CHANGES revision 325335
1 2 OpenSSL CHANGES 3 _______________ 4 5 Changes between 1.0.2k and 1.0.2l [25 May 2017] 6 7 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target 8 platform rather than 'mingw'. 9 [Richard Levitte] 10 11 Changes between 1.0.2j and 1.0.2k [26 Jan 2017] 12 13 *) Truncated packet could crash via OOB read 14 15 If one side of an SSL/TLS path is running on a 32-bit host and a specific 16 cipher is being used, then a truncated packet can cause that host to 17 perform an out-of-bounds read, usually resulting in a crash. 18 19 This issue was reported to OpenSSL by Robert ��wi��cki of Google. 20 (CVE-2017-3731) 21 [Andy Polyakov] 22 23 *) BN_mod_exp may produce incorrect results on x86_64 24 25 There is a carry propagating bug in the x86_64 Montgomery squaring 26 procedure. No EC algorithms are affected. Analysis suggests that attacks 27 against RSA and DSA as a result of this defect would be very difficult to 28 perform and are not believed likely. Attacks against DH are considered just 29 feasible (although very difficult) because most of the work necessary to 30 deduce information about a private key may be performed offline. The amount 31 of resources required for such an attack would be very significant and 32 likely only accessible to a limited number of attackers. An attacker would 33 additionally need online access to an unpatched system using the target 34 private key in a scenario with persistent DH parameters and a private 35 key that is shared between multiple clients. For example this can occur by 36 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very 37 similar to CVE-2015-3193 but must be treated as a separate problem. 38 39 This issue was reported to OpenSSL by the OSS-Fuzz project. 40 (CVE-2017-3732) 41 [Andy Polyakov] 42 43 *) Montgomery multiplication may produce incorrect results 44 45 There is a carry propagating bug in the Broadwell-specific Montgomery 46 multiplication procedure that handles input lengths divisible by, but 47 longer than 256 bits. Analysis suggests that attacks against RSA, DSA 48 and DH private keys are impossible. This is because the subroutine in 49 question is not used in operations with the private key itself and an input 50 of the attacker's direct choice. Otherwise the bug can manifest itself as 51 transient authentication and key negotiation failures or reproducible 52 erroneous outcome of public-key operations with specially crafted input. 53 Among EC algorithms only Brainpool P-512 curves are affected and one 54 presumably can attack ECDH key negotiation. Impact was not analyzed in 55 detail, because pre-requisites for attack are considered unlikely. Namely 56 multiple clients have to choose the curve in question and the server has to 57 share the private key among them, neither of which is default behaviour. 58 Even then only clients that chose the curve will be affected. 59 60 This issue was publicly reported as transient failures and was not 61 initially recognized as a security issue. Thanks to Richard Morgan for 62 providing reproducible case. 63 (CVE-2016-7055) 64 [Andy Polyakov] 65 66 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0 67 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to 68 prevent issues where no progress is being made and the peer continually 69 sends unrecognised record types, using up resources processing them. 70 [Matt Caswell] 71 72 Changes between 1.0.2i and 1.0.2j [26 Sep 2016] 73 74 *) Missing CRL sanity check 75 76 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0 77 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use 78 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception. 79 80 This issue only affects the OpenSSL 1.0.2i 81 (CVE-2016-7052) 82 [Matt Caswell] 83 84 Changes between 1.0.2h and 1.0.2i [22 Sep 2016] 85 86 *) OCSP Status Request extension unbounded memory growth 87 88 A malicious client can send an excessively large OCSP Status Request 89 extension. If that client continually requests renegotiation, sending a 90 large OCSP Status Request extension each time, then there will be unbounded 91 memory growth on the server. This will eventually lead to a Denial Of 92 Service attack through memory exhaustion. Servers with a default 93 configuration are vulnerable even if they do not support OCSP. Builds using 94 the "no-ocsp" build time option are not affected. 95 96 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 97 (CVE-2016-6304) 98 [Matt Caswell] 99 100 *) In order to mitigate the SWEET32 attack, the DES ciphers were moved from 101 HIGH to MEDIUM. 102 103 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan 104 Leurent (INRIA) 105 (CVE-2016-2183) 106 [Rich Salz] 107 108 *) OOB write in MDC2_Update() 109 110 An overflow can occur in MDC2_Update() either if called directly or 111 through the EVP_DigestUpdate() function using MDC2. If an attacker 112 is able to supply very large amounts of input data after a previous 113 call to EVP_EncryptUpdate() with a partial block then a length check 114 can overflow resulting in a heap corruption. 115 116 The amount of data needed is comparable to SIZE_MAX which is impractical 117 on most platforms. 118 119 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 120 (CVE-2016-6303) 121 [Stephen Henson] 122 123 *) Malformed SHA512 ticket DoS 124 125 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a 126 DoS attack where a malformed ticket will result in an OOB read which will 127 ultimately crash. 128 129 The use of SHA512 in TLS session tickets is comparatively rare as it requires 130 a custom server callback and ticket lookup mechanism. 131 132 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 133 (CVE-2016-6302) 134 [Stephen Henson] 135 136 *) OOB write in BN_bn2dec() 137 138 The function BN_bn2dec() does not check the return value of BN_div_word(). 139 This can cause an OOB write if an application uses this function with an 140 overly large BIGNUM. This could be a problem if an overly large certificate 141 or CRL is printed out from an untrusted source. TLS is not affected because 142 record limits will reject an oversized certificate before it is parsed. 143 144 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 145 (CVE-2016-2182) 146 [Stephen Henson] 147 148 *) OOB read in TS_OBJ_print_bio() 149 150 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is 151 the total length the OID text representation would use and not the amount 152 of data written. This will result in OOB reads when large OIDs are 153 presented. 154 155 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 156 (CVE-2016-2180) 157 [Stephen Henson] 158 159 *) Pointer arithmetic undefined behaviour 160 161 Avoid some undefined pointer arithmetic 162 163 A common idiom in the codebase is to check limits in the following manner: 164 "p + len > limit" 165 166 Where "p" points to some malloc'd data of SIZE bytes and 167 limit == p + SIZE 168 169 "len" here could be from some externally supplied data (e.g. from a TLS 170 message). 171 172 The rules of C pointer arithmetic are such that "p + len" is only well 173 defined where len <= SIZE. Therefore the above idiom is actually 174 undefined behaviour. 175 176 For example this could cause problems if some malloc implementation 177 provides an address for "p" such that "p + len" actually overflows for 178 values of len that are too big and therefore p + len < limit. 179 180 This issue was reported to OpenSSL by Guido Vranken 181 (CVE-2016-2177) 182 [Matt Caswell] 183 184 *) Constant time flag not preserved in DSA signing 185 186 Operations in the DSA signing algorithm should run in constant time in 187 order to avoid side channel attacks. A flaw in the OpenSSL DSA 188 implementation means that a non-constant time codepath is followed for 189 certain operations. This has been demonstrated through a cache-timing 190 attack to be sufficient for an attacker to recover the private DSA key. 191 192 This issue was reported by C��sar Pereida (Aalto University), Billy Brumley 193 (Tampere University of Technology), and Yuval Yarom (The University of 194 Adelaide and NICTA). 195 (CVE-2016-2178) 196 [C��sar Pereida] 197 198 *) DTLS buffered message DoS 199 200 In a DTLS connection where handshake messages are delivered out-of-order 201 those messages that OpenSSL is not yet ready to process will be buffered 202 for later use. Under certain circumstances, a flaw in the logic means that 203 those messages do not get removed from the buffer even though the handshake 204 has been completed. An attacker could force up to approx. 15 messages to 205 remain in the buffer when they are no longer required. These messages will 206 be cleared when the DTLS connection is closed. The default maximum size for 207 a message is 100k. Therefore the attacker could force an additional 1500k 208 to be consumed per connection. By opening many simulataneous connections an 209 attacker could cause a DoS attack through memory exhaustion. 210 211 This issue was reported to OpenSSL by Quan Luo. 212 (CVE-2016-2179) 213 [Matt Caswell] 214 215 *) DTLS replay protection DoS 216 217 A flaw in the DTLS replay attack protection mechanism means that records 218 that arrive for future epochs update the replay protection "window" before 219 the MAC for the record has been validated. This could be exploited by an 220 attacker by sending a record for the next epoch (which does not have to 221 decrypt or have a valid MAC), with a very large sequence number. This means 222 that all subsequent legitimate packets are dropped causing a denial of 223 service for a specific DTLS connection. 224 225 This issue was reported to OpenSSL by the OCAP audit team. 226 (CVE-2016-2181) 227 [Matt Caswell] 228 229 *) Certificate message OOB reads 230 231 In OpenSSL 1.0.2 and earlier some missing message length checks can result 232 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a 233 theoretical DoS risk but this has not been observed in practice on common 234 platforms. 235 236 The messages affected are client certificate, client certificate request 237 and server certificate. As a result the attack can only be performed 238 against a client or a server which enables client authentication. 239 240 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 241 (CVE-2016-6306) 242 [Stephen Henson] 243 244 Changes between 1.0.2g and 1.0.2h [3 May 2016] 245 246 *) Prevent padding oracle in AES-NI CBC MAC check 247 248 A MITM attacker can use a padding oracle attack to decrypt traffic 249 when the connection uses an AES CBC cipher and the server support 250 AES-NI. 251 252 This issue was introduced as part of the fix for Lucky 13 padding 253 attack (CVE-2013-0169). The padding check was rewritten to be in 254 constant time by making sure that always the same bytes are read and 255 compared against either the MAC or padding bytes. But it no longer 256 checked that there was enough data to have both the MAC and padding 257 bytes. 258 259 This issue was reported by Juraj Somorovsky using TLS-Attacker. 260 (CVE-2016-2107) 261 [Kurt Roeckx] 262 263 *) Fix EVP_EncodeUpdate overflow 264 265 An overflow can occur in the EVP_EncodeUpdate() function which is used for 266 Base64 encoding of binary data. If an attacker is able to supply very large 267 amounts of input data then a length check can overflow resulting in a heap 268 corruption. 269 270 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by 271 the PEM_write_bio* family of functions. These are mainly used within the 272 OpenSSL command line applications, so any application which processes data 273 from an untrusted source and outputs it as a PEM file should be considered 274 vulnerable to this issue. User applications that call these APIs directly 275 with large amounts of untrusted data may also be vulnerable. 276 277 This issue was reported by Guido Vranken. 278 (CVE-2016-2105) 279 [Matt Caswell] 280 281 *) Fix EVP_EncryptUpdate overflow 282 283 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker 284 is able to supply very large amounts of input data after a previous call to 285 EVP_EncryptUpdate() with a partial block then a length check can overflow 286 resulting in a heap corruption. Following an analysis of all OpenSSL 287 internal usage of the EVP_EncryptUpdate() function all usage is one of two 288 forms. The first form is where the EVP_EncryptUpdate() call is known to be 289 the first called function after an EVP_EncryptInit(), and therefore that 290 specific call must be safe. The second form is where the length passed to 291 EVP_EncryptUpdate() can be seen from the code to be some small value and 292 therefore there is no possibility of an overflow. Since all instances are 293 one of these two forms, it is believed that there can be no overflows in 294 internal code due to this problem. It should be noted that 295 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths. 296 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances 297 of these calls have also been analysed too and it is believed there are no 298 instances in internal usage where an overflow could occur. 299 300 This issue was reported by Guido Vranken. 301 (CVE-2016-2106) 302 [Matt Caswell] 303 304 *) Prevent ASN.1 BIO excessive memory allocation 305 306 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio() 307 a short invalid encoding can casuse allocation of large amounts of memory 308 potentially consuming excessive resources or exhausting memory. 309 310 Any application parsing untrusted data through d2i BIO functions is 311 affected. The memory based functions such as d2i_X509() are *not* affected. 312 Since the memory based functions are used by the TLS library, TLS 313 applications are not affected. 314 315 This issue was reported by Brian Carpenter. 316 (CVE-2016-2109) 317 [Stephen Henson] 318 319 *) EBCDIC overread 320 321 ASN1 Strings that are over 1024 bytes can cause an overread in applications 322 using the X509_NAME_oneline() function on EBCDIC systems. This could result 323 in arbitrary stack data being returned in the buffer. 324 325 This issue was reported by Guido Vranken. 