9449 using the new 'x509' options. 9450 9451 Current auxiliary information includes an "alias" and some trust 9452 settings. The trust settings will ultimately be used in enhanced 9453 certificate chain verification routines: currently a certificate 9454 can only be trusted if it is self signed and then it is trusted 9455 for all purposes. 9456 [Steve Henson] 9457 9458 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 9459 The problem was that one of the replacement routines had not been working 9460 since SSLeay releases. For now the offending routine has been replaced 9461 with non-optimised assembler. Even so, this now gives around 95% 9462 performance improvement for 1024 bit RSA signs. 9463 [Mark Cox] 9464 9465 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 9466 handling. Most clients have the effective key size in bits equal to 9467 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 9468 A few however don't do this and instead use the size of the decrypted key 9469 to determine the RC2 key length and the AlgorithmIdentifier to determine 9470 the effective key length. In this case the effective key length can still 9471 be 40 bits but the key length can be 168 bits for example. This is fixed 9472 by manually forcing an RC2 key into the EVP_PKEY structure because the 9473 EVP code can't currently handle unusual RC2 key sizes: it always assumes 9474 the key length and effective key length are equal. 9475 [Steve Henson] 9476 9477 *) Add a bunch of functions that should simplify the creation of 9478 X509_NAME structures. Now you should be able to do: 9479 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 9480 and have it automatically work out the correct field type and fill in 9481 the structures. The more adventurous can try: 9482 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 9483 and it will (hopefully) work out the correct multibyte encoding. 9484 [Steve Henson] 9485 9486 *) Change the 'req' utility to use the new field handling and multibyte 9487 copy routines. Before the DN field creation was handled in an ad hoc 9488 way in req, ca, and x509 which was rather broken and didn't support 9489 BMPStrings or UTF8Strings. Since some software doesn't implement 9490 BMPStrings or UTF8Strings yet, they can be enabled using the config file 9491 using the dirstring_type option. See the new comment in the default 9492 openssl.cnf for more info. 9493 [Steve Henson] 9494 9495 *) Make crypto/rand/md_rand.c more robust: 9496 - Assure unique random numbers after fork(). 9497 - Make sure that concurrent threads access the global counter and 9498 md serializably so that we never lose entropy in them 9499 or use exactly the same state in multiple threads. 9500 Access to the large state is not always serializable because 9501 the additional locking could be a performance killer, and 9502 md should be large enough anyway. 9503 [Bodo Moeller] 9504 9505 *) New file apps/app_rand.c with commonly needed functionality 9506 for handling the random seed file. 9507 9508 Use the random seed file in some applications that previously did not: 9509 ca, 9510 dsaparam -genkey (which also ignored its '-rand' option), 9511 s_client, 9512 s_server, 9513 x509 (when signing). 9514 Except on systems with /dev/urandom, it is crucial to have a random 9515 seed file at least for key creation, DSA signing, and for DH exchanges; 9516 for RSA signatures we could do without one. 9517 9518 gendh and gendsa (unlike genrsa) used to read only the first byte 9519 of each file listed in the '-rand' option. The function as previously 9520 found in genrsa is now in app_rand.c and is used by all programs 9521 that support '-rand'. 9522 [Bodo Moeller] 9523 9524 *) In RAND_write_file, use mode 0600 for creating files; 9525 don't just chmod when it may be too late. 9526 [Bodo Moeller] 9527 9528 *) Report an error from X509_STORE_load_locations 9529 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 9530 [Bill Perry] 9531 9532 *) New function ASN1_mbstring_copy() this copies a string in either 9533 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 9534 into an ASN1_STRING type. A mask of permissible types is passed 9535 and it chooses the "minimal" type to use or an error if not type 9536 is suitable. 9537 [Steve Henson] 9538 9539 *) Add function equivalents to the various macros in asn1.h. The old 9540 macros are retained with an M_ prefix. Code inside the library can 9541 use the M_ macros. External code (including the openssl utility) 9542 should *NOT* in order to be "shared library friendly". 9543 [Steve Henson] 9544 9545 *) Add various functions that can check a certificate's extensions 9546 to see if it usable for various purposes such as SSL client, 9547 server or S/MIME and CAs of these types. This is currently 9548 VERY EXPERIMENTAL but will ultimately be used for certificate chain 9549 verification. Also added a -purpose flag to x509 utility to 9550 print out all the purposes. 9551 [Steve Henson] 9552 9553 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 9554 functions. 9555 [Steve Henson] 9556 9557 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 9558 for, obtain and decode and extension and obtain its critical flag. 9559 This allows all the necessary extension code to be handled in a 9560 single function call. 9561 [Steve Henson] 9562 9563 *) RC4 tune-up featuring 30-40% performance improvement on most RISC 9564 platforms. See crypto/rc4/rc4_enc.c for further details. 9565 [Andy Polyakov] 9566 9567 *) New -noout option to asn1parse. This causes no output to be produced 9568 its main use is when combined with -strparse and -out to extract data 9569 from a file (which may not be in ASN.1 format). 9570 [Steve Henson] 9571 9572 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 9573 when producing the local key id. 9574 [Richard Levitte <levitte@stacken.kth.se>] 9575 9576 *) New option -dhparam in s_server. This allows a DH parameter file to be 9577 stated explicitly. If it is not stated then it tries the first server 9578 certificate file. The previous behaviour hard coded the filename 9579 "server.pem". 9580 [Steve Henson] 9581 9582 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 9583 a public key to be input or output. For example: 9584 openssl rsa -in key.pem -pubout -out pubkey.pem 9585 Also added necessary DSA public key functions to handle this. 9586 [Steve Henson] 9587 9588 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 9589 in the message. This was handled by allowing 9590 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 9591 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 9592 9593 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 9594 to the end of the strings whereas this didn't. This would cause problems 9595 if strings read with d2i_ASN1_bytes() were later modified. 9596 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 9597 9598 *) Fix for base64 decode bug. When a base64 bio reads only one line of 9599 data and it contains EOF it will end up returning an error. This is 9600 caused by input 46 bytes long. The cause is due to the way base64 9601 BIOs find the start of base64 encoded data. They do this by trying a 9602 trial decode on each line until they find one that works. When they 9603 do a flag is set and it starts again knowing it can pass all the 9604 data directly through the decoder. Unfortunately it doesn't reset 9605 the context it uses. This means that if EOF is reached an attempt 9606 is made to pass two EOFs through the context and this causes the 9607 resulting error. This can also cause other problems as well. As is 9608 usual with these problems it takes *ages* to find and the fix is 9609 trivial: move one line. 9610 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 9611 9612 *) Ugly workaround to get s_client and s_server working under Windows. The 9613 old code wouldn't work because it needed to select() on sockets and the 9614 tty (for keypresses and to see if data could be written). Win32 only 9615 supports select() on sockets so we select() with a 1s timeout on the 9616 sockets and then see if any characters are waiting to be read, if none 9617 are present then we retry, we also assume we can always write data to 9618 the tty. This isn't nice because the code then blocks until we've 9619 received a complete line of data and it is effectively polling the 9620 keyboard at 1s intervals: however it's quite a bit better than not 9621 working at all :-) A dedicated Windows application might handle this 9622 with an event loop for example. 9623 [Steve Henson] 9624 9625 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 9626 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 9627 will be called when RSA_sign() and RSA_verify() are used. This is useful 9628 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 9629 For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 9630 should *not* be used: RSA_sign() and RSA_verify() must be used instead. 9631 This necessitated the support of an extra signature type NID_md5_sha1 9632 for SSL signatures and modifications to the SSL library to use it instead 9633 of calling RSA_public_decrypt() and RSA_private_encrypt(). 9634 [Steve Henson] 9635 9636 *) Add new -verify -CAfile and -CApath options to the crl program, these 9637 will lookup a CRL issuers certificate and verify the signature in a 9638 similar way to the verify program. Tidy up the crl program so it 9639 no longer accesses structures directly. Make the ASN1 CRL parsing a bit 9640 less strict. It will now permit CRL extensions even if it is not 9641 a V2 CRL: this will allow it to tolerate some broken CRLs. 9642 [Steve Henson] 9643 9644 *) Initialize all non-automatic variables each time one of the openssl 9645 sub-programs is started (this is necessary as they may be started 9646 multiple times from the "OpenSSL>" prompt). 9647 [Lennart Bang, Bodo Moeller] 9648 9649 *) Preliminary compilation option RSA_NULL which disables RSA crypto without 9650 removing all other RSA functionality (this is what NO_RSA does). This 9651 is so (for example) those in the US can disable those operations covered 9652 by the RSA patent while allowing storage and parsing of RSA keys and RSA 9653 key generation. 9654 [Steve Henson] 9655 9656 *) Non-copying interface to BIO pairs. 