326 (CVE-2016-2176) 327 [Matt Caswell] 328 329 *) Modify behavior of ALPN to invoke callback after SNI/servername 330 callback, such that updates to the SSL_CTX affect ALPN. 331 [Todd Short] 332 333 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the 334 default. 335 [Kurt Roeckx] 336 337 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the 338 methods are enabled and ssl2 is disabled the methods return NULL. 339 [Kurt Roeckx] 340 341 Changes between 1.0.2f and 1.0.2g [1 Mar 2016] 342 343 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL. 344 Builds that are not configured with "enable-weak-ssl-ciphers" will not 345 provide any "EXPORT" or "LOW" strength ciphers. 346 [Viktor Dukhovni] 347 348 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2 349 is by default disabled at build-time. Builds that are not configured with 350 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used, 351 users who want to negotiate SSLv2 via the version-flexible SSLv23_method() 352 will need to explicitly call either of: 353 354 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2); 355 or 356 SSL_clear_options(ssl, SSL_OP_NO_SSLv2); 357 358 as appropriate. Even if either of those is used, or the application 359 explicitly uses the version-specific SSLv2_method() or its client and 360 server variants, SSLv2 ciphers vulnerable to exhaustive search key 361 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT 362 ciphers, and SSLv2 56-bit DES are no longer available. 363 (CVE-2016-0800) 364 [Viktor Dukhovni] 365 366 *) Fix a double-free in DSA code 367 368 A double free bug was discovered when OpenSSL parses malformed DSA private 369 keys and could lead to a DoS attack or memory corruption for applications 370 that receive DSA private keys from untrusted sources. This scenario is 371 considered rare. 372 373 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using 374 libFuzzer. 375 (CVE-2016-0705) 376 [Stephen Henson] 377 378 *) Disable SRP fake user seed to address a server memory leak. 379 380 Add a new method SRP_VBASE_get1_by_user that handles the seed properly. 381 382 SRP_VBASE_get_by_user had inconsistent memory management behaviour. 383 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user 384 was changed to ignore the "fake user" SRP seed, even if the seed 385 is configured. 386 387 Users should use SRP_VBASE_get1_by_user instead. Note that in 388 SRP_VBASE_get1_by_user, caller must free the returned value. Note 389 also that even though configuring the SRP seed attempts to hide 390 invalid usernames by continuing the handshake with fake 391 credentials, this behaviour is not constant time and no strong 392 guarantees are made that the handshake is indistinguishable from 393 that of a valid user. 394 (CVE-2016-0798) 395 [Emilia K��sper] 396 397 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption 398 399 In the BN_hex2bn function the number of hex digits is calculated using an 400 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For 401 large values of |i| this can result in |bn_expand| not allocating any 402 memory because |i * 4| is negative. This can leave the internal BIGNUM data 403 field as NULL leading to a subsequent NULL ptr deref. For very large values 404 of |i|, the calculation |i * 4| could be a positive value smaller than |i|. 405 In this case memory is allocated to the internal BIGNUM data field, but it 406 is insufficiently sized leading to heap corruption. A similar issue exists 407 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn 408 is ever called by user applications with very large untrusted hex/dec data. 409 This is anticipated to be a rare occurrence. 410 411 All OpenSSL internal usage of these functions use data that is not expected 412 to be untrusted, e.g. config file data or application command line 413 arguments. If user developed applications generate config file data based 414 on untrusted data then it is possible that this could also lead to security 415 consequences. This is also anticipated to be rare. 416 417 This issue was reported to OpenSSL by Guido Vranken. 418 (CVE-2016-0797) 419 [Matt Caswell] 420 421 *) Fix memory issues in BIO_*printf functions 422 423 The internal |fmtstr| function used in processing a "%s" format string in 424 the BIO_*printf functions could overflow while calculating the length of a 425 string and cause an OOB read when printing very long strings. 426 427 Additionally the internal |doapr_outch| function can attempt to write to an 428 OOB memory location (at an offset from the NULL pointer) in the event of a 429 memory allocation failure. In 1.0.2 and below this could be caused where 430 the size of a buffer to be allocated is greater than INT_MAX. E.g. this 431 could be in processing a very long "%s" format string. Memory leaks can 432 also occur. 433 434 The first issue may mask the second issue dependent on compiler behaviour. 435 These problems could enable attacks where large amounts of untrusted data 436 is passed to the BIO_*printf functions. If applications use these functions 437 in this way then they could be vulnerable. OpenSSL itself uses these 438 functions when printing out human-readable dumps of ASN.1 data. Therefore 439 applications that print this data could be vulnerable if the data is from 440 untrusted sources. OpenSSL command line applications could also be 441 vulnerable where they print out ASN.1 data, or if untrusted data is passed 442 as command line arguments. 443 444 Libssl is not considered directly vulnerable. Additionally certificates etc 445 received via remote connections via libssl are also unlikely to be able to 446 trigger these issues because of message size limits enforced within libssl. 447 448 This issue was reported to OpenSSL Guido Vranken. 449 (CVE-2016-0799) 450 [Matt Caswell] 451 452 *) Side channel attack on modular exponentiation 453 454 A side-channel attack was found which makes use of cache-bank conflicts on 455 the Intel Sandy-Bridge microarchitecture which could lead to the recovery 456 of RSA keys. The ability to exploit this issue is limited as it relies on 457 an attacker who has control of code in a thread running on the same 458 hyper-threaded core as the victim thread which is performing decryptions. 459 460 This issue was reported to OpenSSL by Yuval Yarom, The University of 461 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and 462 Nadia Heninger, University of Pennsylvania with more information at 463 http://cachebleed.info. 464 (CVE-2016-0702) 465 [Andy Polyakov] 466 467 *) Change the req app to generate a 2048-bit RSA/DSA key by default, 468 if no keysize is specified with default_bits. This fixes an 469 omission in an earlier change that changed all RSA/DSA key generation 470 apps to use 2048 bits by default. 471 [Emilia K��sper] 472 473 Changes between 1.0.2e and 1.0.2f [28 Jan 2016] 474 475 *) DH small subgroups 476 477 Historically OpenSSL only ever generated DH parameters based on "safe" 478 primes. More recently (in version 1.0.2) support was provided for 479 generating X9.42 style parameter files such as those required for RFC 5114 480 support. The primes used in such files may not be "safe". Where an 481 application is using DH configured with parameters based on primes that are 482 not "safe" then an attacker could use this fact to find a peer's private 483 DH exponent. This attack requires that the attacker complete multiple 484 handshakes in which the peer uses the same private DH exponent. For example 485 this could be used to discover a TLS server's private DH exponent if it's 486 reusing the private DH exponent or it's using a static DH ciphersuite. 487 488 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in 489 TLS. It is not on by default. If the option is not set then the server 490 reuses the same private DH exponent for the life of the server process and 491 would be vulnerable to this attack. It is believed that many popular 492 applications do set this option and would therefore not be at risk. 493 494 The fix for this issue adds an additional check where a "q" parameter is 495 available (as is the case in X9.42 based parameters). This detects the 496 only known attack, and is the only possible defense for static DH 497 ciphersuites. This could have some performance impact. 498 499 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by 500 default and cannot be disabled. This could have some performance impact. 501 502 This issue was reported to OpenSSL by Antonio Sanso (Adobe). 503 (CVE-2016-0701) 504 [Matt Caswell] 505 506 *) SSLv2 doesn't block disabled ciphers 507 508 A malicious client can negotiate SSLv2 ciphers that have been disabled on 509 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have 510 been disabled, provided that the SSLv2 protocol was not also disabled via 511 SSL_OP_NO_SSLv2. 512 513 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram 514 and Sebastian Schinzel. 515 (CVE-2015-3197) 516 [Viktor Dukhovni] 517 518 *) Reject DH handshakes with parameters shorter than 1024 bits. 519 [Kurt Roeckx] 520 521 Changes between 1.0.2d and 1.0.2e [3 Dec 2015] 522 523 *) BN_mod_exp may produce incorrect results on x86_64 524 525 There is a carry propagating bug in the x86_64 Montgomery squaring 526 procedure. No EC algorithms are affected. Analysis suggests that attacks 527 against RSA and DSA as a result of this defect would be very difficult to 528 perform and are not believed likely. Attacks against DH are considered just 529 feasible (although very difficult) because most of the work necessary to 530 deduce information about a private key may be performed offline. The amount 531 of resources required for such an attack would be very significant and 532 likely only accessible to a limited number of attackers. An attacker would 533 additionally need online access to an unpatched system using the target 534 private key in a scenario with persistent DH parameters and a private 535 key that is shared between multiple clients. For example this can occur by 536 default in OpenSSL DHE based SSL/TLS ciphersuites. 537 538 This issue was reported to OpenSSL by Hanno B��ck. 539 (CVE-2015-3193) 540 [Andy Polyakov] 541 542 *) Certificate verify crash with missing PSS parameter 543 544 The signature verification routines will crash with a NULL pointer 545 dereference if presented with an ASN.1 signature using the RSA PSS 546 algorithm and absent mask generation function parameter. Since these 547 routines are used to verify certificate signature algorithms this can be 548 used to crash any certificate verification operation and exploited in a 549 DoS attack. Any application which performs certificate verification is 550 vulnerable including OpenSSL clients and servers which enable client 551 authentication. 552 553 This issue was reported to OpenSSL by Lo��c Jonas Etienne (Qnective AG). 554 (CVE-2015-3194) 555 [Stephen Henson] 556 557 *) X509_ATTRIBUTE memory leak 558 559 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak 560 memory. This structure is used by the PKCS#7 and CMS routines so any 561 application which reads PKCS#7 or CMS data from untrusted sources is 562 affected. SSL/TLS is not affected. 563 564 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using 565 libFuzzer. 566 (CVE-2015-3195) 567 [Stephen Henson] 568 569 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs. 570 This changes the decoding behaviour for some invalid messages, 571 though the change is mostly in the more lenient direction, and 572 legacy behaviour is preserved as much as possible. 573 [Emilia K��sper] 574 575 *) In DSA_generate_parameters_ex, if the provided seed is too short, 576 use a random seed, as already documented. 577 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>] 578 579 Changes between 1.0.2c and 1.0.2d [9 Jul 2015] 580 581 *) Alternate chains certificate forgery 582 583 During certificate verfification, OpenSSL will attempt to find an 584 alternative certificate chain if the first attempt to build such a chain 585 fails. An error in the implementation of this logic can mean that an 586 attacker could cause certain checks on untrusted certificates to be 587 bypassed, such as the CA flag, enabling them to use a valid leaf 588 certificate to act as a CA and "issue" an invalid certificate. 589 590 This issue was reported to OpenSSL by Adam Langley/David Benjamin 591 (Google/BoringSSL). 592 (CVE-2015-1793) 593 [Matt Caswell] 594 595 *) Race condition handling PSK identify hint 596 597 If PSK identity hints are received by a multi-threaded client then 598 the values are wrongly updated in the parent SSL_CTX structure. This can 599 result in a race condition potentially leading to a double free of the 600 identify hint data. 601 (CVE-2015-3196) 602 [Stephen Henson] 603 604 Changes between 1.0.2b and 1.0.2c [12 Jun 2015] 605 606 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI 607 incompatibility in the handling of HMAC. The previous ABI has now been 608 restored. 609 610 Changes between 1.0.2a and 1.0.2b [11 Jun 2015] 611 612 *) Malformed ECParameters causes infinite loop 613 614 When processing an ECParameters structure OpenSSL enters an infinite loop 615 if the curve specified is over a specially malformed binary polynomial 616 field. 617 618 This can be used to perform denial of service against any 619 system which processes public keys, certificate requests or 620 certificates. This includes TLS clients and TLS servers with 621 client authentication enabled. 622 623 This issue was reported to OpenSSL by Joseph Barr-Pixton. 624 (CVE-2015-1788) 625 [Andy Polyakov] 626 627 *) Exploitable out-of-bounds read in X509_cmp_time 628 629 X509_cmp_time does not properly check the length of the ASN1_TIME 630 string and can read a few bytes out of bounds. In addition, 631 X509_cmp_time accepts an arbitrary number of fractional seconds in the 632 time string. 633 634 An attacker can use this to craft malformed certificates and CRLs of 635 various sizes and potentially cause a segmentation fault, resulting in 636 a DoS on applications that verify certificates or CRLs. TLS clients 637 that verify CRLs are affected. TLS clients and servers with client 638 authentication enabled may be affected if they use custom verification 639 callbacks. 