9657 (still largely untested) 9658 [Bodo Moeller] 9659 9660 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 9661 ASCII string. This was handled independently in various places before. 9662 [Steve Henson] 9663 9664 *) New functions UTF8_getc() and UTF8_putc() that parse and generate 9665 UTF8 strings a character at a time. 9666 [Steve Henson] 9667 9668 *) Use client_version from client hello to select the protocol 9669 (s23_srvr.c) and for RSA client key exchange verification 9670 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 9671 [Bodo Moeller] 9672 9673 *) Add various utility functions to handle SPKACs, these were previously 9674 handled by poking round in the structure internals. Added new function 9675 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 9676 print, verify and generate SPKACs. Based on an original idea from 9677 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 9678 [Steve Henson] 9679 9680 *) RIPEMD160 is operational on all platforms and is back in 'make test'. 9681 [Andy Polyakov] 9682 9683 *) Allow the config file extension section to be overwritten on the 9684 command line. Based on an original idea from Massimiliano Pala 9685 <madwolf@comune.modena.it>. The new option is called -extensions 9686 and can be applied to ca, req and x509. Also -reqexts to override 9687 the request extensions in req and -crlexts to override the crl extensions 9688 in ca. 9689 [Steve Henson] 9690 9691 *) Add new feature to the SPKAC handling in ca. Now you can include 9692 the same field multiple times by preceding it by "XXXX." for example: 9693 1.OU="Unit name 1" 9694 2.OU="Unit name 2" 9695 this is the same syntax as used in the req config file. 9696 [Steve Henson] 9697 9698 *) Allow certificate extensions to be added to certificate requests. These 9699 are specified in a 'req_extensions' option of the req section of the 9700 config file. They can be printed out with the -text option to req but 9701 are otherwise ignored at present. 9702 [Steve Henson] 9703 9704 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 9705 data read consists of only the final block it would not decrypted because 9706 EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 9707 A misplaced 'break' also meant the decrypted final block might not be 9708 copied until the next read. 9709 [Steve Henson] 9710 9711 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 9712 a few extra parameters to the DH structure: these will be useful if 9713 for example we want the value of 'q' or implement X9.42 DH. 9714 [Steve Henson] 9715 9716 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 9717 provides hooks that allow the default DSA functions or functions on a 9718 "per key" basis to be replaced. This allows hardware acceleration and 9719 hardware key storage to be handled without major modification to the 9720 library. Also added low level modexp hooks and CRYPTO_EX structure and 9721 associated functions. 9722 [Steve Henson] 9723 9724 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 9725 as "read only": it can't be written to and the buffer it points to will 9726 not be freed. Reading from a read only BIO is much more efficient than 9727 a normal memory BIO. This was added because there are several times when 9728 an area of memory needs to be read from a BIO. The previous method was 9729 to create a memory BIO and write the data to it, this results in two 9730 copies of the data and an O(n^2) reading algorithm. There is a new 9731 function BIO_new_mem_buf() which creates a read only memory BIO from 9732 an area of memory. Also modified the PKCS#7 routines to use read only 9733 memory BIOs. 9734 [Steve Henson] 9735 9736 *) Bugfix: ssl23_get_client_hello did not work properly when called in 9737 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 9738 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 9739 but a retry condition occured while trying to read the rest. 9740 [Bodo Moeller] 9741 9742 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 9743 NID_pkcs7_encrypted by default: this was wrong since this should almost 9744 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 9745 the encrypted data type: this is a more sensible place to put it and it 9746 allows the PKCS#12 code to be tidied up that duplicated this 9747 functionality. 9748 [Steve Henson] 9749 9750 *) Changed obj_dat.pl script so it takes its input and output files on 9751 the command line. This should avoid shell escape redirection problems 9752 under Win32. 9753 [Steve Henson] 9754 9755 *) Initial support for certificate extension requests, these are included 9756 in things like Xenroll certificate requests. Included functions to allow 9757 extensions to be obtained and added. 9758 [Steve Henson] 9759 9760 *) -crlf option to s_client and s_server for sending newlines as 9761 CRLF (as required by many protocols). 9762 [Bodo Moeller] 9763 9764 Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 9765 9766 *) Install libRSAglue.a when OpenSSL is built with RSAref. 9767 [Ralf S. Engelschall] 9768 9769 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 9770 [Andrija Antonijevic <TheAntony2@bigfoot.com>] 9771 9772 *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 9773 program. 9774 [Steve Henson] 9775 9776 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 9777 DH parameters/keys (q is lost during that conversion, but the resulting 9778 DH parameters contain its length). 9779 9780 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 9781 much faster than DH_generate_parameters (which creates parameters 9782 where p = 2*q + 1), and also the smaller q makes DH computations 9783 much more efficient (160-bit exponentiation instead of 1024-bit 9784 exponentiation); so this provides a convenient way to support DHE 9785 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 9786 utter importance to use 9787 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 9788 or 9789 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 9790 when such DH parameters are used, because otherwise small subgroup 9791 attacks may become possible! 9792 [Bodo Moeller] 9793 9794 *) Avoid memory leak in i2d_DHparams. 9795 [Bodo Moeller] 9796 9797 *) Allow the -k option to be used more than once in the enc program: 9798 this allows the same encrypted message to be read by multiple recipients. 9799 [Steve Henson] 9800 9801 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 9802 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 9803 it will always use the numerical form of the OID, even if it has a short 9804 or long name. 9805 [Steve Henson] 9806 9807 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 9808 method only got called if p,q,dmp1,dmq1,iqmp components were present, 9809 otherwise bn_mod_exp was called. In the case of hardware keys for example 9810 no private key components need be present and it might store extra data 9811 in the RSA structure, which cannot be accessed from bn_mod_exp. 9812 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 9813 private key operations. 9814 [Steve Henson] 9815 9816 *) Added support for SPARC Linux. 9817 [Andy Polyakov] 9818 9819 *) pem_password_cb function type incompatibly changed from 9820 typedef int pem_password_cb(char *buf, int size, int rwflag); 9821 to 9822 ....(char *buf, int size, int rwflag, void *userdata); 9823 so that applications can pass data to their callbacks: 9824 The PEM[_ASN1]_{read,write}... functions and macros now take an 9825 additional void * argument, which is just handed through whenever 9826 the password callback is called. 9827 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 9828 9829 New function SSL_CTX_set_default_passwd_cb_userdata. 9830 9831 Compatibility note: As many C implementations push function arguments 9832 onto the stack in reverse order, the new library version is likely to 9833 interoperate with programs that have been compiled with the old 9834 pem_password_cb definition (PEM_whatever takes some data that 9835 happens to be on the stack as its last argument, and the callback 9836 just ignores this garbage); but there is no guarantee whatsoever that 9837 this will work. 9838 9839 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 9840 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 9841 problems not only on Windows, but also on some Unix platforms. 9842 To avoid problematic command lines, these definitions are now in an 9843 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 9844 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 9845 [Bodo Moeller] 9846 9847 *) MIPS III/IV assembler module is reimplemented. 9848 [Andy Polyakov] 9849 9850 *) More DES library cleanups: remove references to srand/rand and 9851 delete an unused file. 9852 [Ulf M��ller] 9853 9854 *) Add support for the the free Netwide assembler (NASM) under Win32, 9855 since not many people have MASM (ml) and it can be hard to obtain. 9856 This is currently experimental but it seems to work OK and pass all 9857 the tests. Check out INSTALL.W32 for info. 9858 [Steve Henson] 9859 9860 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 9861 without temporary keys kept an extra copy of the server key, 9862 and connections with temporary keys did not free everything in case 9863 of an error. 9864 [Bodo Moeller] 9865 9866 *) New function RSA_check_key and new openssl rsa option -check 9867 for verifying the consistency of RSA keys. 9868 [Ulf Moeller, Bodo Moeller] 9869 9870 *) Various changes to make Win32 compile work: 9871 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 9872 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 9873 comparison" warnings. 9874 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 9875 [Steve Henson] 9876 9877 *) Add a debugging option to PKCS#5 v2 key generation function: when 9878 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 9879 derived keys are printed to stderr. 9880 [Steve Henson] 9881 9882 *) Copy the flags in ASN1_STRING_dup(). 9883 [Roman E. Pavlov <pre@mo.msk.ru>] 9884 9885 *) The x509 application mishandled signing requests containing DSA 9886 keys when the signing key was also DSA and the parameters didn't match. 9887 9888 It was supposed to omit the parameters when they matched the signing key: 9889 the verifying software was then supposed to automatically use the CA's 9890 parameters if they were absent from the end user certificate. 9891 9892 Omitting parameters is no longer recommended. The test was also 9893 the wrong way round! This was probably due to unusual behaviour in 9894 EVP_cmp_parameters() which returns 1 if the parameters match. 9895 This meant that parameters were omitted when they *didn't* match and 9896 the certificate was useless. Certificates signed with 'ca' didn't have 9897 this bug. 9898 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 9899 9900 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 9901 The interface is as follows: 9902 Applications can use 9903 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 9904 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 9905 "off" is now the default. 9906 The library internally uses 9907 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 9908 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 9909 to disable memory-checking temporarily. 9910 9911 Some inconsistent states that previously were possible (and were 9912 even the default) are now avoided. 9913 9914 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 9915 with each memory chunk allocated; this is occasionally more helpful 9916 than just having a counter. 9917 9918 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 9919 9920 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 9921 extensions. 9922 [Bodo Moeller] 9923 9924 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 9925 which largely parallels "options", but is for changing API behaviour, 9926 whereas "options" are about protocol behaviour. 9927 Initial "mode" flags are: 9928 9929 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 9930 a single record has been written. 9931 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 9932 retries use the same buffer location. 9933 (But all of the contents must be 9934 copied!) 9935 [Bodo Moeller] 9936 9937 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 9938 worked. 9939 9940 *) Fix problems with no-hmac etc. 9941 [Ulf M��ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 9942 9943 *) New functions RSA_get_default_method(), RSA_set_method() and 9944 RSA_get_method(). These allows replacement of RSA_METHODs without having 9945 to mess around with the internals of an RSA structure. 9946 [Steve Henson] 9947 9948 *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 9949 Also really enable memory leak checks in openssl.c and in some 9950 test programs. 9951 [Chad C. Mulligan, Bodo Moeller] 9952 9953 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 9954 up the length of negative integers. This has now been simplified to just 9955 store the length when it is first determined and use it later, rather 9956 than trying to keep track of where data is copied and updating it to 9957 point to the end. 9958 [Steve Henson, reported by Brien Wheeler 9959 <bwheeler@authentica-security.com>] 9960 9961 *) Add a new function PKCS7_signatureVerify. This allows the verification 9962 of a PKCS#7 signature but with the signing certificate passed to the 9963 function itself. This contrasts with PKCS7_dataVerify which assumes the 9964 certificate is present in the PKCS#7 structure. This isn't always the 9965 case: certificates can be omitted from a PKCS#7 structure and be 9966 distributed by "out of band" means (such as a certificate database). 9967 [Steve Henson] 9968 9969 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 9970 function prototypes in pem.h, also change util/mkdef.pl to add the 9971 necessary function names. 9972 [Steve Henson] 9973 9974 *) mk1mf.pl (used by Windows builds) did not properly read the 9975 options set by Configure in the top level Makefile, and Configure 9976 was not even able to write more than one option correctly. 9977 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 9978 [Bodo Moeller] 9979 9980 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 9981 file to be loaded from a BIO or FILE pointer. The BIO version will 9982 for example allow memory BIOs to contain config info. 9983 [Steve Henson] 9984 9985 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 9986 Whoever hopes to achieve shared-library compatibility across versions 9987 must use this, not the compile-time macro. 9988 (Exercise 0.9.4: Which is the minimum library version required by 9989 such programs?) 9990 Note: All this applies only to multi-threaded programs, others don't 9991 need locks. 9992 [Bodo Moeller] 9993 9994 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 9995 through a BIO pair triggered the default case, i.e. 9996 SSLerr(...,SSL_R_UNKNOWN_STATE). 9997 [Bodo Moeller] 9998 9999 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 10000 can use the SSL library even if none of the specific BIOs is 10001 appropriate. 10002 [Bodo Moeller] 10003 10004 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 10005 for the encoded length. 10006 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 10007 10008 *) Add initial documentation of the X509V3 functions. 10009 [Steve Henson] 10010 10011 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 10012 PEM_write_bio_PKCS8PrivateKey() that are equivalent to 10013 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 10014 secure PKCS#8 private key format with a high iteration count. 10015 [Steve Henson] 10016 10017 *) Fix determination of Perl interpreter: A perl or perl5 10018 _directory_ in $PATH was also accepted as the interpreter. 10019 [Ralf S. Engelschall] 10020 10021 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 10022 wrong with it but it was very old and did things like calling 10023 PEM_ASN1_read() directly and used MD5 for the hash not to mention some 10024 unusual formatting. 10025 [Steve Henson] 10026 10027 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 10028 to use the new extension code. 10029 [Steve Henson] 10030 10031 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 10032 with macros. This should make it easier to change their form, add extra 10033 arguments etc. Fix a few PEM prototypes which didn't have cipher as a 10034 constant. 10035 [Steve Henson] 10036 10037 *) Add to configuration table a new entry that can specify an alternative 10038 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 10039 according to Mark Crispin <MRC@Panda.COM>. 10040 [Bodo Moeller] 10041 10042#if 0 10043 *) DES CBC did not update the IV. Weird. 10044 [Ben Laurie] 10045#else 10046 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 10047 Changing the behaviour of the former might break existing programs -- 10048 where IV updating is needed, des_ncbc_encrypt can be used. 10049#endif 10050 10051 *) When bntest is run from "make test" it drives bc to check its 10052 calculations, as well as internally checking them. If an internal check 10053 fails, it needs to cause bc to give a non-zero result or make test carries 10054 on without noticing the failure. Fixed. 10055 [Ben Laurie] 10056 10057 *) DES library cleanups. 10058 [Ulf M��ller] 10059 10060 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 10061 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 10062 ciphers. NOTE: although the key derivation function has been verified 10063 against some published test vectors it has not been extensively tested 10064 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 10065 of v2.0. 10066 [Steve Henson] 10067 10068 *) Instead of "mkdir -p", which is not fully portable, use new 10069 Perl script "util/mkdir-p.pl". 10070 [Bodo Moeller] 10071 10072 *) Rewrite the way password based encryption (PBE) is handled. It used to 10073 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 10074 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 10075 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 10076 the 'parameter' field of the AlgorithmIdentifier is passed to the 10077 underlying key generation function so it must do its own ASN1 parsing. 10078 This has also changed the EVP_PBE_CipherInit() function which now has a 10079 'parameter' argument instead of literal salt and iteration count values 10080 and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 10081 [Steve Henson] 10082 10083 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 10084 and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 10085 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 10086 KEY" because this clashed with PKCS#8 unencrypted string. Since this 10087 value was just used as a "magic string" and not used directly its 10088 value doesn't matter. 10089 [Steve Henson] 10090 10091 *) Introduce some semblance of const correctness to BN. Shame C doesn't 10092 support mutable. 10093 [Ben Laurie] 10094 10095 *) "linux-sparc64" configuration (ultrapenguin). 10096 [Ray Miller <ray.miller@oucs.ox.ac.uk>] 10097 "linux-sparc" configuration. 10098 [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 10099 10100 *) config now generates no-xxx options for missing ciphers. 10101 [Ulf M��ller] 10102 10103 *) Support the EBCDIC character set (work in progress). 10104 File ebcdic.c not yet included because it has a different license. 10105 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 10106 10107 *) Support BS2000/OSD-POSIX. 10108 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 10109 10110 *) Make callbacks for key generation use void * instead of char *. 10111 [Ben Laurie] 10112 10113 *) Make S/MIME samples compile (not yet tested). 10114 [Ben Laurie] 10115 10116 *) Additional typesafe stacks. 10117 [Ben Laurie] 10118 10119 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 10120 [Bodo Moeller] 10121 10122 10123 Changes between 0.9.3 and 0.9.3a [29 May 1999] 10124 10125 *) New configuration variant "sco5-gcc". 10126 10127 *) Updated some demos. 10128 [Sean O Riordain, Wade Scholine] 10129 10130 *) Add missing BIO_free at exit of pkcs12 application. 10131 [Wu Zhigang] 10132 10133 *) Fix memory leak in conf.c. 10134 [Steve Henson] 10135 10136 *) Updates for Win32 to assembler version of MD5. 10137 [Steve Henson] 10138 10139 *) Set #! path to perl in apps/der_chop to where we found it 10140 instead of using a fixed path. 10141 [Bodo Moeller] 10142 10143 *) SHA library changes for irix64-mips4-cc. 10144 [Andy Polyakov] 10145 10146 *) Improvements for VMS support. 