640 641 This issue was reported to OpenSSL by Robert Swiecki (Google), and 642 independently by Hanno B��ck. 643 (CVE-2015-1789) 644 [Emilia K��sper] 645 646 *) PKCS7 crash with missing EnvelopedContent 647 648 The PKCS#7 parsing code does not handle missing inner EncryptedContent 649 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs 650 with missing content and trigger a NULL pointer dereference on parsing. 651 652 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7 653 structures from untrusted sources are affected. OpenSSL clients and 654 servers are not affected. 655 656 This issue was reported to OpenSSL by Michal Zalewski (Google). 657 (CVE-2015-1790) 658 [Emilia K��sper] 659 660 *) CMS verify infinite loop with unknown hash function 661 662 When verifying a signedData message the CMS code can enter an infinite loop 663 if presented with an unknown hash function OID. This can be used to perform 664 denial of service against any system which verifies signedData messages using 665 the CMS code. 666 This issue was reported to OpenSSL by Johannes Bauer. 667 (CVE-2015-1792) 668 [Stephen Henson] 669 670 *) Race condition handling NewSessionTicket 671 672 If a NewSessionTicket is received by a multi-threaded client when attempting to 673 reuse a previous ticket then a race condition can occur potentially leading to 674 a double free of the ticket data. 675 (CVE-2015-1791) 676 [Matt Caswell] 677 678 *) Removed support for the two export grade static DH ciphersuites 679 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites 680 were newly added (along with a number of other static DH ciphersuites) to 681 1.0.2. However the two export ones have *never* worked since they were 682 introduced. It seems strange in any case to be adding new export 683 ciphersuites, and given "logjam" it also does not seem correct to fix them. 684 [Matt Caswell] 685 686 *) Only support 256-bit or stronger elliptic curves with the 687 'ecdh_auto' setting (server) or by default (client). Of supported 688 curves, prefer P-256 (both). 689 [Emilia Kasper] 690 691 *) Reject DH handshakes with parameters shorter than 768 bits. 692 [Kurt Roeckx and Emilia Kasper] 693 694 Changes between 1.0.2 and 1.0.2a [19 Mar 2015] 695 696 *) ClientHello sigalgs DoS fix 697 698 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an 699 invalid signature algorithms extension a NULL pointer dereference will 700 occur. This can be exploited in a DoS attack against the server. 701 702 This issue was was reported to OpenSSL by David Ramos of Stanford 703 University. 704 (CVE-2015-0291) 705 [Stephen Henson and Matt Caswell] 706 707 *) Multiblock corrupted pointer fix 708 709 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This 710 feature only applies on 64 bit x86 architecture platforms that support AES 711 NI instructions. A defect in the implementation of "multiblock" can cause 712 OpenSSL's internal write buffer to become incorrectly set to NULL when 713 using non-blocking IO. Typically, when the user application is using a 714 socket BIO for writing, this will only result in a failed connection. 715 However if some other BIO is used then it is likely that a segmentation 716 fault will be triggered, thus enabling a potential DoS attack. 717 718 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller. 719 (CVE-2015-0290) 720 [Matt Caswell] 721 722 *) Segmentation fault in DTLSv1_listen fix 723 724 The DTLSv1_listen function is intended to be stateless and processes the 725 initial ClientHello from many peers. It is common for user code to loop 726 over the call to DTLSv1_listen until a valid ClientHello is received with 727 an associated cookie. A defect in the implementation of DTLSv1_listen means 728 that state is preserved in the SSL object from one invocation to the next 729 that can lead to a segmentation fault. Errors processing the initial 730 ClientHello can trigger this scenario. An example of such an error could be 731 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only 732 server. 733 734 This issue was reported to OpenSSL by Per Allansson. 735 (CVE-2015-0207) 736 [Matt Caswell] 737 738 *) Segmentation fault in ASN1_TYPE_cmp fix 739 740 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is 741 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check 742 certificate signature algorithm consistency this can be used to crash any 743 certificate verification operation and exploited in a DoS attack. Any 744 application which performs certificate verification is vulnerable including 745 OpenSSL clients and servers which enable client authentication. 746 (CVE-2015-0286) 747 [Stephen Henson] 748 749 *) Segmentation fault for invalid PSS parameters fix 750 751 The signature verification routines will crash with a NULL pointer 752 dereference if presented with an ASN.1 signature using the RSA PSS 753 algorithm and invalid parameters. Since these routines are used to verify 754 certificate signature algorithms this can be used to crash any 755 certificate verification operation and exploited in a DoS attack. Any 756 application which performs certificate verification is vulnerable including 757 OpenSSL clients and servers which enable client authentication. 758 759 This issue was was reported to OpenSSL by Brian Carpenter. 760 (CVE-2015-0208) 761 [Stephen Henson] 762 763 *) ASN.1 structure reuse memory corruption fix 764 765 Reusing a structure in ASN.1 parsing may allow an attacker to cause 766 memory corruption via an invalid write. Such reuse is and has been 767 strongly discouraged and is believed to be rare. 768 769 Applications that parse structures containing CHOICE or ANY DEFINED BY 770 components may be affected. Certificate parsing (d2i_X509 and related 771 functions) are however not affected. OpenSSL clients and servers are 772 not affected. 773 (CVE-2015-0287) 774 [Stephen Henson] 775 776 *) PKCS7 NULL pointer dereferences fix 777 778 The PKCS#7 parsing code does not handle missing outer ContentInfo 779 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with 780 missing content and trigger a NULL pointer dereference on parsing. 781 782 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or 783 otherwise parse PKCS#7 structures from untrusted sources are 784 affected. OpenSSL clients and servers are not affected. 785 786 This issue was reported to OpenSSL by Michal Zalewski (Google). 787 (CVE-2015-0289) 788 [Emilia K��sper] 789 790 *) DoS via reachable assert in SSLv2 servers fix 791 792 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in 793 servers that both support SSLv2 and enable export cipher suites by sending 794 a specially crafted SSLv2 CLIENT-MASTER-KEY message. 795 796 This issue was discovered by Sean Burford (Google) and Emilia K��sper 797 (OpenSSL development team). 798 (CVE-2015-0293) 799 [Emilia K��sper] 800 801 *) Empty CKE with client auth and DHE fix 802 803 If client auth is used then a server can seg fault in the event of a DHE 804 ciphersuite being selected and a zero length ClientKeyExchange message 805 being sent by the client. This could be exploited in a DoS attack. 806 (CVE-2015-1787) 807 [Matt Caswell] 808 809 *) Handshake with unseeded PRNG fix 810 811 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake 812 with an unseeded PRNG. The conditions are: 813 - The client is on a platform where the PRNG has not been seeded 814 automatically, and the user has not seeded manually 815 - A protocol specific client method version has been used (i.e. not 816 SSL_client_methodv23) 817 - A ciphersuite is used that does not require additional random data from 818 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA). 819 820 If the handshake succeeds then the client random that has been used will 821 have been generated from a PRNG with insufficient entropy and therefore the 822 output may be predictable. 823 824 For example using the following command with an unseeded openssl will 825 succeed on an unpatched platform: 826 827 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA 828 (CVE-2015-0285) 829 [Matt Caswell] 830 831 *) Use After Free following d2i_ECPrivatekey error fix 832 833 A malformed EC private key file consumed via the d2i_ECPrivateKey function 834 could cause a use after free condition. This, in turn, could cause a double 835 free in several private key parsing functions (such as d2i_PrivateKey 836 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption 837 for applications that receive EC private keys from untrusted 838 sources. This scenario is considered rare. 839 840 This issue was discovered by the BoringSSL project and fixed in their 841 commit 517073cd4b. 842 (CVE-2015-0209) 843 [Matt Caswell] 844 845 *) X509_to_X509_REQ NULL pointer deref fix 846 847 The function X509_to_X509_REQ will crash with a NULL pointer dereference if 848 the certificate key is invalid. This function is rarely used in practice. 849 850 This issue was discovered by Brian Carpenter. 851 (CVE-2015-0288) 852 [Stephen Henson] 853 854 *) Removed the export ciphers from the DEFAULT ciphers 855 [Kurt Roeckx] 856 857 Changes between 1.0.1l and 1.0.2 [22 Jan 2015] 858 859 *) Change RSA and DH/DSA key generation apps to generate 2048-bit 860 keys by default. 861 [Kurt Roeckx] 862 863 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g. 864 ARMv5 through ARMv8, as opposite to "locking" it to single one. 865 So far those who have to target multiple plaforms would compromise 866 and argue that binary targeting say ARMv5 would still execute on 867 ARMv8. "Universal" build resolves this compromise by providing 868 near-optimal performance even on newer platforms. 869 [Andy Polyakov] 870 871 *) Accelerated NIST P-256 elliptic curve implementation for x86_64 872 (other platforms pending). 873 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov] 874 875 *) Add support for the SignedCertificateTimestampList certificate and 876 OCSP response extensions from RFC6962. 877 [Rob Stradling] 878 879 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) 880 for corner cases. (Certain input points at infinity could lead to 881 bogus results, with non-infinity inputs mapped to infinity too.) 882 [Bodo Moeller] 883 884 *) Initial support for PowerISA 2.0.7, first implemented in POWER8. 885 This covers AES, SHA256/512 and GHASH. "Initial" means that most 886 common cases are optimized and there still is room for further 887 improvements. Vector Permutation AES for Altivec is also added. 888 [Andy Polyakov] 889 890 *) Add support for little-endian ppc64 Linux target. 891 [Marcelo Cerri (IBM)] 892 893 *) Initial support for AMRv8 ISA crypto extensions. This covers AES, 894 SHA1, SHA256 and GHASH. "Initial" means that most common cases 895 are optimized and there still is room for further improvements. 896 Both 32- and 64-bit modes are supported. 897 [Andy Polyakov, Ard Biesheuvel (Linaro)] 898 899 *) Improved ARMv7 NEON support. 900 [Andy Polyakov] 901 902 *) Support for SPARC Architecture 2011 crypto extensions, first 903 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1, 904 SHA256/512, MD5, GHASH and modular exponentiation. 905 [Andy Polyakov, David Miller] 906 907 *) Accelerated modular exponentiation for Intel processors, a.k.a. 908 RSAZ. 909 [Shay Gueron & Vlad Krasnov (Intel Corp)] 910 911 *) Support for new and upcoming Intel processors, including AVX2, 912 BMI and SHA ISA extensions. This includes additional "stitched" 913 implementations, AESNI-SHA256 and GCM, and multi-buffer support 914 for TLS encrypt. 915 916 This work was sponsored by Intel Corp. 917 [Andy Polyakov] 918 919 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method() 920 supports both DTLS 1.2 and 1.0 and should use whatever version the peer 921 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only. 922 [Steve Henson] 923 924 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file(): 925 this fixes a limiation in previous versions of OpenSSL. 926 [Steve Henson] 927 928 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest, 929 MGF1 digest and OAEP label. 930 [Steve Henson] 931 932 *) Add EVP support for key wrapping algorithms, to avoid problems with 933 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in 934 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap 935 algorithms and include tests cases. 936 [Steve Henson] 937 938 *) Add functions to allocate and set the fields of an ECDSA_METHOD 939 structure. 940 [Douglas E. Engert, Steve Henson] 941 942 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the 943 difference in days and seconds between two tm or ASN1_TIME structures. 944 [Steve Henson] 945 946 *) Add -rev test option to s_server to just reverse order of characters 947 received by client and send back to server. Also prints an abbreviated 948 summary of the connection parameters. 949 [Steve Henson] 950 951 *) New option -brief for s_client and s_server to print out a brief summary 952 of connection parameters. 953 [Steve Henson] 954 955 *) Add callbacks for arbitrary TLS extensions. 956 [Trevor Perrin <trevp@trevp.net> and Ben Laurie] 957 958 *) New option -crl_download in several openssl utilities to download CRLs 959 from CRLDP extension in certificates. 960 [Steve Henson] 961 962 *) New options -CRL and -CRLform for s_client and s_server for CRLs. 963 [Steve Henson] 964 965 *) New function X509_CRL_diff to generate a delta CRL from the difference 966 of two full CRLs. Add support to "crl" utility. 967 [Steve Henson] 968 969 *) New functions to set lookup_crls function and to retrieve 970 X509_STORE from X509_STORE_CTX. 971 [Steve Henson] 972 973 *) Print out deprecated issuer and subject unique ID fields in 974 certificates. 975 [Steve Henson] 976 977 *) Extend OCSP I/O functions so they can be used for simple general purpose 978 HTTP as well as OCSP. New wrapper function which can be used to download 979 CRLs using the OCSP API. 980 [Steve Henson] 981 982 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs. 983 [Steve Henson] 984 985 *) SSL_CONF* functions. These provide a common framework for application 986 configuration using configuration files or command lines. 987 [Steve Henson] 988 989 *) SSL/TLS tracing code. This parses out SSL/TLS records using the 990 message callback and prints the results. Needs compile time option 991 "enable-ssl-trace". New options to s_client and s_server to enable 992 tracing. 