10147 [Richard Levitte] 10148 10149 10150 Changes between 0.9.2b and 0.9.3 [24 May 1999] 10151 10152 *) Bignum library bug fix. IRIX 6 passes "make test" now! 10153 This also avoids the problems with SC4.2 and unpatched SC5. 10154 [Andy Polyakov <appro@fy.chalmers.se>] 10155 10156 *) New functions sk_num, sk_value and sk_set to replace the previous macros. 10157 These are required because of the typesafe stack would otherwise break 10158 existing code. If old code used a structure member which used to be STACK 10159 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 10160 sk_num or sk_value it would produce an error because the num, data members 10161 are not present in STACK_OF. Now it just produces a warning. sk_set 10162 replaces the old method of assigning a value to sk_value 10163 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 10164 that does this will no longer work (and should use sk_set instead) but 10165 this could be regarded as a "questionable" behaviour anyway. 10166 [Steve Henson] 10167 10168 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 10169 correctly handle encrypted S/MIME data. 10170 [Steve Henson] 10171 10172 *) Change type of various DES function arguments from des_cblock 10173 (which means, in function argument declarations, pointer to char) 10174 to des_cblock * (meaning pointer to array with 8 char elements), 10175 which allows the compiler to do more typechecking; it was like 10176 that back in SSLeay, but with lots of ugly casts. 10177 10178 Introduce new type const_des_cblock. 10179 [Bodo Moeller] 10180 10181 *) Reorganise the PKCS#7 library and get rid of some of the more obvious 10182 problems: find RecipientInfo structure that matches recipient certificate 10183 and initialise the ASN1 structures properly based on passed cipher. 10184 [Steve Henson] 10185 10186 *) Belatedly make the BN tests actually check the results. 10187 [Ben Laurie] 10188 10189 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 10190 to and from BNs: it was completely broken. New compilation option 10191 NEG_PUBKEY_BUG to allow for some broken certificates that encode public 10192 key elements as negative integers. 10193 [Steve Henson] 10194 10195 *) Reorganize and speed up MD5. 10196 [Andy Polyakov <appro@fy.chalmers.se>] 10197 10198 *) VMS support. 10199 [Richard Levitte <richard@levitte.org>] 10200 10201 *) New option -out to asn1parse to allow the parsed structure to be 10202 output to a file. This is most useful when combined with the -strparse 10203 option to examine the output of things like OCTET STRINGS. 10204 [Steve Henson] 10205 10206 *) Make SSL library a little more fool-proof by not requiring any longer 10207 that SSL_set_{accept,connect}_state be called before 10208 SSL_{accept,connect} may be used (SSL_set_..._state is omitted 10209 in many applications because usually everything *appeared* to work as 10210 intended anyway -- now it really works as intended). 10211 [Bodo Moeller] 10212 10213 *) Move openssl.cnf out of lib/. 10214 [Ulf M��ller] 10215 10216 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 10217 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 10218 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 10219 [Ralf S. Engelschall] 10220 10221 *) Various fixes to the EVP and PKCS#7 code. It may now be able to 10222 handle PKCS#7 enveloped data properly. 10223 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 10224 10225 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 10226 copying pointers. The cert_st handling is changed by this in 10227 various ways (and thus what used to be known as ctx->default_cert 10228 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 10229 any longer when s->cert does not give us what we need). 10230 ssl_cert_instantiate becomes obsolete by this change. 10231 As soon as we've got the new code right (possibly it already is?), 10232 we have solved a couple of bugs of the earlier code where s->cert 10233 was used as if it could not have been shared with other SSL structures. 10234 10235 Note that using the SSL API in certain dirty ways now will result 10236 in different behaviour than observed with earlier library versions: 10237 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 10238 does not influence s as it used to. 10239 10240 In order to clean up things more thoroughly, inside SSL_SESSION 10241 we don't use CERT any longer, but a new structure SESS_CERT 10242 that holds per-session data (if available); currently, this is 10243 the peer's certificate chain and, for clients, the server's certificate 10244 and temporary key. CERT holds only those values that can have 10245 meaningful defaults in an SSL_CTX. 10246 [Bodo Moeller] 10247 10248 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 10249 from the internal representation. Various PKCS#7 fixes: remove some 10250 evil casts and set the enc_dig_alg field properly based on the signing 10251 key type. 10252 [Steve Henson] 10253 10254 *) Allow PKCS#12 password to be set from the command line or the 10255 environment. Let 'ca' get its config file name from the environment 10256 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 10257 and 'x509'). 10258 [Steve Henson] 10259 10260 *) Allow certificate policies extension to use an IA5STRING for the 10261 organization field. This is contrary to the PKIX definition but 10262 VeriSign uses it and IE5 only recognises this form. Document 'x509' 10263 extension option. 10264 [Steve Henson] 10265 10266 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 10267 without disallowing inline assembler and the like for non-pedantic builds. 10268 [Ben Laurie] 10269 10270 *) Support Borland C++ builder. 10271 [Janez Jere <jj@void.si>, modified by Ulf M��ller] 10272 10273 *) Support Mingw32. 10274 [Ulf M��ller] 10275 10276 *) SHA-1 cleanups and performance enhancements. 10277 [Andy Polyakov <appro@fy.chalmers.se>] 10278 10279 *) Sparc v8plus assembler for the bignum library. 10280 [Andy Polyakov <appro@fy.chalmers.se>] 10281 10282 *) Accept any -xxx and +xxx compiler options in Configure. 10283 [Ulf M��ller] 10284 10285 *) Update HPUX configuration. 10286 [Anonymous] 10287 10288 *) Add missing sk_<type>_unshift() function to safestack.h 10289 [Ralf S. Engelschall] 10290 10291 *) New function SSL_CTX_use_certificate_chain_file that sets the 10292 "extra_cert"s in addition to the certificate. (This makes sense 10293 only for "PEM" format files, as chains as a whole are not 10294 DER-encoded.) 10295 [Bodo Moeller] 10296 10297 *) Support verify_depth from the SSL API. 10298 x509_vfy.c had what can be considered an off-by-one-error: 10299 Its depth (which was not part of the external interface) 10300 was actually counting the number of certificates in a chain; 10301 now it really counts the depth. 10302 [Bodo Moeller] 10303 10304 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 10305 instead of X509err, which often resulted in confusing error 10306 messages since the error codes are not globally unique 10307 (e.g. an alleged error in ssl3_accept when a certificate 10308 didn't match the private key). 10309 10310 *) New function SSL_CTX_set_session_id_context that allows to set a default 10311 value (so that you don't need SSL_set_session_id_context for each 10312 connection using the SSL_CTX). 10313 [Bodo Moeller] 10314 10315 *) OAEP decoding bug fix. 10316 [Ulf M��ller] 10317 10318 *) Support INSTALL_PREFIX for package builders, as proposed by 10319 David Harris. 10320 [Bodo Moeller] 10321 10322 *) New Configure options "threads" and "no-threads". For systems 10323 where the proper compiler options are known (currently Solaris 10324 and Linux), "threads" is the default. 10325 [Bodo Moeller] 10326 10327 *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 10328 [Bodo Moeller] 10329 10330 *) Install various scripts to $(OPENSSLDIR)/misc, not to 10331 $(INSTALLTOP)/bin -- they shouldn't clutter directories 10332 such as /usr/local/bin. 10333 [Bodo Moeller] 10334 10335 *) "make linux-shared" to build shared libraries. 10336 [Niels Poppe <niels@netbox.org>] 10337 10338 *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 10339 [Ulf M��ller] 10340 10341 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 10342 extension adding in x509 utility. 10343 [Steve Henson] 10344 10345 *) Remove NOPROTO sections and error code comments. 10346 [Ulf M��ller] 10347 10348 *) Partial rewrite of the DEF file generator to now parse the ANSI 10349 prototypes. 10350 [Steve Henson] 10351 10352 *) New Configure options --prefix=DIR and --openssldir=DIR. 10353 [Ulf M��ller] 10354 10355 *) Complete rewrite of the error code script(s). It is all now handled 10356 by one script at the top level which handles error code gathering, 10357 header rewriting and C source file generation. It should be much better 10358 than the old method: it now uses a modified version of Ulf's parser to 10359 read the ANSI prototypes in all header files (thus the old K&R definitions 10360 aren't needed for error creation any more) and do a better job of 10361 translating function codes into names. The old 'ASN1 error code imbedded 10362 in a comment' is no longer necessary and it doesn't use .err files which 10363 have now been deleted. Also the error code call doesn't have to appear all 10364 on one line (which resulted in some large lines...). 10365 [Steve Henson] 10366 10367 *) Change #include filenames from <foo.h> to <openssl/foo.h>. 10368 [Bodo Moeller] 10369 10370 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 10371 0 (which usually indicates a closed connection), but continue reading. 10372 [Bodo Moeller] 10373 10374 *) Fix some race conditions. 10375 [Bodo Moeller] 10376 10377 *) Add support for CRL distribution points extension. Add Certificate 10378 Policies and CRL distribution points documentation. 10379 [Steve Henson] 10380 10381 *) Move the autogenerated header file parts to crypto/opensslconf.h. 10382 [Ulf M��ller] 10383 10384 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 10385 8 of keying material. Merlin has also confirmed interop with this fix 10386 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 10387 [Merlin Hughes <merlin@baltimore.ie>] 10388 10389 *) Fix lots of warnings. 