993 [Steve Henson] 994 995 *) New ctrl and macro to retrieve supported points extensions. 996 Print out extension in s_server and s_client. 997 [Steve Henson] 998 999 *) New functions to retrieve certificate signature and signature 1000 OID NID. 1001 [Steve Henson] 1002 1003 *) Add functions to retrieve and manipulate the raw cipherlist sent by a 1004 client to OpenSSL. 1005 [Steve Henson] 1006 1007 *) New Suite B modes for TLS code. These use and enforce the requirements 1008 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and 1009 only use Suite B curves. The Suite B modes can be set by using the 1010 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring. 1011 [Steve Henson] 1012 1013 *) New chain verification flags for Suite B levels of security. Check 1014 algorithms are acceptable when flags are set in X509_verify_cert. 1015 [Steve Henson] 1016 1017 *) Make tls1_check_chain return a set of flags indicating checks passed 1018 by a certificate chain. Add additional tests to handle client 1019 certificates: checks for matching certificate type and issuer name 1020 comparison. 1021 [Steve Henson] 1022 1023 *) If an attempt is made to use a signature algorithm not in the peer 1024 preference list abort the handshake. If client has no suitable 1025 signature algorithms in response to a certificate request do not 1026 use the certificate. 1027 [Steve Henson] 1028 1029 *) If server EC tmp key is not in client preference list abort handshake. 1030 [Steve Henson] 1031 1032 *) Add support for certificate stores in CERT structure. This makes it 1033 possible to have different stores per SSL structure or one store in 1034 the parent SSL_CTX. Include distint stores for certificate chain 1035 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN 1036 to build and store a certificate chain in CERT structure: returing 1037 an error if the chain cannot be built: this will allow applications 1038 to test if a chain is correctly configured. 1039 1040 Note: if the CERT based stores are not set then the parent SSL_CTX 1041 store is used to retain compatibility with existing behaviour. 1042 1043 [Steve Henson] 1044 1045 *) New function ssl_set_client_disabled to set a ciphersuite disabled 1046 mask based on the current session, check mask when sending client 1047 hello and checking the requested ciphersuite. 1048 [Steve Henson] 1049 1050 *) New ctrls to retrieve and set certificate types in a certificate 1051 request message. Print out received values in s_client. If certificate 1052 types is not set with custom values set sensible values based on 1053 supported signature algorithms. 1054 [Steve Henson] 1055 1056 *) Support for distinct client and server supported signature algorithms. 1057 [Steve Henson] 1058 1059 *) Add certificate callback. If set this is called whenever a certificate 1060 is required by client or server. An application can decide which 1061 certificate chain to present based on arbitrary criteria: for example 1062 supported signature algorithms. Add very simple example to s_server. 1063 This fixes many of the problems and restrictions of the existing client 1064 certificate callback: for example you can now clear an existing 1065 certificate and specify the whole chain. 1066 [Steve Henson] 1067 1068 *) Add new "valid_flags" field to CERT_PKEY structure which determines what 1069 the certificate can be used for (if anything). Set valid_flags field 1070 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used 1071 to have similar checks in it. 1072 1073 Add new "cert_flags" field to CERT structure and include a "strict mode". 1074 This enforces some TLS certificate requirements (such as only permitting 1075 certificate signature algorithms contained in the supported algorithms 1076 extension) which some implementations ignore: this option should be used 1077 with caution as it could cause interoperability issues. 1078 [Steve Henson] 1079 1080 *) Update and tidy signature algorithm extension processing. Work out 1081 shared signature algorithms based on preferences and peer algorithms 1082 and print them out in s_client and s_server. Abort handshake if no 1083 shared signature algorithms. 1084 [Steve Henson] 1085 1086 *) Add new functions to allow customised supported signature algorithms 1087 for SSL and SSL_CTX structures. Add options to s_client and s_server 1088 to support them. 1089 [Steve Henson] 1090 1091 *) New function SSL_certs_clear() to delete all references to certificates 1092 from an SSL structure. Before this once a certificate had been added 1093 it couldn't be removed. 1094 [Steve Henson] 1095 1096 *) Integrate hostname, email address and IP address checking with certificate 1097 verification. New verify options supporting checking in opensl utility. 1098 [Steve Henson] 1099 1100 *) Fixes and wildcard matching support to hostname and email checking 1101 functions. Add manual page. 1102 [Florian Weimer (Red Hat Product Security Team)] 1103 1104 *) New functions to check a hostname email or IP address against a 1105 certificate. Add options x509 utility to print results of checks against 1106 a certificate. 1107 [Steve Henson] 1108 1109 *) Fix OCSP checking. 1110 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie] 1111 1112 *) Initial experimental support for explicitly trusted non-root CAs. 1113 OpenSSL still tries to build a complete chain to a root but if an 1114 intermediate CA has a trust setting included that is used. The first 1115 setting is used: whether to trust (e.g., -addtrust option to the x509 1116 utility) or reject. 1117 [Steve Henson] 1118 1119 *) Add -trusted_first option which attempts to find certificates in the 1120 trusted store even if an untrusted chain is also supplied. 1121 [Steve Henson] 1122 1123 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE, 1124 platform support for Linux and Android. 1125 [Andy Polyakov] 1126 1127 *) Support for linux-x32, ILP32 environment in x86_64 framework. 1128 [Andy Polyakov] 1129 1130 *) Experimental multi-implementation support for FIPS capable OpenSSL. 1131 When in FIPS mode the approved implementations are used as normal, 1132 when not in FIPS mode the internal unapproved versions are used instead. 1133 This means that the FIPS capable OpenSSL isn't forced to use the 1134 (often lower perfomance) FIPS implementations outside FIPS mode. 1135 [Steve Henson] 1136 1137 *) Transparently support X9.42 DH parameters when calling 1138 PEM_read_bio_DHparameters. This means existing applications can handle 1139 the new parameter format automatically. 1140 [Steve Henson] 1141 1142 *) Initial experimental support for X9.42 DH parameter format: mainly 1143 to support use of 'q' parameter for RFC5114 parameters. 1144 [Steve Henson] 1145 1146 *) Add DH parameters from RFC5114 including test data to dhtest. 1147 [Steve Henson] 1148 1149 *) Support for automatic EC temporary key parameter selection. If enabled 1150 the most preferred EC parameters are automatically used instead of 1151 hardcoded fixed parameters. Now a server just has to call: 1152 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically 1153 support ECDH and use the most appropriate parameters. 1154 [Steve Henson] 1155 1156 *) Enhance and tidy EC curve and point format TLS extension code. Use 1157 static structures instead of allocation if default values are used. 1158 New ctrls to set curves we wish to support and to retrieve shared curves. 1159 Print out shared curves in s_server. New options to s_server and s_client 1160 to set list of supported curves. 1161 [Steve Henson] 1162 1163 *) New ctrls to retrieve supported signature algorithms and 1164 supported curve values as an array of NIDs. Extend openssl utility 1165 to print out received values. 1166 [Steve Henson] 1167 1168 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert 1169 between NIDs and the more common NIST names such as "P-256". Enhance 1170 ecparam utility and ECC method to recognise the NIST names for curves. 1171 [Steve Henson] 1172 1173 *) Enhance SSL/TLS certificate chain handling to support different 1174 chains for each certificate instead of one chain in the parent SSL_CTX. 1175 [Steve Henson] 1176 1177 *) Support for fixed DH ciphersuite client authentication: where both 1178 server and client use DH certificates with common parameters. 1179 [Steve Henson] 1180 1181 *) Support for fixed DH ciphersuites: those requiring DH server 1182 certificates. 1183 [Steve Henson] 1184 1185 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of 1186 the certificate. 1187 Note: Related 1.0.2-beta specific macros X509_get_cert_info, 1188 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and 1189 X509_CINF_get_signature were reverted post internal team review. 1190 1191 Changes between 1.0.1k and 1.0.1l [15 Jan 2015] 1192 1193 *) Build fixes for the Windows and OpenVMS platforms 1194 [Matt Caswell and Richard Levitte] 1195 1196 Changes between 1.0.1j and 1.0.1k [8 Jan 2015] 1197 1198 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS 1199 message can cause a segmentation fault in OpenSSL due to a NULL pointer 1200 dereference. This could lead to a Denial Of Service attack. Thanks to 1201 Markus Stenberg of Cisco Systems, Inc. for reporting this issue. 1202 (CVE-2014-3571) 1203 [Steve Henson] 1204 1205 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the 1206 dtls1_buffer_record function under certain conditions. In particular this 1207 could occur if an attacker sent repeated DTLS records with the same 1208 sequence number but for the next epoch. The memory leak could be exploited 1209 by an attacker in a Denial of Service attack through memory exhaustion. 1210 Thanks to Chris Mueller for reporting this issue. 1211 (CVE-2015-0206) 1212 [Matt Caswell] 1213 1214 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is 1215 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl 1216 method would be set to NULL which could later result in a NULL pointer 1217 dereference. Thanks to Frank Schmirler for reporting this issue. 1218 (CVE-2014-3569) 1219 [Kurt Roeckx] 1220 1221 *) Abort handshake if server key exchange message is omitted for ephemeral 1222 ECDH ciphersuites. 1223 1224 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for 1225 reporting this issue. 1226 (CVE-2014-3572) 1227 [Steve Henson] 1228 1229 *) Remove non-export ephemeral RSA code on client and server. This code 1230 violated the TLS standard by allowing the use of temporary RSA keys in 1231 non-export ciphersuites and could be used by a server to effectively 1232 downgrade the RSA key length used to a value smaller than the server 1233 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at 1234 INRIA or reporting this issue. 1235 (CVE-2015-0204) 1236 [Steve Henson] 1237 1238 *) Fixed issue where DH client certificates are accepted without verification. 1239 An OpenSSL server will accept a DH certificate for client authentication 1240 without the certificate verify message. This effectively allows a client to 1241 authenticate without the use of a private key. This only affects servers 1242 which trust a client certificate authority which issues certificates 1243 containing DH keys: these are extremely rare and hardly ever encountered. 1244 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting 1245 this issue. 1246 (CVE-2015-0205) 1247 [Steve Henson] 1248 1249 *) Ensure that the session ID context of an SSL is updated when its 1250 SSL_CTX is updated via SSL_set_SSL_CTX. 1251 1252 The session ID context is typically set from the parent SSL_CTX, 1253 and can vary with the CTX. 1254 [Adam Langley] 1255 1256 *) Fix various certificate fingerprint issues. 1257 1258 By using non-DER or invalid encodings outside the signed portion of a 1259 certificate the fingerprint can be changed without breaking the signature. 1260 Although no details of the signed portion of the certificate can be changed 1261 this can cause problems with some applications: e.g. those using the 1262 certificate fingerprint for blacklists. 1263 1264 1. Reject signatures with non zero unused bits. 1265 1266 If the BIT STRING containing the signature has non zero unused bits reject 1267 the signature. All current signature algorithms require zero unused bits. 1268 1269 2. Check certificate algorithm consistency. 1270 1271 Check the AlgorithmIdentifier inside TBS matches the one in the 1272 certificate signature. NB: this will result in signature failure 1273 errors for some broken certificates. 1274 1275 Thanks to Konrad Kraszewski from Google for reporting this issue. 1276 1277 3. Check DSA/ECDSA signatures use DER. 1278 1279 Reencode DSA/ECDSA signatures and compare with the original received 1280 signature. Return an error if there is a mismatch. 1281 1282 This will reject various cases including garbage after signature 1283 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS 1284 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs 1285 (negative or with leading zeroes). 1286 1287 Further analysis was conducted and fixes were developed by Stephen Henson 1288 of the OpenSSL core team. 1289 1290 (CVE-2014-8275) 1291 [Steve Henson] 1292 1293 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect 1294 results on some platforms, including x86_64. This bug occurs at random 1295 with a very low probability, and is not known to be exploitable in any 1296 way, though its exact impact is difficult to determine. Thanks to Pieter 1297 Wuille (Blockstream) who reported this issue and also suggested an initial 1298 fix. Further analysis was conducted by the OpenSSL development team and 1299 Adam Langley of Google. The final fix was developed by Andy Polyakov of 1300 the OpenSSL core team. 1301 (CVE-2014-3570) 1302 [Andy Polyakov] 1303 1304 *) Do not resume sessions on the server if the negotiated protocol 1305 version does not match the session's version. Resuming with a different 1306 version, while not strictly forbidden by the RFC, is of questionable 1307 sanity and breaks all known clients. 1308 [David Benjamin, Emilia K��sper] 1309 1310 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject 1311 early CCS messages during renegotiation. (Note that because 1312 renegotiation is encrypted, this early CCS was not exploitable.) 1313 [Emilia K��sper] 1314 1315 *) Tighten client-side session ticket handling during renegotiation: 1316 ensure that the client only accepts a session ticket if the server sends 1317 the extension anew in the ServerHello. Previously, a TLS client would 1318 reuse the old extension state and thus accept a session ticket if one was 1319 announced in the initial ServerHello. 1320 1321 Similarly, ensure that the client requires a session ticket if one 1322 was advertised in the ServerHello. Previously, a TLS client would 1323 ignore a missing NewSessionTicket message. 1324 [Emilia K��sper] 1325 1326 Changes between 1.0.1i and 1.0.1j [15 Oct 2014] 1327 1328 *) SRTP Memory Leak. 1329 1330 A flaw in the DTLS SRTP extension parsing code allows an attacker, who 1331 sends a carefully crafted handshake message, to cause OpenSSL to fail 1332 to free up to 64k of memory causing a memory leak. This could be 1333 exploited in a Denial Of Service attack. This issue affects OpenSSL 1334 1.0.1 server implementations for both SSL/TLS and DTLS regardless of 1335 whether SRTP is used or configured. Implementations of OpenSSL that 1336 have been compiled with OPENSSL_NO_SRTP defined are not affected. 1337 1338 The fix was developed by the OpenSSL team. 1339 (CVE-2014-3513) 1340 [OpenSSL team] 1341 1342 *) Session Ticket Memory Leak. 1343 1344 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the 1345 integrity of that ticket is first verified. In the event of a session 1346 ticket integrity check failing, OpenSSL will fail to free memory 1347 causing a memory leak. By sending a large number of invalid session 1348 tickets an attacker could exploit this issue in a Denial Of Service 1349 attack. 1350 (CVE-2014-3567) 1351 [Steve Henson] 1352 1353 *) Build option no-ssl3 is incomplete. 1354 1355 When OpenSSL is configured with "no-ssl3" as a build option, servers 1356 could accept and complete a SSL 3.0 handshake, and clients could be 1357 configured to send them. 1358 (CVE-2014-3568) 1359 [Akamai and the OpenSSL team] 1360 1361 *) Add support for TLS_FALLBACK_SCSV. 1362 Client applications doing fallback retries should call 1363 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV). 1364 (CVE-2014-3566) 1365 [Adam Langley, Bodo Moeller] 1366 1367 *) Add additional DigestInfo checks. 1368 1369 Reencode DigestInto in DER and check against the original when 1370 verifying RSA signature: this will reject any improperly encoded 1371 DigestInfo structures. 1372 1373 Note: this is a precautionary measure and no attacks are currently known. 1374 1375 [Steve Henson] 1376 1377 Changes between 1.0.1h and 1.0.1i [6 Aug 2014] 1378 1379 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the 1380 SRP code can be overrun an internal buffer. Add sanity check that 1381 g, A, B < N to SRP code. 1382 1383 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC 1384 Group for discovering this issue. 1385 (CVE-2014-3512) 1386 [Steve Henson] 1387 1388 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate 1389 TLS 1.0 instead of higher protocol versions when the ClientHello message 1390 is badly fragmented. This allows a man-in-the-middle attacker to force a 1391 downgrade to TLS 1.0 even if both the server and the client support a 1392 higher protocol version, by modifying the client's TLS records. 1393 1394 Thanks to David Benjamin and Adam Langley (Google) for discovering and 1395 researching this issue. 1396 (CVE-2014-3511) 1397 [David Benjamin] 1398 1399 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject 1400 to a denial of service attack. A malicious server can crash the client 1401 with a null pointer dereference (read) by specifying an anonymous (EC)DH 1402 ciphersuite and sending carefully crafted handshake messages. 1403 1404 Thanks to Felix Gr��bert (Google) for discovering and researching this 1405 issue. 1406 (CVE-2014-3510) 1407 [Emilia K��sper] 1408 1409 *) By sending carefully crafted DTLS packets an attacker could cause openssl 1410 to leak memory. This can be exploited through a Denial of Service attack. 1411 Thanks to Adam Langley for discovering and researching this issue. 1412 (CVE-2014-3507) 1413 [Adam Langley] 1414 1415 *) An attacker can force openssl to consume large amounts of memory whilst 1416 processing DTLS handshake messages. This can be exploited through a 1417 Denial of Service attack. 1418 Thanks to Adam Langley for discovering and researching this issue. 1419 (CVE-2014-3506) 1420 [Adam Langley] 1421 1422 *) An attacker can force an error condition which causes openssl to crash 1423 whilst processing DTLS packets due to memory being freed twice. This 1424 can be exploited through a Denial of Service attack. 1425 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching 1426 this issue. 1427 (CVE-2014-3505) 1428 [Adam Langley] 1429 1430 *) If a multithreaded client connects to a malicious server using a resumed 1431 session and the server sends an ec point format extension it could write 1432 up to 255 bytes to freed memory. 1433 1434 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this 1435 issue. 1436 (CVE-2014-3509) 1437 [Gabor Tyukasz] 1438 1439 *) A malicious server can crash an OpenSSL client with a null pointer 1440 dereference (read) by specifying an SRP ciphersuite even though it was not 1441 properly negotiated with the client. This can be exploited through a 1442 Denial of Service attack. 1443 1444 Thanks to Joonas Kuorilehto and Riku Hietam��ki (Codenomicon) for 1445 discovering and researching this issue. 1446 (CVE-2014-5139) 1447 [Steve Henson] 1448 1449 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as 1450 X509_name_oneline, X509_name_print_ex et al. to leak some information 1451 from the stack. Applications may be affected if they echo pretty printing 1452 output to the attacker. 1453 1454 Thanks to Ivan Fratric (Google) for discovering this issue. 1455 (CVE-2014-3508) 1456 [Emilia K��sper, and Steve Henson] 1457 1458 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) 1459 for corner cases. (Certain input points at infinity could lead to 1460 bogus results, with non-infinity inputs mapped to infinity too.) 1461 [Bodo Moeller] 1462 1463 Changes between 1.0.1g and 1.0.1h [5 Jun 2014] 1464 1465 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted 1466 handshake can force the use of weak keying material in OpenSSL 1467 SSL/TLS clients and servers. 1468 1469 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and 1470 researching this issue. (CVE-2014-0224) 1471 [KIKUCHI Masashi, Steve Henson] 1472 1473 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an 1474 OpenSSL DTLS client the code can be made to recurse eventually crashing 1475 in a DoS attack. 1476 1477 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue. 1478 (CVE-2014-0221) 1479 [Imre Rad, Steve Henson] 1480 1481 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can 1482 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS 1483 client or server. This is potentially exploitable to run arbitrary 1484 code on a vulnerable client or server. 1485 1486 Thanks to J��ri Aedla for reporting this issue. (CVE-2014-0195) 1487 [J��ri Aedla, Steve Henson] 1488 1489 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites 1490 are subject to a denial of service attack. 1491 1492 Thanks to Felix Gr��bert and Ivan Fratric at Google for discovering 1493 this issue. (CVE-2014-3470) 1494 [Felix Gr��bert, Ivan Fratric, Steve Henson] 1495 1496 *) Harmonize version and its documentation. -f flag is used to display 1497 compilation flags. 1498 [mancha <mancha1@zoho.com>] 1499 1500 *) Fix eckey_priv_encode so it immediately returns an error upon a failure 1501 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue. 1502 [mancha <mancha1@zoho.com>] 1503 1504 *) Fix some double frees. These are not thought to be exploitable. 1505 [mancha <mancha1@zoho.com>] 1506 1507 Changes between 1.0.1f and 1.0.1g [7 Apr 2014] 1508 1509 *) A missing bounds check in the handling of the TLS heartbeat extension 1510 can be used to reveal up to 64k of memory to a connected client or 1511 server. 1512 1513 Thanks for Neel Mehta of Google Security for discovering this bug and to 1514 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for 1515 preparing the fix (CVE-2014-0160) 1516 [Adam Langley, Bodo Moeller] 1517 1518 *) Fix for the attack described in the paper "Recovering OpenSSL 1519 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack" 1520 by Yuval Yarom and Naomi Benger. Details can be obtained from: 1521 http://eprint.iacr.org/2014/140 1522 1523 Thanks to Yuval Yarom and Naomi Benger for discovering this 1524 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076) 1525 [Yuval Yarom and Naomi Benger] 1526 1527 *) TLS pad extension: draft-agl-tls-padding-03 1528 1529 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the 1530 TLS client Hello record length value would otherwise be > 255 and 1531 less that 512 pad with a dummy extension containing zeroes so it 1532 is at least 512 bytes long. 1533 1534 [Adam Langley, Steve Henson] 1535 1536 Changes between 1.0.1e and 1.0.1f [6 Jan 2014] 1537 1538 *) Fix for TLS record tampering bug. A carefully crafted invalid 1539 handshake could crash OpenSSL with a NULL pointer exception. 1540 Thanks to Anton Johansson for reporting this issues. 1541 (CVE-2013-4353) 1542 1543 *) Keep original DTLS digest and encryption contexts in retransmission 1544 structures so we can use the previous session parameters if they need 1545 to be resent. (CVE-2013-6450) 1546 [Steve Henson] 1547 1548 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which 1549 avoids preferring ECDHE-ECDSA ciphers when the client appears to be 1550 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for 1551 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug 1552 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing 1553 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer. 1554 [Rob Stradling, Adam Langley] 1555 1556 Changes between 1.0.1d and 1.0.1e [11 Feb 2013] 1557 1558 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI 1559 supporting platforms or when small records were transferred. 1560 [Andy Polyakov, Steve Henson] 1561 1562 Changes between 1.0.1c and 1.0.1d [5 Feb 2013] 1563 1564 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 1565 1566 This addresses the flaw in CBC record processing discovered by 1567 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 1568 at: http://www.isg.rhul.ac.uk/tls/ 1569 1570 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 1571 Security Group at Royal Holloway, University of London 1572 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 1573 Emilia K��sper for the initial patch. 1574 (CVE-2013-0169) 1575 [Emilia K��sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 1576 1577 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode 1578 ciphersuites which can be exploited in a denial of service attack. 1579 Thanks go to and to Adam Langley <agl@chromium.org> for discovering 1580 and detecting this bug and to Wolfgang Ettlinger 1581 <wolfgang.ettlinger@gmail.com> for independently discovering this issue. 1582 (CVE-2012-2686) 1583 [Adam Langley] 1584 1585 *) Return an error when checking OCSP signatures when key is NULL. 1586 This fixes a DoS attack. (CVE-2013-0166) 1587 [Steve Henson] 1588 1589 *) Make openssl verify return errors. 1590 [Chris Palmer <palmer@google.com> and Ben Laurie] 1591 1592 *) Call OCSP Stapling callback after ciphersuite has been chosen, so 1593 the right response is stapled. Also change SSL_get_certificate() 1594 so it returns the certificate actually sent. 1595 See http://rt.openssl.org/Ticket/Display.html?id=2836. 1596 [Rob Stradling <rob.stradling@comodo.com>] 1597 1598 *) Fix possible deadlock when decoding public keys. 1599 [Steve Henson] 1600 1601 *) Don't use TLS 1.0 record version number in initial client hello 1602 if renegotiating. 1603 [Steve Henson] 1604 1605 Changes between 1.0.1b and 1.0.1c [10 May 2012] 1606 1607 *) Sanity check record length before skipping explicit IV in TLS 1608 1.2, 1.1 and DTLS to fix DoS attack. 1609 1610 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 1611 fuzzing as a service testing platform. 1612 (CVE-2012-2333) 1613 [Steve Henson] 1614 1615 *) Initialise tkeylen properly when encrypting CMS messages. 1616 Thanks to Solar Designer of Openwall for reporting this issue. 1617 [Steve Henson] 1618 1619 *) In FIPS mode don't try to use composite ciphers as they are not 1620 approved. 1621 [Steve Henson] 1622 1623 Changes between 1.0.1a and 1.0.1b [26 Apr 2012] 1624 1625 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and 1626 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately 1627 mean any application compiled against OpenSSL 1.0.0 headers setting 1628 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng 1629 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to 1630 0x10000000L Any application which was previously compiled against 1631 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1 1632 will need to be recompiled as a result. Letting be results in 1633 inability to disable specifically TLS 1.1 and in client context, 1634 in unlike event, limit maximum offered version to TLS 1.0 [see below]. 1635 [Steve Henson] 1636 1637 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not 1638 disable just protocol X, but all protocols above X *if* there are 1639 protocols *below* X still enabled. In more practical terms it means 1640 that if application wants to disable TLS1.0 in favor of TLS1.1 and 1641 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass 1642 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to 1643 client side. 1644 [Andy Polyakov] 1645 1646 Changes between 1.0.1 and 1.0.1a [19 Apr 2012] 1647 1648 *) Check for potentially exploitable overflows in asn1_d2i_read_bio 1649 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 1650 in CRYPTO_realloc_clean. 1651 1652 Thanks to Tavis Ormandy, Google Security Team, for discovering this 1653 issue and to Adam Langley <agl@chromium.org> for fixing it. 1654 (CVE-2012-2110) 1655 [Adam Langley (Google), Tavis Ormandy, Google Security Team] 1656 1657 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections. 