10390 [Richard Levitte <levitte@stacken.kth.se>] 10391 10392 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 10393 the directory spec didn't end with a LIST_SEPARATOR_CHAR. 10394 [Richard Levitte <levitte@stacken.kth.se>] 10395 10396 *) Fix problems with sizeof(long) == 8. 10397 [Andy Polyakov <appro@fy.chalmers.se>] 10398 10399 *) Change functions to ANSI C. 10400 [Ulf M��ller] 10401 10402 *) Fix typos in error codes. 10403 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M��ller] 10404 10405 *) Remove defunct assembler files from Configure. 10406 [Ulf M��ller] 10407 10408 *) SPARC v8 assembler BIGNUM implementation. 10409 [Andy Polyakov <appro@fy.chalmers.se>] 10410 10411 *) Support for Certificate Policies extension: both print and set. 10412 Various additions to support the r2i method this uses. 10413 [Steve Henson] 10414 10415 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 10416 return a const string when you are expecting an allocated buffer. 10417 [Ben Laurie] 10418 10419 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 10420 types DirectoryString and DisplayText. 10421 [Steve Henson] 10422 10423 *) Add code to allow r2i extensions to access the configuration database, 10424 add an LHASH database driver and add several ctx helper functions. 10425 [Steve Henson] 10426 10427 *) Fix an evil bug in bn_expand2() which caused various BN functions to 10428 fail when they extended the size of a BIGNUM. 10429 [Steve Henson] 10430 10431 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 10432 support typesafe stack. 10433 [Steve Henson] 10434 10435 *) Fix typo in SSL_[gs]et_options(). 10436 [Nils Frostberg <nils@medcom.se>] 10437 10438 *) Delete various functions and files that belonged to the (now obsolete) 10439 old X509V3 handling code. 10440 [Steve Henson] 10441 10442 *) New Configure option "rsaref". 10443 [Ulf M��ller] 10444 10445 *) Don't auto-generate pem.h. 10446 [Bodo Moeller] 10447 10448 *) Introduce type-safe ASN.1 SETs. 10449 [Ben Laurie] 10450 10451 *) Convert various additional casted stacks to type-safe STACK_OF() variants. 10452 [Ben Laurie, Ralf S. Engelschall, Steve Henson] 10453 10454 *) Introduce type-safe STACKs. This will almost certainly break lots of code 10455 that links with OpenSSL (well at least cause lots of warnings), but fear 10456 not: the conversion is trivial, and it eliminates loads of evil casts. A 10457 few STACKed things have been converted already. Feel free to convert more. 10458 In the fullness of time, I'll do away with the STACK type altogether. 10459 [Ben Laurie] 10460 10461 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 10462 specified in <certfile> by updating the entry in the index.txt file. 10463 This way one no longer has to edit the index.txt file manually for 10464 revoking a certificate. The -revoke option does the gory details now. 10465 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 10466 10467 *) Fix `openssl crl -noout -text' combination where `-noout' killed the 10468 `-text' option at all and this way the `-noout -text' combination was 10469 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 10470 [Ralf S. Engelschall] 10471 10472 *) Make sure a corresponding plain text error message exists for the 10473 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 10474 verify callback function determined that a certificate was revoked. 10475 [Ralf S. Engelschall] 10476 10477 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 10478 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 10479 all available cipers including rc5, which was forgotten until now. 10480 In order to let the testing shell script know which algorithms 10481 are available, a new (up to now undocumented) command 10482 "openssl list-cipher-commands" is used. 10483 [Bodo Moeller] 10484 10485 *) Bugfix: s_client occasionally would sleep in select() when 10486 it should have checked SSL_pending() first. 10487 [Bodo Moeller] 10488 10489 *) New functions DSA_do_sign and DSA_do_verify to provide access to 10490 the raw DSA values prior to ASN.1 encoding. 10491 [Ulf M��ller] 10492 10493 *) Tweaks to Configure 10494 [Niels Poppe <niels@netbox.org>] 10495 10496 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 10497 yet... 10498 [Steve Henson] 10499 10500 *) New variables $(RANLIB) and $(PERL) in the Makefiles. 10501 [Ulf M��ller] 10502 10503 *) New config option to avoid instructions that are illegal on the 80386. 10504 The default code is faster, but requires at least a 486. 10505 [Ulf M��ller] 10506 10507 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 10508 SSL2_SERVER_VERSION (not used at all) macros, which are now the 10509 same as SSL2_VERSION anyway. 10510 [Bodo Moeller] 10511 10512 *) New "-showcerts" option for s_client. 10513 [Bodo Moeller] 10514 10515 *) Still more PKCS#12 integration. Add pkcs12 application to openssl 10516 application. Various cleanups and fixes. 10517 [Steve Henson] 10518 10519 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 10520 modify error routines to work internally. Add error codes and PBE init 10521 to library startup routines. 10522 [Steve Henson] 10523 10524 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 10525 packing functions to asn1 and evp. Changed function names and error 10526 codes along the way. 10527 [Steve Henson] 10528 10529 *) PKCS12 integration: and so it begins... First of several patches to 10530 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 10531 objects to objects.h 10532 [Steve Henson] 10533 10534 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 10535 and display support for Thawte strong extranet extension. 10536 [Steve Henson] 10537 10538 *) Add LinuxPPC support. 10539 [Jeff Dubrule <igor@pobox.org>] 10540 10541 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 10542 bn_div_words in alpha.s. 10543 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 10544 10545 *) Make sure the RSA OAEP test is skipped under -DRSAref because 10546 OAEP isn't supported when OpenSSL is built with RSAref. 10547 [Ulf Moeller <ulf@fitug.de>] 10548 10549 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 10550 so they no longer are missing under -DNOPROTO. 10551 [Soren S. Jorvang <soren@t.dk>] 10552 10553 10554 Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 10555 10556 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 10557 doesn't work when the session is reused. Coming soon! 10558 [Ben Laurie] 10559 10560 *) Fix a security hole, that allows sessions to be reused in the wrong 10561 context thus bypassing client cert protection! All software that uses 10562 client certs and session caches in multiple contexts NEEDS PATCHING to 10563 allow session reuse! A fuller solution is in the works. 10564 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 10565 10566 *) Some more source tree cleanups (removed obsolete files 10567 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 10568 permission on "config" script to be executable) and a fix for the INSTALL 10569 document. 10570 [Ulf Moeller <ulf@fitug.de>] 10571 10572 *) Remove some legacy and erroneous uses of malloc, free instead of 10573 Malloc, Free. 10574 [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 10575 10576 *) Make rsa_oaep_test return non-zero on error. 10577 [Ulf Moeller <ulf@fitug.de>] 10578 10579 *) Add support for native Solaris shared libraries. Configure 10580 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 10581 if someone would make that last step automatic. 10582 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 10583 10584 *) ctx_size was not built with the right compiler during "make links". Fixed. 10585 [Ben Laurie] 10586 10587 *) Change the meaning of 'ALL' in the cipher list. It now means "everything 10588 except NULL ciphers". This means the default cipher list will no longer 10589 enable NULL ciphers. They need to be specifically enabled e.g. with 10590 the string "DEFAULT:eNULL". 10591 [Steve Henson] 10592 10593 *) Fix to RSA private encryption routines: if p < q then it would 10594 occasionally produce an invalid result. This will only happen with 10595 externally generated keys because OpenSSL (and SSLeay) ensure p > q. 10596 [Steve Henson] 10597 10598 *) Be less restrictive and allow also `perl util/perlpath.pl 10599 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 10600 because this way one can also use an interpreter named `perl5' (which is 10601 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 10602 installed as `perl'). 10603 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 10604 10605 *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 10606 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 10607 10608 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 10609 advapi32.lib to Win32 build and change the pem test comparision 10610 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 10611 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 10612 and crypto/des/ede_cbcm_enc.c. 10613 [Steve Henson] 10614 10615 *) DES quad checksum was broken on big-endian architectures. Fixed. 10616 [Ben Laurie] 10617 10618 *) Comment out two functions in bio.h that aren't implemented. Fix up the 10619 Win32 test batch file so it (might) work again. The Win32 test batch file 10620 is horrible: I feel ill.... 10621 [Steve Henson] 10622 10623 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 10624 in e_os.h. Audit of header files to check ANSI and non ANSI 10625 sections: 10 functions were absent from non ANSI section and not exported 10626 from Windows DLLs. Fixed up libeay.num for new functions. 10627 [Steve Henson] 10628 10629 *) Make `openssl version' output lines consistent. 10630 [Ralf S. Engelschall] 10631 10632 *) Fix Win32 symbol export lists for BIO functions: Added 10633 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 10634 to ms/libeay{16,32}.def. 10635 [Ralf S. Engelschall] 10636 10637 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 10638 fine under Unix and passes some trivial tests I've now added. But the 10639 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 10640 added to make sure no one expects that this stuff really works in the 10641 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 10642 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 10643 openssl_bio.xs. 10644 [Ralf S. Engelschall] 10645 10646 *) Fix the generation of two part addresses in perl. 10647 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 10648 10649 *) Add config entry for Linux on MIPS. 10650 [John Tobey <jtobey@channel1.com>] 10651 10652 *) Make links whenever Configure is run, unless we are on Windoze. 10653 [Ben Laurie] 10654 10655 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 10656 Currently only issuerAltName and AuthorityKeyIdentifier make any sense 10657 in CRLs. 10658 [Steve Henson] 10659 10660 *) Add a useful kludge to allow package maintainers to specify compiler and 10661 other platforms details on the command line without having to patch the 10662 Configure script everytime: One now can use ``perl Configure 10663 <id>:<details>'', i.e. platform ids are allowed to have details appended 10664 to them (seperated by colons). This is treated as there would be a static 10665 pre-configured entry in Configure's %table under key <id> with value 10666 <details> and ``perl Configure <id>'' is called. So, when you want to 10667 perform a quick test-compile under FreeBSD 3.1 with pgcc and without 10668 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 10669 now, which overrides the FreeBSD-elf entry on-the-fly. 10670 [Ralf S. Engelschall] 10671 10672 *) Disable new TLS1 ciphersuites by default: they aren't official yet. 10673 [Ben Laurie] 10674 10675 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 10676 on the `perl Configure ...' command line. This way one can compile 10677 OpenSSL libraries with Position Independent Code (PIC) which is needed 10678 for linking it into DSOs. 10679 [Ralf S. Engelschall] 10680 10681 *) Remarkably, export ciphers were totally broken and no-one had noticed! 10682 Fixed. 10683 [Ben Laurie] 10684 10685 *) Cleaned up the LICENSE document: The official contact for any license 10686 questions now is the OpenSSL core team under openssl-core@openssl.org. 10687 And add a paragraph about the dual-license situation to make sure people 10688 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 10689 to the OpenSSL toolkit. 10690 [Ralf S. Engelschall] 10691 10692 *) General source tree makefile cleanups: Made `making xxx in yyy...' 10693 display consistent in the source tree and replaced `/bin/rm' by `rm'. 10694 Additonally cleaned up the `make links' target: Remove unnecessary 10695 semicolons, subsequent redundant removes, inline point.sh into mklink.sh 10696 to speed processing and no longer clutter the display with confusing 10697 stuff. Instead only the actually done links are displayed. 10698 [Ralf S. Engelschall] 10699 10700 *) Permit null encryption ciphersuites, used for authentication only. It used 10701 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 10702 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 10703 encryption. 10704 [Ben Laurie] 10705 10706 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 10707 signed attributes when verifying signatures (this would break them), 10708 the detached data encoding was wrong and public keys obtained using 10709 X509_get_pubkey() weren't freed. 10710 [Steve Henson] 10711 10712 *) Add text documentation for the BUFFER functions. Also added a work around 10713 to a Win95 console bug. This was triggered by the password read stuff: the 10714 last character typed gets carried over to the next fread(). If you were 10715 generating a new cert request using 'req' for example then the last 10716 character of the passphrase would be CR which would then enter the first 10717 field as blank. 10718 [Steve Henson] 10719 10720 *) Added the new `Includes OpenSSL Cryptography Software' button as 10721 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 10722 button and can be used by applications based on OpenSSL to show the 10723 relationship to the OpenSSL project. 10724 [Ralf S. Engelschall] 10725 10726 *) Remove confusing variables in function signatures in files 10727 ssl/ssl_lib.c and ssl/ssl.h. 10728 [Lennart Bong <lob@kulthea.stacken.kth.se>] 10729 10730 *) Don't install bss_file.c under PREFIX/include/ 10731 [Lennart Bong <lob@kulthea.stacken.kth.se>] 10732 10733 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 10734 functions that return function pointers and has support for NT specific 10735 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 10736 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 10737 unsigned to signed types: this was killing the Win32 compile. 10738 [Steve Henson] 10739 10740 *) Add new certificate file to stack functions, 10741 SSL_add_dir_cert_subjects_to_stack() and 10742 SSL_add_file_cert_subjects_to_stack(). These largely supplant 10743 SSL_load_client_CA_file(), and can be used to add multiple certs easily 10744 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 10745 This means that Apache-SSL and similar packages don't have to mess around 10746 to add as many CAs as they want to the preferred list. 10747 [Ben Laurie] 10748 10749 *) Experiment with doxygen documentation. Currently only partially applied to 10750 ssl/ssl_lib.c. 10751 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 10752 openssl.doxy as the configuration file. 10753 [Ben Laurie] 10754 10755 *) Get rid of remaining C++-style comments which strict C compilers hate. 10756 [Ralf S. Engelschall, pointed out by Carlos Amengual] 10757 10758 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 10759 compiled in by default: it has problems with large keys. 10760 [Steve Henson] 10761 10762 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 10763 DH private keys and/or callback functions which directly correspond to 10764 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 10765 is needed for applications which have to configure certificates on a 10766 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 10767 (e.g. s_server). 10768 For the RSA certificate situation is makes no difference, but 10769 for the DSA certificate situation this fixes the "no shared cipher" 10770 problem where the OpenSSL cipher selection procedure failed because the 10771 temporary keys were not overtaken from the context and the API provided 10772 no way to reconfigure them. 10773 The new functions now let applications reconfigure the stuff and they 10774 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 10775 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 10776 non-public-API function ssl_cert_instantiate() is used as a helper 10777 function and also to reduce code redundancy inside ssl_rsa.c. 10778 [Ralf S. Engelschall] 10779 10780 *) Move s_server -dcert and -dkey options out of the undocumented feature 10781 area because they are useful for the DSA situation and should be 10782 recognized by the users. 10783 [Ralf S. Engelschall] 10784 10785 *) Fix the cipher decision scheme for export ciphers: the export bits are 10786 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 10787 SSL_EXP_MASK. So, the original variable has to be used instead of the 10788 already masked variable. 10789 [Richard Levitte <levitte@stacken.kth.se>] 10790 10791 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 10792 [Richard Levitte <levitte@stacken.kth.se>] 10793 10794 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 10795 from `int' to `unsigned int' because it's a length and initialized by 10796 EVP_DigestFinal() which expects an `unsigned int *'. 10797 [Richard Levitte <levitte@stacken.kth.se>] 10798 10799 *) Don't hard-code path to Perl interpreter on shebang line of Configure 10800 script. Instead use the usual Shell->Perl transition trick. 10801 [Ralf S. Engelschall] 10802 10803 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 10804 (in addition to RSA certificates) to match the behaviour of `openssl dsa 10805 -noout -modulus' as it's already the case for `openssl rsa -noout 10806 -modulus'. For RSA the -modulus is the real "modulus" while for DSA 10807 currently the public key is printed (a decision which was already done by 10808 `openssl dsa -modulus' in the past) which serves a similar purpose. 10809 Additionally the NO_RSA no longer completely removes the whole -modulus 10810 option; it now only avoids using the RSA stuff. Same applies to NO_DSA 10811 now, too. 10812 [Ralf S. Engelschall] 10813 10814 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 10815 BIO. See the source (crypto/evp/bio_ok.c) for more info. 10816 [Arne Ansper <arne@ats.cyber.ee>] 10817 10818 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 10819 to be added. Now both 'req' and 'ca' can use new objects defined in the 10820 config file. 10821 [Steve Henson] 10822 10823 *) Add cool BIO that does syslog (or event log on NT). 10824 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 10825 10826 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 10827 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 10828 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 10829 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 10830 [Ben Laurie] 10831 10832 *) Add preliminary config info for new extension code. 10833 [Steve Henson] 10834 10835 *) Make RSA_NO_PADDING really use no padding. 10836 [Ulf Moeller <ulf@fitug.de>] 10837 10838 *) Generate errors when private/public key check is done. 10839 [Ben Laurie] 10840 10841 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 10842 for some CRL extensions and new objects added. 10843 [Steve Henson] 10844 10845 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 10846 key usage extension and fuller support for authority key id. 10847 [Steve Henson] 10848 10849 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 10850 padding method for RSA, which is recommended for new applications in PKCS 10851 #1 v2.0 (RFC 2437, October 1998). 