1658 [Adam Langley] 1659 1660 *) Workarounds for some broken servers that "hang" if a client hello 1661 record length exceeds 255 bytes. 1662 1663 1. Do not use record version number > TLS 1.0 in initial client 1664 hello: some (but not all) hanging servers will now work. 1665 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate 1666 the number of ciphers sent in the client hello. This should be 1667 set to an even number, such as 50, for example by passing: 1668 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure. 1669 Most broken servers should now work. 1670 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable 1671 TLS 1.2 client support entirely. 1672 [Steve Henson] 1673 1674 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH. 1675 [Andy Polyakov] 1676 1677 Changes between 1.0.0h and 1.0.1 [14 Mar 2012] 1678 1679 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET 1680 STRING form instead of a DigestInfo. 1681 [Steve Henson] 1682 1683 *) The format used for MDC2 RSA signatures is inconsistent between EVP 1684 and the RSA_sign/RSA_verify functions. This was made more apparent when 1685 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular 1686 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect 1687 the correct format in RSA_verify so both forms transparently work. 1688 [Steve Henson] 1689 1690 *) Some servers which support TLS 1.0 can choke if we initially indicate 1691 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA 1692 encrypted premaster secret. As a workaround use the maximum pemitted 1693 client version in client hello, this should keep such servers happy 1694 and still work with previous versions of OpenSSL. 1695 [Steve Henson] 1696 1697 *) Add support for TLS/DTLS heartbeats. 1698 [Robin Seggelmann <seggelmann@fh-muenster.de>] 1699 1700 *) Add support for SCTP. 1701 [Robin Seggelmann <seggelmann@fh-muenster.de>] 1702 1703 *) Improved PRNG seeding for VOS. 1704 [Paul Green <Paul.Green@stratus.com>] 1705 1706 *) Extensive assembler packs updates, most notably: 1707 1708 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support; 1709 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES); 1710 - x86_64: bit-sliced AES implementation; 1711 - ARM: NEON support, contemporary platforms optimizations; 1712 - s390x: z196 support; 1713 - *: GHASH and GF(2^m) multiplication implementations; 1714 1715 [Andy Polyakov] 1716 1717 *) Make TLS-SRP code conformant with RFC 5054 API cleanup 1718 (removal of unnecessary code) 1719 [Peter Sylvester <peter.sylvester@edelweb.fr>] 1720 1721 *) Add TLS key material exporter from RFC 5705. 1722 [Eric Rescorla] 1723 1724 *) Add DTLS-SRTP negotiation from RFC 5764. 1725 [Eric Rescorla] 1726 1727 *) Add Next Protocol Negotiation, 1728 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be 1729 disabled with a no-npn flag to config or Configure. Code donated 1730 by Google. 1731 [Adam Langley <agl@google.com> and Ben Laurie] 1732 1733 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224, 1734 NIST-P256, NIST-P521, with constant-time single point multiplication on 1735 typical inputs. Compiler support for the nonstandard type __uint128_t is 1736 required to use this (present in gcc 4.4 and later, for 64-bit builds). 1737 Code made available under Apache License version 2.0. 1738 1739 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command 1740 line to include this in your build of OpenSSL, and run "make depend" (or 1741 "make update"). This enables the following EC_METHODs: 1742 1743 EC_GFp_nistp224_method() 1744 EC_GFp_nistp256_method() 1745 EC_GFp_nistp521_method() 1746 1747 EC_GROUP_new_by_curve_name() will automatically use these (while 1748 EC_GROUP_new_curve_GFp() currently prefers the more flexible 1749 implementations). 1750 [Emilia K��sper, Adam Langley, Bodo Moeller (Google)] 1751 1752 *) Use type ossl_ssize_t instad of ssize_t which isn't available on 1753 all platforms. Move ssize_t definition from e_os.h to the public 1754 header file e_os2.h as it now appears in public header file cms.h 1755 [Steve Henson] 1756 1757 *) New -sigopt option to the ca, req and x509 utilities. Additional 1758 signature parameters can be passed using this option and in 1759 particular PSS. 1760 [Steve Henson] 1761 1762 *) Add RSA PSS signing function. This will generate and set the 1763 appropriate AlgorithmIdentifiers for PSS based on those in the 1764 corresponding EVP_MD_CTX structure. No application support yet. 1765 [Steve Henson] 1766 1767 *) Support for companion algorithm specific ASN1 signing routines. 1768 New function ASN1_item_sign_ctx() signs a pre-initialised 1769 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on 1770 the appropriate parameters. 1771 [Steve Henson] 1772 1773 *) Add new algorithm specific ASN1 verification initialisation function 1774 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1 1775 handling will be the same no matter what EVP_PKEY_METHOD is used. 1776 Add a PSS handler to support verification of PSS signatures: checked 1777 against a number of sample certificates. 1778 [Steve Henson] 1779 1780 *) Add signature printing for PSS. Add PSS OIDs. 1781 [Steve Henson, Martin Kaiser <lists@kaiser.cx>] 1782 1783 *) Add algorithm specific signature printing. An individual ASN1 method 1784 can now print out signatures instead of the standard hex dump. 1785 1786 More complex signatures (e.g. PSS) can print out more meaningful 1787 information. Include DSA version that prints out the signature 1788 parameters r, s. 1789 [Steve Henson] 1790 1791 *) Password based recipient info support for CMS library: implementing 1792 RFC3211. 1793 [Steve Henson] 1794 1795 *) Split password based encryption into PBES2 and PBKDF2 functions. This 1796 neatly separates the code into cipher and PBE sections and is required 1797 for some algorithms that split PBES2 into separate pieces (such as 1798 password based CMS). 1799 [Steve Henson] 1800 1801 *) Session-handling fixes: 1802 - Fix handling of connections that are resuming with a session ID, 1803 but also support Session Tickets. 1804 - Fix a bug that suppressed issuing of a new ticket if the client 1805 presented a ticket with an expired session. 1806 - Try to set the ticket lifetime hint to something reasonable. 1807 - Make tickets shorter by excluding irrelevant information. 1808 - On the client side, don't ignore renewed tickets. 1809 [Adam Langley, Bodo Moeller (Google)] 1810 1811 *) Fix PSK session representation. 1812 [Bodo Moeller] 1813 1814 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations. 1815 1816 This work was sponsored by Intel. 1817 [Andy Polyakov] 1818 1819 *) Add GCM support to TLS library. Some custom code is needed to split 1820 the IV between the fixed (from PRF) and explicit (from TLS record) 1821 portions. This adds all GCM ciphersuites supported by RFC5288 and 1822 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and 1823 add a special AESGCM string for GCM only. 1824 [Steve Henson] 1825 1826 *) Expand range of ctrls for AES GCM. Permit setting invocation 1827 field on decrypt and retrieval of invocation field only on encrypt. 1828 [Steve Henson] 1829 1830 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support. 1831 As required by RFC5289 these ciphersuites cannot be used if for 1832 versions of TLS earlier than 1.2. 1833 [Steve Henson] 1834 1835 *) For FIPS capable OpenSSL interpret a NULL default public key method 1836 as unset and return the appopriate default but do *not* set the default. 1837 This means we can return the appopriate method in applications that 1838 swicth between FIPS and non-FIPS modes. 1839 [Steve Henson] 1840 1841 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an 1842 ENGINE is used then we cannot handle that in the FIPS module so we 1843 keep original code iff non-FIPS operations are allowed. 1844 [Steve Henson] 1845 1846 *) Add -attime option to openssl utilities. 1847 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson] 1848 1849 *) Redirect DSA and DH operations to FIPS module in FIPS mode. 1850 [Steve Henson] 1851 1852 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use 1853 FIPS EC methods unconditionally for now. 1854 [Steve Henson] 1855 1856 *) New build option no-ec2m to disable characteristic 2 code. 1857 [Steve Henson] 1858 1859 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not 1860 all cases can be covered as some introduce binary incompatibilities. 1861 [Steve Henson] 1862 1863 *) Redirect RSA operations to FIPS module including keygen, 1864 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods. 1865 [Steve Henson] 1866 1867 *) Add similar low level API blocking to ciphers. 1868 [Steve Henson] 1869 1870 *) Low level digest APIs are not approved in FIPS mode: any attempt 1871 to use these will cause a fatal error. Applications that *really* want 1872 to use them can use the private_* version instead. 1873 [Steve Henson] 1874 1875 *) Redirect cipher operations to FIPS module for FIPS builds. 1876 [Steve Henson] 1877 1878 *) Redirect digest operations to FIPS module for FIPS builds. 1879 [Steve Henson] 1880 1881 *) Update build system to add "fips" flag which will link in fipscanister.o 1882 for static and shared library builds embedding a signature if needed. 1883 [Steve Henson] 1884 1885 *) Output TLS supported curves in preference order instead of numerical 1886 order. This is currently hardcoded for the highest order curves first. 1887 This should be configurable so applications can judge speed vs strength. 1888 [Steve Henson] 1889 1890 *) Add TLS v1.2 server support for client authentication. 1891 [Steve Henson] 1892 1893 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers 1894 and enable MD5. 1895 [Steve Henson] 1896 1897 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying 1898 FIPS modules versions. 1899 [Steve Henson] 1900 1901 *) Add TLS v1.2 client side support for client authentication. Keep cache 1902 of handshake records longer as we don't know the hash algorithm to use 1903 until after the certificate request message is received. 1904 [Steve Henson] 1905 1906 *) Initial TLS v1.2 client support. Add a default signature algorithms 1907 extension including all the algorithms we support. Parse new signature 1908 format in client key exchange. Relax some ECC signing restrictions for 1909 TLS v1.2 as indicated in RFC5246. 1910 [Steve Henson] 1911 1912 *) Add server support for TLS v1.2 signature algorithms extension. Switch 1913 to new signature format when needed using client digest preference. 1914 All server ciphersuites should now work correctly in TLS v1.2. No client 1915 support yet and no support for client certificates. 1916 [Steve Henson] 1917 1918 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch 1919 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based 1920 ciphersuites. At present only RSA key exchange ciphersuites work with 1921 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete 1922 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods 1923 and version checking. 1924 [Steve Henson] 1925 1926 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled 1927 with this defined it will not be affected by any changes to ssl internal 1928 structures. Add several utility functions to allow openssl application 1929 to work with OPENSSL_NO_SSL_INTERN defined. 1930 [Steve Henson] 1931 1932 *) Add SRP support. 1933 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie] 1934 1935 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id. 1936 [Steve Henson] 1937 1938 *) Permit abbreviated handshakes when renegotiating using the function 1939 SSL_renegotiate_abbreviated(). 1940 [Robin Seggelmann <seggelmann@fh-muenster.de>] 1941 1942 *) Add call to ENGINE_register_all_complete() to 1943 ENGINE_load_builtin_engines(), so some implementations get used 1944 automatically instead of needing explicit application support. 1945 [Steve Henson] 1946 1947 *) Add support for TLS key exporter as described in RFC5705. 1948 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson] 1949 1950 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only 1951 a few changes are required: 1952 1953 Add SSL_OP_NO_TLSv1_1 flag. 1954 Add TLSv1_1 methods. 1955 Update version checking logic to handle version 1.1. 1956 Add explicit IV handling (ported from DTLS code). 1957 Add command line options to s_client/s_server. 1958 [Steve Henson] 1959 1960 Changes between 1.0.0g and 1.0.0h [12 Mar 2012] 1961 1962 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 1963 in CMS and PKCS7 code. When RSA decryption fails use a random key for 1964 content decryption and always return the same error. Note: this attack 1965 needs on average 2^20 messages so it only affects automated senders. The 1966 old behaviour can be reenabled in the CMS code by setting the 1967 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 1968 an MMA defence is not necessary. 1969 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 1970 this issue. (CVE-2012-0884) 1971 [Steve Henson] 1972 1973 *) Fix CVE-2011-4619: make sure we really are receiving a 1974 client hello before rejecting multiple SGC restarts. Thanks to 1975 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 1976 [Steve Henson] 1977 1978 Changes between 1.0.0f and 1.0.0g [18 Jan 2012] 1979 1980 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 1981 Thanks to Antonio Martin, Enterprise Secure Access Research and 1982 Development, Cisco Systems, Inc. for discovering this bug and 1983 preparing a fix. (CVE-2012-0050) 1984 [Antonio Martin] 1985 1986 Changes between 1.0.0e and 1.0.0f [4 Jan 2012] 1987 1988 *) Nadhem Alfardan and Kenny Paterson have discovered an extension 1989 of the Vaudenay padding oracle attack on CBC mode encryption 1990 which enables an efficient plaintext recovery attack against 1991 the OpenSSL implementation of DTLS. Their attack exploits timing 1992 differences arising during decryption processing. A research 1993 paper describing this attack can be found at: 1994 http://www.isg.rhul.ac.uk/~kp/dtls.pdf 1995 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 1996 Security Group at Royal Holloway, University of London 1997 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 1998 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 1999 for preparing the fix. (CVE-2011-4108) 2000 [Robin Seggelmann, Michael Tuexen] 2001 2002 *) Clear bytes used for block padding of SSL 3.0 records. 