10852 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 10853 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 10854 against Bleichbacher's attack on RSA. 10855 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 10856 Ben Laurie] 10857 10858 *) Updates to the new SSL compression code 10859 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 10860 10861 *) Fix so that the version number in the master secret, when passed 10862 via RSA, checks that if TLS was proposed, but we roll back to SSLv3 10863 (because the server will not accept higher), that the version number 10864 is 0x03,0x01, not 0x03,0x00 10865 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 10866 10867 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 10868 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 10869 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 10870 [Steve Henson] 10871 10872 *) Support for RAW extensions where an arbitrary extension can be 10873 created by including its DER encoding. See apps/openssl.cnf for 10874 an example. 10875 [Steve Henson] 10876 10877 *) Make sure latest Perl versions don't interpret some generated C array 10878 code as Perl array code in the crypto/err/err_genc.pl script. 10879 [Lars Weber <3weber@informatik.uni-hamburg.de>] 10880 10881 *) Modify ms/do_ms.bat to not generate assembly language makefiles since 10882 not many people have the assembler. Various Win32 compilation fixes and 10883 update to the INSTALL.W32 file with (hopefully) more accurate Win32 10884 build instructions. 10885 [Steve Henson] 10886 10887 *) Modify configure script 'Configure' to automatically create crypto/date.h 10888 file under Win32 and also build pem.h from pem.org. New script 10889 util/mkfiles.pl to create the MINFO file on environments that can't do a 10890 'make files': perl util/mkfiles.pl >MINFO should work. 10891 [Steve Henson] 10892 10893 *) Major rework of DES function declarations, in the pursuit of correctness 10894 and purity. As a result, many evil casts evaporated, and some weirdness, 10895 too. You may find this causes warnings in your code. Zapping your evil 10896 casts will probably fix them. Mostly. 10897 [Ben Laurie] 10898 10899 *) Fix for a typo in asn1.h. Bug fix to object creation script 10900 obj_dat.pl. It considered a zero in an object definition to mean 10901 "end of object": none of the objects in objects.h have any zeros 10902 so it wasn't spotted. 10903 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 10904 10905 *) Add support for Triple DES Cipher Block Chaining with Output Feedback 10906 Masking (CBCM). In the absence of test vectors, the best I have been able 10907 to do is check that the decrypt undoes the encrypt, so far. Send me test 10908 vectors if you have them. 10909 [Ben Laurie] 10910 10911 *) Correct calculation of key length for export ciphers (too much space was 10912 allocated for null ciphers). This has not been tested! 10913 [Ben Laurie] 10914 10915 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 10916 message is now correct (it understands "crypto" and "ssl" on its 10917 command line). There is also now an "update" option. This will update 10918 the util/ssleay.num and util/libeay.num files with any new functions. 10919 If you do a: 10920 perl util/mkdef.pl crypto ssl update 10921 it will update them. 10922 [Steve Henson] 10923 10924 *) Overhauled the Perl interface (perl/*): 10925 - ported BN stuff to OpenSSL's different BN library 10926 - made the perl/ source tree CVS-aware 10927 - renamed the package from SSLeay to OpenSSL (the files still contain 10928 their history because I've copied them in the repository) 10929 - removed obsolete files (the test scripts will be replaced 10930 by better Test::Harness variants in the future) 10931 [Ralf S. Engelschall] 10932 10933 *) First cut for a very conservative source tree cleanup: 10934 1. merge various obsolete readme texts into doc/ssleay.txt 10935 where we collect the old documents and readme texts. 10936 2. remove the first part of files where I'm already sure that we no 10937 longer need them because of three reasons: either they are just temporary 10938 files which were left by Eric or they are preserved original files where 10939 I've verified that the diff is also available in the CVS via "cvs diff 10940 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 10941 the crypto/md/ stuff). 10942 [Ralf S. Engelschall] 10943 10944 *) More extension code. Incomplete support for subject and issuer alt 10945 name, issuer and authority key id. Change the i2v function parameters 10946 and add an extra 'crl' parameter in the X509V3_CTX structure: guess 10947 what that's for :-) Fix to ASN1 macro which messed up 10948 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 10949 [Steve Henson] 10950 10951 *) Preliminary support for ENUMERATED type. This is largely copied from the 10952 INTEGER code. 10953 [Steve Henson] 10954 10955 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 10956 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 10957 10958 *) Make sure `make rehash' target really finds the `openssl' program. 10959 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 10960 10961 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 10962 like to hear about it if this slows down other processors. 10963 [Ben Laurie] 10964 10965 *) Add CygWin32 platform information to Configure script. 10966 [Alan Batie <batie@aahz.jf.intel.com>] 10967 10968 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 10969 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 10970 10971 *) New program nseq to manipulate netscape certificate sequences 10972 [Steve Henson] 10973 10974 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 10975 few typos. 10976 [Steve Henson] 10977 10978 *) Fixes to BN code. Previously the default was to define BN_RECURSION 10979 but the BN code had some problems that would cause failures when 10980 doing certificate verification and some other functions. 10981 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 10982 10983 *) Add ASN1 and PEM code to support netscape certificate sequences. 10984 [Steve Henson] 10985 10986 *) Add ASN1 and PEM code to support netscape certificate sequences. 10987 [Steve Henson] 10988 10989 *) Add several PKIX and private extended key usage OIDs. 10990 [Steve Henson] 10991 10992 *) Modify the 'ca' program to handle the new extension code. Modify 10993 openssl.cnf for new extension format, add comments. 10994 [Steve Henson] 10995 10996 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 10997 and add a sample to openssl.cnf so req -x509 now adds appropriate 10998 CA extensions. 10999 [Steve Henson] 11000 11001 *) Continued X509 V3 changes. Add to other makefiles, integrate with the 11002 error code, add initial support to X509_print() and x509 application. 11003 [Steve Henson] 11004 11005 *) Takes a deep breath and start addding X509 V3 extension support code. Add 11006 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 11007 stuff is currently isolated and isn't even compiled yet. 11008 [Steve Henson] 11009 11010 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 11011 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 11012 Removed the versions check from X509 routines when loading extensions: 11013 this allows certain broken certificates that don't set the version 11014 properly to be processed. 11015 [Steve Henson] 11016 11017 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 11018 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 11019 can still be regenerated with "make depend". 11020 [Ben Laurie] 11021 11022 *) Spelling mistake in C version of CAST-128. 11023 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 11024 11025 *) Changes to the error generation code. The perl script err-code.pl 11026 now reads in the old error codes and retains the old numbers, only 11027 adding new ones if necessary. It also only changes the .err files if new 11028 codes are added. The makefiles have been modified to only insert errors 11029 when needed (to avoid needlessly modifying header files). This is done 11030 by only inserting errors if the .err file is newer than the auto generated 11031 C file. To rebuild all the error codes from scratch (the old behaviour) 11032 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 11033 or delete all the .err files. 11034 [Steve Henson] 11035 11036 *) CAST-128 was incorrectly implemented for short keys. The C version has 11037 been fixed, but is untested. The assembler versions are also fixed, but 11038 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 11039 to regenerate it if needed. 11040 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 11041 Hagino <itojun@kame.net>] 11042 11043 *) File was opened incorrectly in randfile.c. 11044 [Ulf M��ller <ulf@fitug.de>] 11045 11046 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 11047 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 11048 GeneralizedTime. ASN1_TIME is the proper type used in certificates et 11049 al: it's just almost always a UTCTime. Note this patch adds new error 11050 codes so do a "make errors" if there are problems. 11051 [Steve Henson] 11052 11053 *) Correct Linux 1 recognition in config. 11054 [Ulf M��ller <ulf@fitug.de>] 11055 11056 *) Remove pointless MD5 hash when using DSA keys in ca. 11057 [Anonymous <nobody@replay.com>] 11058 11059 *) Generate an error if given an empty string as a cert directory. Also 11060 generate an error if handed NULL (previously returned 0 to indicate an 11061 error, but didn't set one). 11062 [Ben Laurie, reported by Anonymous <nobody@replay.com>] 11063 11064 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 11065 [Ben Laurie] 11066 11067 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 11068 parameters. This was causing a warning which killed off the Win32 compile. 11069 [Steve Henson] 11070 11071 *) Remove C++ style comments from crypto/bn/bn_local.h. 11072 [Neil Costigan <neil.costigan@celocom.com>] 11073 11074 *) The function OBJ_txt2nid was broken. It was supposed to return a nid 11075 based on a text string, looking up short and long names and finally 11076 "dot" format. The "dot" format stuff didn't work. Added new function 11077 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 11078 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 11079 OID is not part of the table. 11080 [Steve Henson] 11081 11082 *) Add prototypes to X509 lookup/verify methods, fixing a bug in 11083 X509_LOOKUP_by_alias(). 11084 [Ben Laurie] 11085 11086 *) Sort openssl functions by name. 11087 [Ben Laurie] 11088 11089 *) Get the gendsa program working (hopefully) and add it to app list. Remove 11090 encryption from sample DSA keys (in case anyone is interested the password 11091 was "1234"). 