2003 (CVE-2011-4576) 2004 [Adam Langley (Google)] 2005 2006 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 2007 Kadianakis <desnacked@gmail.com> for discovering this issue and 2008 Adam Langley for preparing the fix. (CVE-2011-4619) 2009 [Adam Langley (Google)] 2010 2011 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027) 2012 [Andrey Kulikov <amdeich@gmail.com>] 2013 2014 *) Prevent malformed RFC3779 data triggering an assertion failure. 2015 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 2016 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 2017 [Rob Austein <sra@hactrn.net>] 2018 2019 *) Improved PRNG seeding for VOS. 2020 [Paul Green <Paul.Green@stratus.com>] 2021 2022 *) Fix ssl_ciph.c set-up race. 2023 [Adam Langley (Google)] 2024 2025 *) Fix spurious failures in ecdsatest.c. 2026 [Emilia K��sper (Google)] 2027 2028 *) Fix the BIO_f_buffer() implementation (which was mixing different 2029 interpretations of the '..._len' fields). 2030 [Adam Langley (Google)] 2031 2032 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 2033 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 2034 threads won't reuse the same blinding coefficients. 2035 2036 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 2037 lock to call BN_BLINDING_invert_ex, and avoids one use of 2038 BN_BLINDING_update for each BN_BLINDING structure (previously, 2039 the last update always remained unused). 2040 [Emilia K��sper (Google)] 2041 2042 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf. 2043 [Bob Buckholz (Google)] 2044 2045 Changes between 1.0.0d and 1.0.0e [6 Sep 2011] 2046 2047 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted 2048 by initialising X509_STORE_CTX properly. (CVE-2011-3207) 2049 [Kaspar Brand <ossl@velox.ch>] 2050 2051 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 2052 for multi-threaded use of ECDH. (CVE-2011-3210) 2053 [Adam Langley (Google)] 2054 2055 *) Fix x509_name_ex_d2i memory leak on bad inputs. 2056 [Bodo Moeller] 2057 2058 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check 2059 signature public key algorithm by using OID xref utilities instead. 2060 Before this you could only use some ECC ciphersuites with SHA1 only. 2061 [Steve Henson] 2062 2063 *) Add protection against ECDSA timing attacks as mentioned in the paper 2064 by Billy Bob Brumley and Nicola Tuveri, see: 2065 2066 http://eprint.iacr.org/2011/232.pdf 2067 2068 [Billy Bob Brumley and Nicola Tuveri] 2069 2070 Changes between 1.0.0c and 1.0.0d [8 Feb 2011] 2071 2072 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 2073 [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 2074 2075 *) Fix bug in string printing code: if *any* escaping is enabled we must 2076 escape the escape character (backslash) or the resulting string is 2077 ambiguous. 2078 [Steve Henson] 2079 2080 Changes between 1.0.0b and 1.0.0c [2 Dec 2010] 2081 2082 *) Disable code workaround for ancient and obsolete Netscape browsers 2083 and servers: an attacker can use it in a ciphersuite downgrade attack. 2084 Thanks to Martin Rex for discovering this bug. CVE-2010-4180 2085 [Steve Henson] 2086 2087 *) Fixed J-PAKE implementation error, originally discovered by 2088 Sebastien Martini, further info and confirmation from Stefan 2089 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 2090 [Ben Laurie] 2091 2092 Changes between 1.0.0a and 1.0.0b [16 Nov 2010] 2093 2094 *) Fix extension code to avoid race conditions which can result in a buffer 2095 overrun vulnerability: resumed sessions must not be modified as they can 2096 be shared by multiple threads. CVE-2010-3864 2097 [Steve Henson] 2098 2099 *) Fix WIN32 build system to correctly link an ENGINE directory into 2100 a DLL. 2101 [Steve Henson] 2102 2103 Changes between 1.0.0 and 1.0.0a [01 Jun 2010] 2104 2105 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover 2106 (CVE-2010-1633) 2107 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>] 2108 2109 Changes between 0.9.8n and 1.0.0 [29 Mar 2010] 2110 2111 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher 2112 context. The operation can be customised via the ctrl mechanism in 2113 case ENGINEs want to include additional functionality. 2114 [Steve Henson] 2115 2116 *) Tolerate yet another broken PKCS#8 key format: private key value negative. 2117 [Steve Henson] 2118 2119 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to 2120 output hashes compatible with older versions of OpenSSL. 2121 [Willy Weisz <weisz@vcpc.univie.ac.at>] 2122 2123 *) Fix compression algorithm handling: if resuming a session use the 2124 compression algorithm of the resumed session instead of determining 2125 it from client hello again. Don't allow server to change algorithm. 2126 [Steve Henson] 2127 2128 *) Add load_crls() function to apps tidying load_certs() too. Add option 2129 to verify utility to allow additional CRLs to be included. 2130 [Steve Henson] 2131 2132 *) Update OCSP request code to permit adding custom headers to the request: 2133 some responders need this. 2134 [Steve Henson] 2135 2136 *) The function EVP_PKEY_sign() returns <=0 on error: check return code 2137 correctly. 2138 [Julia Lawall <julia@diku.dk>] 2139 2140 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it 2141 needlessly dereferenced structures, used obsolete functions and 2142 didn't handle all updated verify codes correctly. 2143 [Steve Henson] 2144 2145 *) Disable MD2 in the default configuration. 2146 [Steve Henson] 2147 2148 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to 2149 indicate the initial BIO being pushed or popped. This makes it possible 2150 to determine whether the BIO is the one explicitly called or as a result 2151 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so 2152 it handles reference counts correctly and doesn't zero out the I/O bio 2153 when it is not being explicitly popped. WARNING: applications which 2154 included workarounds for the old buggy behaviour will need to be modified 2155 or they could free up already freed BIOs. 2156 [Steve Henson] 2157 2158 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni 2159 renaming to all platforms (within the 0.9.8 branch, this was 2160 done conditionally on Netware platforms to avoid a name clash). 2161 [Guenter <lists@gknw.net>] 2162 2163 *) Add ECDHE and PSK support to DTLS. 2164 [Michael Tuexen <tuexen@fh-muenster.de>] 2165 2166 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't 2167 be used on C++. 2168 [Steve Henson] 2169 2170 *) Add "missing" function EVP_MD_flags() (without this the only way to 2171 retrieve a digest flags is by accessing the structure directly. Update 2172 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest 2173 or cipher is registered as in the "from" argument. Print out all 2174 registered digests in the dgst usage message instead of manually 2175 attempting to work them out. 2176 [Steve Henson] 2177 2178 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello: 2179 this allows the use of compression and extensions. Change default cipher 2180 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2 2181 by default unless an application cipher string requests it. 2182 [Steve Henson] 2183 2184 *) Alter match criteria in PKCS12_parse(). It used to try to use local 2185 key ids to find matching certificates and keys but some PKCS#12 files 2186 don't follow the (somewhat unwritten) rules and this strategy fails. 2187 Now just gather all certificates together and the first private key 2188 then look for the first certificate that matches the key. 2189 [Steve Henson] 2190 2191 *) Support use of registered digest and cipher names for dgst and cipher 2192 commands instead of having to add each one as a special case. So now 2193 you can do: 2194 2195 openssl sha256 foo 2196 2197 as well as: 2198 2199 openssl dgst -sha256 foo 2200 2201 and this works for ENGINE based algorithms too. 2202 2203 [Steve Henson] 2204 2205 *) Update Gost ENGINE to support parameter files. 2206 [Victor B. Wagner <vitus@cryptocom.ru>] 2207 2208 *) Support GeneralizedTime in ca utility. 2209 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson] 2210 2211 *) Enhance the hash format used for certificate directory links. The new 2212 form uses the canonical encoding (meaning equivalent names will work 2213 even if they aren't identical) and uses SHA1 instead of MD5. This form 2214 is incompatible with the older format and as a result c_rehash should 2215 be used to rebuild symbolic links. 2216 [Steve Henson] 2217 2218 *) Make PKCS#8 the default write format for private keys, replacing the 2219 traditional format. This form is standardised, more secure and doesn't 2220 include an implicit MD5 dependency. 2221 [Steve Henson] 2222 2223 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code 2224 committed to OpenSSL should pass this lot as a minimum. 2225 [Steve Henson] 2226 2227 *) Add session ticket override functionality for use by EAP-FAST. 2228 [Jouni Malinen <j@w1.fi>] 2229 2230 *) Modify HMAC functions to return a value. Since these can be implemented 2231 in an ENGINE errors can occur. 2232 [Steve Henson] 2233 2234 *) Type-checked OBJ_bsearch_ex. 2235 [Ben Laurie] 2236 2237 *) Type-checked OBJ_bsearch. Also some constification necessitated 2238 by type-checking. Still to come: TXT_DB, bsearch(?), 2239 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING, 2240 CONF_VALUE. 2241 [Ben Laurie] 2242 2243 *) New function OPENSSL_gmtime_adj() to add a specific number of days and 2244 seconds to a tm structure directly, instead of going through OS 2245 specific date routines. This avoids any issues with OS routines such 2246 as the year 2038 bug. New *_adj() functions for ASN1 time structures 2247 and X509_time_adj_ex() to cover the extended range. The existing 2248 X509_time_adj() is still usable and will no longer have any date issues. 2249 [Steve Henson] 2250 2251 *) Delta CRL support. New use deltas option which will attempt to locate 2252 and search any appropriate delta CRLs available. 2253 2254 This work was sponsored by Google. 2255 [Steve Henson] 2256 2257 *) Support for CRLs partitioned by reason code. Reorganise CRL processing 2258 code and add additional score elements. Validate alternate CRL paths 2259 as part of the CRL checking and indicate a new error "CRL path validation 2260 error" in this case. Applications wanting additional details can use 2261 the verify callback and check the new "parent" field. If this is not 2262 NULL CRL path validation is taking place. Existing applications wont 2263 see this because it requires extended CRL support which is off by 2264 default. 2265 2266 This work was sponsored by Google. 2267 [Steve Henson] 2268 2269 *) Support for freshest CRL extension. 2270 2271 This work was sponsored by Google. 2272 [Steve Henson] 2273 2274 *) Initial indirect CRL support. Currently only supported in the CRLs 2275 passed directly and not via lookup. Process certificate issuer 2276 CRL entry extension and lookup CRL entries by bother issuer name 2277 and serial number. Check and process CRL issuer entry in IDP extension. 2278 2279 This work was sponsored by Google. 2280 [Steve Henson] 2281 2282 *) Add support for distinct certificate and CRL paths. The CRL issuer 2283 certificate is validated separately in this case. Only enabled if 2284 an extended CRL support flag is set: this flag will enable additional 2285 CRL functionality in future. 2286 2287 This work was sponsored by Google. 2288 [Steve Henson] 2289 2290 *) Add support for policy mappings extension. 2291 2292 This work was sponsored by Google. 2293 [Steve Henson] 2294 2295 *) Fixes to pathlength constraint, self issued certificate handling, 2296 policy processing to align with RFC3280 and PKITS tests. 2297 2298 This work was sponsored by Google. 2299 [Steve Henson] 2300 2301 *) Support for name constraints certificate extension. DN, email, DNS 2302 and URI types are currently supported. 2303 2304 This work was sponsored by Google. 2305 [Steve Henson] 2306 2307 *) To cater for systems that provide a pointer-based thread ID rather 2308 than numeric, deprecate the current numeric thread ID mechanism and 2309 replace it with a structure and associated callback type. This 2310 mechanism allows a numeric "hash" to be extracted from a thread ID in 2311 either case, and on platforms where pointers are larger than 'long', 2312 mixing is done to help ensure the numeric 'hash' is usable even if it 2313 can't be guaranteed unique. The default mechanism is to use "&errno" 2314 as a pointer-based thread ID to distinguish between threads. 2315 2316 Applications that want to provide their own thread IDs should now use 2317 CRYPTO_THREADID_set_callback() to register a callback that will call 2318 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer(). 2319 2320 Note that ERR_remove_state() is now deprecated, because it is tied 2321 to the assumption that thread IDs are numeric. ERR_remove_state(0) 2322 to free the current thread's error state should be replaced by 2323 ERR_remove_thread_state(NULL). 2324 2325 (This new approach replaces the functions CRYPTO_set_idptr_callback(), 2326 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in 2327 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an 2328 application was previously providing a numeric thread callback that 2329 was inappropriate for distinguishing threads, then uniqueness might 2330 have been obtained with &errno that happened immediately in the 2331 intermediate development versions of OpenSSL; this is no longer the 2332 case, the numeric thread callback will now override the automatic use 2333 of &errno.) 2334 [Geoff Thorpe, with help from Bodo Moeller] 2335 2336 *) Initial support for different CRL issuing certificates. This covers a 2337 simple case where the self issued certificates in the chain exist and 2338 the real CRL issuer is higher in the existing chain. 