11092 [Steve Henson] 11093 11094 *) Make _all_ *_free functions accept a NULL pointer. 11095 [Frans Heymans <fheymans@isaserver.be>] 11096 11097 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 11098 NULL pointers. 11099 [Anonymous <nobody@replay.com>] 11100 11101 *) s_server should send the CAfile as acceptable CAs, not its own cert. 11102 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 11103 11104 *) Don't blow it for numeric -newkey arguments to apps/req. 11105 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 11106 11107 *) Temp key "for export" tests were wrong in s3_srvr.c. 11108 [Anonymous <nobody@replay.com>] 11109 11110 *) Add prototype for temp key callback functions 11111 SSL_CTX_set_tmp_{rsa,dh}_callback(). 11112 [Ben Laurie] 11113 11114 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 11115 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 11116 [Steve Henson] 11117 11118 *) X509_name_add_entry() freed the wrong thing after an error. 11119 [Arne Ansper <arne@ats.cyber.ee>] 11120 11121 *) rsa_eay.c would attempt to free a NULL context. 11122 [Arne Ansper <arne@ats.cyber.ee>] 11123 11124 *) BIO_s_socket() had a broken should_retry() on Windoze. 11125 [Arne Ansper <arne@ats.cyber.ee>] 11126 11127 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 11128 [Arne Ansper <arne@ats.cyber.ee>] 11129 11130 *) Make sure the already existing X509_STORE->depth variable is initialized 11131 in X509_STORE_new(), but document the fact that this variable is still 11132 unused in the certificate verification process. 11133 [Ralf S. Engelschall] 11134 11135 *) Fix the various library and apps files to free up pkeys obtained from 11136 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 11137 [Steve Henson] 11138 11139 *) Fix reference counting in X509_PUBKEY_get(). This makes 11140 demos/maurice/example2.c work, amongst others, probably. 11141 [Steve Henson and Ben Laurie] 11142 11143 *) First cut of a cleanup for apps/. First the `ssleay' program is now named 11144 `openssl' and second, the shortcut symlinks for the `openssl <command>' 11145 are no longer created. This way we have a single and consistent command 11146 line interface `openssl <command>', similar to `cvs <command>'. 11147 [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 11148 11149 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 11150 BIT STRING wrapper always have zero unused bits. 11151 [Steve Henson] 11152 11153 *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 11154 [Steve Henson] 11155 11156 *) Make the top-level INSTALL documentation easier to understand. 11157 [Paul Sutton] 11158 11159 *) Makefiles updated to exit if an error occurs in a sub-directory 11160 make (including if user presses ^C) [Paul Sutton] 11161 11162 *) Make Montgomery context stuff explicit in RSA data structure. 11163 [Ben Laurie] 11164 11165 *) Fix build order of pem and err to allow for generated pem.h. 11166 [Ben Laurie] 11167 11168 *) Fix renumbering bug in X509_NAME_delete_entry(). 11169 [Ben Laurie] 11170 11171 *) Enhanced the err-ins.pl script so it makes the error library number 11172 global and can add a library name. This is needed for external ASN1 and 11173 other error libraries. 11174 [Steve Henson] 11175 11176 *) Fixed sk_insert which never worked properly. 11177 [Steve Henson] 11178 11179 *) Fix ASN1 macros so they can handle indefinite length construted 11180 EXPLICIT tags. Some non standard certificates use these: they can now 11181 be read in. 11182 [Steve Henson] 11183 11184 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 11185 into a single doc/ssleay.txt bundle. This way the information is still 11186 preserved but no longer messes up this directory. Now it's new room for 11187 the new set of documenation files. 11188 [Ralf S. Engelschall] 11189 11190 *) SETs were incorrectly DER encoded. This was a major pain, because they 11191 shared code with SEQUENCEs, which aren't coded the same. This means that 11192 almost everything to do with SETs or SEQUENCEs has either changed name or 11193 number of arguments. 11194 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 11195 11196 *) Fix test data to work with the above. 11197 [Ben Laurie] 11198 11199 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 11200 was already fixed by Eric for 0.9.1 it seems. 11201 [Ben Laurie - pointed out by Ulf M��ller <ulf@fitug.de>] 11202 11203 *) Autodetect FreeBSD3. 11204 [Ben Laurie] 11205 11206 *) Fix various bugs in Configure. This affects the following platforms: 11207 nextstep 11208 ncr-scde 11209 unixware-2.0 11210 unixware-2.0-pentium 11211 sco5-cc. 11212 [Ben Laurie] 11213 11214 *) Eliminate generated files from CVS. Reorder tests to regenerate files 11215 before they are needed. 11216 [Ben Laurie] 11217 11218 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 11219 [Ben Laurie] 11220 11221 11222 Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 11223 11224 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 11225 changed SSLeay to OpenSSL in version strings. 11226 [Ralf S. Engelschall] 11227 11228 *) Some fixups to the top-level documents. 11229 [Paul Sutton] 11230 11231 *) Fixed the nasty bug where rsaref.h was not found under compile-time 11232 because the symlink to include/ was missing. 11233 [Ralf S. Engelschall] 11234 11235 *) Incorporated the popular no-RSA/DSA-only patches 11236 which allow to compile a RSA-free SSLeay. 11237 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 11238 11239 *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 11240 when "ssleay" is still not found. 11241 [Ralf S. Engelschall] 11242 11243 *) Added more platforms to Configure: Cray T3E, HPUX 11, 11244 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 11245 11246 *) Updated the README file. 11247 [Ralf S. Engelschall] 11248 11249 *) Added various .cvsignore files in the CVS repository subdirs 11250 to make a "cvs update" really silent. 11251 [Ralf S. Engelschall] 11252 11253 *) Recompiled the error-definition header files and added 11254 missing symbols to the Win32 linker tables. 11255 [Ralf S. Engelschall] 11256 11257 *) Cleaned up the top-level documents; 11258 o new files: CHANGES and LICENSE 11259 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 11260 o merged COPYRIGHT into LICENSE 11261 o removed obsolete TODO file 11262 o renamed MICROSOFT to INSTALL.W32 11263 [Ralf S. Engelschall] 11264 11265 *) Removed dummy files from the 0.9.1b source tree: 11266 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 11267 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 11268 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 11269 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 11270 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 11271 [Ralf S. Engelschall] 11272 11273 *) Added various platform portability fixes. 11274 [Mark J. Cox] 11275 11276 *) The Genesis of the OpenSSL rpject: 11277 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 11278 Young and Tim J. Hudson created while they were working for C2Net until 11279 summer 1998. 11280 [The OpenSSL Project] 11281 11282 11283 Changes between 0.9.0b and 0.9.1b [not released] 11284 11285 *) Updated a few CA certificates under certs/ 11286 [Eric A. Young] 11287 11288 *) Changed some BIGNUM api stuff. 11289 [Eric A. Young] 11290 11291 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 11292 DGUX x86, Linux Alpha, etc. 11293 [Eric A. Young] 11294 11295 *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 11296 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 11297 available). 11298 [Eric A. Young] 11299 11300 *) Add -strparse option to asn1pars program which parses nested 11301 binary structures 11302 [Dr Stephen Henson <shenson@bigfoot.com>] 11303 11304 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 11305 [Eric A. Young] 11306 11307 *) DSA fix for "ca" program. 11308 [Eric A. Young] 11309 11310 *) Added "-genkey" option to "dsaparam" program. 11311 [Eric A. Young] 11312 11313 *) Added RIPE MD160 (rmd160) message digest. 11314 [Eric A. Young] 11315 11316 *) Added -a (all) option to "ssleay version" command. 11317 [Eric A. Young] 11318 11319 *) Added PLATFORM define which is the id given to Configure. 11320 [Eric A. Young] 11321 11322 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 11323 [Eric A. Young] 11324 11325 *) Extended the ASN.1 parser routines. 11326 [Eric A. Young] 11327 11328 *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 11329 [Eric A. Young] 11330 11331 *) Added a BN_CTX to the BN library. 11332 [Eric A. Young] 11333 11334 *) Fixed the weak key values in DES library 11335 [Eric A. Young] 11336 11337 *) Changed API in EVP library for cipher aliases. 11338 [Eric A. Young] 11339 11340 *) Added support for RC2/64bit cipher. 11341 [Eric A. Young] 11342 11343 *) Converted the lhash library to the crypto/mem.c functions. 11344 [Eric A. Young] 11345 11346 *) Added more recognized ASN.1 object ids. 11347 [Eric A. Young] 11348 11349 *) Added more RSA padding checks for SSL/TLS. 11350 [Eric A. Young] 11351 11352 *) Added BIO proxy/filter functionality. 11353 [Eric A. Young] 11354 11355 *) Added extra_certs to SSL_CTX which can be used 11356 send extra CA certificates to the client in the CA cert chain sending 11357 process. It can be configured with SSL_CTX_add_extra_chain_cert(). 11358 [Eric A. Young] 11359 11360 *) Now Fortezza is denied in the authentication phase because 11361 this is key exchange mechanism is not supported by SSLeay at all. 11362 [Eric A. Young] 11363 11364 *) Additional PKCS1 checks. 11365 [Eric A. Young] 11366 11367 *) Support the string "TLSv1" for all TLS v1 ciphers. 11368 [Eric A. Young] 11369 11370 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 11371 ex_data index of the SSL context in the X509_STORE_CTX ex_data. 11372 [Eric A. Young] 11373 11374 *) Fixed a few memory leaks. 11375 [Eric A. Young] 11376 11377 *) Fixed various code and comment typos. 11378 [Eric A. Young] 11379 11380 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 11381 bytes sent in the client random. 11382 [Edward Bishop <ebishop@spyglass.com>] 11383 11384