2339 2340 This work was sponsored by Google. 2341 [Steve Henson] 2342 2343 *) Removed effectively defunct crypto/store from the build. 2344 [Ben Laurie] 2345 2346 *) Revamp of STACK to provide stronger type-checking. Still to come: 2347 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE, 2348 ASN1_STRING, CONF_VALUE. 2349 [Ben Laurie] 2350 2351 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer 2352 RAM on SSL connections. This option can save about 34k per idle SSL. 2353 [Nick Mathewson] 2354 2355 *) Revamp of LHASH to provide stronger type-checking. Still to come: 2356 STACK, TXT_DB, bsearch, qsort. 2357 [Ben Laurie] 2358 2359 *) Initial support for Cryptographic Message Syntax (aka CMS) based 2360 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility, 2361 support for data, signedData, compressedData, digestedData and 2362 encryptedData, envelopedData types included. Scripts to check against 2363 RFC4134 examples draft and interop and consistency checks of many 2364 content types and variants. 2365 [Steve Henson] 2366 2367 *) Add options to enc utility to support use of zlib compression BIO. 2368 [Steve Henson] 2369 2370 *) Extend mk1mf to support importing of options and assembly language 2371 files from Configure script, currently only included in VC-WIN32. 2372 The assembly language rules can now optionally generate the source 2373 files from the associated perl scripts. 2374 [Steve Henson] 2375 2376 *) Implement remaining functionality needed to support GOST ciphersuites. 2377 Interop testing has been performed using CryptoPro implementations. 2378 [Victor B. Wagner <vitus@cryptocom.ru>] 2379 2380 *) s390x assembler pack. 2381 [Andy Polyakov] 2382 2383 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU 2384 "family." 2385 [Andy Polyakov] 2386 2387 *) Implement Opaque PRF Input TLS extension as specified in 2388 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an 2389 official specification yet and no extension type assignment by 2390 IANA exists, this extension (for now) will have to be explicitly 2391 enabled when building OpenSSL by providing the extension number 2392 to use. For example, specify an option 2393 2394 -DTLSEXT_TYPE_opaque_prf_input=0x9527 2395 2396 to the "config" or "Configure" script to enable the extension, 2397 assuming extension number 0x9527 (which is a completely arbitrary 2398 and unofficial assignment based on the MD5 hash of the Internet 2399 Draft). Note that by doing so, you potentially lose 2400 interoperability with other TLS implementations since these might 2401 be using the same extension number for other purposes. 2402 2403 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the 2404 opaque PRF input value to use in the handshake. This will create 2405 an interal copy of the length-'len' string at 'src', and will 2406 return non-zero for success. 2407 2408 To get more control and flexibility, provide a callback function 2409 by using 2410 2411 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb) 2412 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg) 2413 2414 where 2415 2416 int (*cb)(SSL *, void *peerinput, size_t len, void *arg); 2417 void *arg; 2418 2419 Callback function 'cb' will be called in handshakes, and is 2420 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate. 2421 Argument 'arg' is for application purposes (the value as given to 2422 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly 2423 be provided to the callback function). The callback function 2424 has to return non-zero to report success: usually 1 to use opaque 2425 PRF input just if possible, or 2 to enforce use of the opaque PRF 2426 input. In the latter case, the library will abort the handshake 2427 if opaque PRF input is not successfully negotiated. 2428 2429 Arguments 'peerinput' and 'len' given to the callback function 2430 will always be NULL and 0 in the case of a client. A server will 2431 see the client's opaque PRF input through these variables if 2432 available (NULL and 0 otherwise). Note that if the server 2433 provides an opaque PRF input, the length must be the same as the 2434 length of the client's opaque PRF input. 2435 2436 Note that the callback function will only be called when creating 2437 a new session (session resumption can resume whatever was 2438 previously negotiated), and will not be called in SSL 2.0 2439 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or 2440 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended 2441 for applications that need to enforce opaque PRF input. 2442 2443 [Bodo Moeller] 2444 2445 *) Update ssl code to support digests other than SHA1+MD5 for handshake 2446 MAC. 2447 2448 [Victor B. Wagner <vitus@cryptocom.ru>] 2449 2450 *) Add RFC4507 support to OpenSSL. This includes the corrections in 2451 RFC4507bis. The encrypted ticket format is an encrypted encoded 2452 SSL_SESSION structure, that way new session features are automatically 2453 supported. 2454 2455 If a client application caches session in an SSL_SESSION structure 2456 support is transparent because tickets are now stored in the encoded 2457 SSL_SESSION. 2458 2459 The SSL_CTX structure automatically generates keys for ticket 2460 protection in servers so again support should be possible 2461 with no application modification. 2462 2463 If a client or server wishes to disable RFC4507 support then the option 2464 SSL_OP_NO_TICKET can be set. 2465 2466 Add a TLS extension debugging callback to allow the contents of any client 2467 or server extensions to be examined. 2468 2469 This work was sponsored by Google. 2470 [Steve Henson] 2471 2472 *) Final changes to avoid use of pointer pointer casts in OpenSSL. 2473 OpenSSL should now compile cleanly on gcc 4.2 2474 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson] 2475 2476 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC 2477 support including streaming MAC support: this is required for GOST 2478 ciphersuite support. 2479 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson] 2480 2481 *) Add option -stream to use PKCS#7 streaming in smime utility. New 2482 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream() 2483 to output in BER and PEM format. 2484 [Steve Henson] 2485 2486 *) Experimental support for use of HMAC via EVP_PKEY interface. This 2487 allows HMAC to be handled via the EVP_DigestSign*() interface. The 2488 EVP_PKEY "key" in this case is the HMAC key, potentially allowing 2489 ENGINE support for HMAC keys which are unextractable. New -mac and 2490 -macopt options to dgst utility. 2491 [Steve Henson] 2492 2493 *) New option -sigopt to dgst utility. Update dgst to use 2494 EVP_Digest{Sign,Verify}*. These two changes make it possible to use 2495 alternative signing paramaters such as X9.31 or PSS in the dgst 2496 utility. 2497 [Steve Henson] 2498 2499 *) Change ssl_cipher_apply_rule(), the internal function that does 2500 the work each time a ciphersuite string requests enabling 2501 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or 2502 removing ("!foo+bar") a class of ciphersuites: Now it maintains 2503 the order of disabled ciphersuites such that those ciphersuites 2504 that most recently went from enabled to disabled not only stay 2505 in order with respect to each other, but also have higher priority 2506 than other disabled ciphersuites the next time ciphersuites are 2507 enabled again. 2508 2509 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable 2510 the same ciphersuites as with "HIGH" alone, but in a specific 2511 order where the PSK ciphersuites come first (since they are the 2512 most recently disabled ciphersuites when "HIGH" is parsed). 2513 2514 Also, change ssl_create_cipher_list() (using this new 2515 funcionality) such that between otherwise identical 2516 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in 2517 the default order. 2518 [Bodo Moeller] 2519 2520 *) Change ssl_create_cipher_list() so that it automatically 2521 arranges the ciphersuites in reasonable order before starting 2522 to process the rule string. Thus, the definition for "DEFAULT" 2523 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but 2524 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH". 2525 This makes it much easier to arrive at a reasonable default order 2526 in applications for which anonymous ciphers are OK (meaning 2527 that you can't actually use DEFAULT). 2528 [Bodo Moeller; suggested by Victor Duchovni] 2529 2530 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string 2531 processing) into multiple integers instead of setting 2532 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK", 2533 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer. 2534 (These masks as well as the individual bit definitions are hidden 2535 away into the non-exported interface ssl/ssl_locl.h, so this 2536 change to the definition of the SSL_CIPHER structure shouldn't 2537 affect applications.) This give us more bits for each of these 2538 categories, so there is no longer a need to coagulate AES128 and 2539 AES256 into a single algorithm bit, and to coagulate Camellia128 2540 and Camellia256 into a single algorithm bit, which has led to all 2541 kinds of kludges. 2542 2543 Thus, among other things, the kludge introduced in 0.9.7m and 2544 0.9.8e for masking out AES256 independently of AES128 or masking 2545 out Camellia256 independently of AES256 is not needed here in 0.9.9. 2546 2547 With the change, we also introduce new ciphersuite aliases that 2548 so far were missing: "AES128", "AES256", "CAMELLIA128", and 2549 "CAMELLIA256". 2550 [Bodo Moeller] 2551 2552 *) Add support for dsa-with-SHA224 and dsa-with-SHA256. 2553 Use the leftmost N bytes of the signature input if the input is 2554 larger than the prime q (with N being the size in bytes of q). 2555 [Nils Larsch] 2556 2557 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses 2558 it yet and it is largely untested. 2559 [Steve Henson] 2560 2561 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types. 2562 [Nils Larsch] 2563 2564 *) Initial incomplete changes to avoid need for function casts in OpenSSL 2565 some compilers (gcc 4.2 and later) reject their use. Safestack is 2566 reimplemented. Update ASN1 to avoid use of legacy functions. 2567 [Steve Henson] 2568 2569 *) Win32/64 targets are linked with Winsock2. 2570 [Andy Polyakov] 2571 2572 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected 2573 to external functions. This can be used to increase CRL handling 2574 efficiency especially when CRLs are very large by (for example) storing 2575 the CRL revoked certificates in a database. 2576 [Steve Henson] 2577 2578 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so 2579 new CRLs added to a directory can be used. New command line option 2580 -verify_return_error to s_client and s_server. This causes real errors 2581 to be returned by the verify callback instead of carrying on no matter 2582 what. This reflects the way a "real world" verify callback would behave. 2583 [Steve Henson] 2584 2585 *) GOST engine, supporting several GOST algorithms and public key formats. 2586 Kindly donated by Cryptocom. 2587 [Cryptocom] 2588 2589 *) Partial support for Issuing Distribution Point CRL extension. CRLs 2590 partitioned by DP are handled but no indirect CRL or reason partitioning 2591 (yet). Complete overhaul of CRL handling: now the most suitable CRL is 2592 selected via a scoring technique which handles IDP and AKID in CRLs. 2593 [Steve Henson] 2594 2595 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which 2596 will ultimately be used for all verify operations: this will remove the 2597 X509_STORE dependency on certificate verification and allow alternative 2598 lookup methods. X509_STORE based implementations of these two callbacks. 2599 [Steve Henson] 2600 2601 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names. 2602 Modify get_crl() to find a valid (unexpired) CRL if possible. 2603 [Steve Henson] 2604 2605 *) New function X509_CRL_match() to check if two CRLs are identical. Normally 2606 this would be called X509_CRL_cmp() but that name is already used by 2607 a function that just compares CRL issuer names. Cache several CRL 2608 extensions in X509_CRL structure and cache CRLDP in X509. 2609 [Steve Henson] 2610 2611 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name) 2612 this maps equivalent X509_NAME structures into a consistent structure. 2613 Name comparison can then be performed rapidly using memcmp(). 2614 [Steve Henson] 2615 2616 *) Non-blocking OCSP request processing. Add -timeout option to ocsp 2617 utility. 2618 [Steve Henson] 2619 2620 *) Allow digests to supply their own micalg string for S/MIME type using 2621 the ctrl EVP_MD_CTRL_MICALG. 2622 [Steve Henson] 2623 2624 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the 2625 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN 2626 ctrl. It can then customise the structure before and/or after signing 2627 if necessary. 2628 [Steve Henson] 2629 2630 *) New function OBJ_add_sigid() to allow application defined signature OIDs 2631 to be added to OpenSSLs internal tables. New function OBJ_sigid_free() 2632 to free up any added signature OIDs. 2633 [Steve Henson] 2634 2635 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(), 2636 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal 2637 digest and cipher tables. New options added to openssl utility: 2638 list-message-digest-algorithms and list-cipher-algorithms. 2639 [Steve Henson] 2640 2641 *) Change the array representation of binary polynomials: the list 2642 of degrees of non-zero coefficients is now terminated with -1. 2643 Previously it was terminated with 0, which was also part of the 2644 value; thus, the array representation was not applicable to 2645 polynomials where t^0 has coefficient zero. This change makes 2646 the array representation useful in a more general context. 2647 [Douglas Stebila] 2648 2649 *) Various modifications and fixes to SSL/TLS cipher string 2650 handling. For ECC, the code now distinguishes between fixed ECDH 2651 with RSA certificates on the one hand and with ECDSA certificates 2652 on the other hand, since these are separate ciphersuites. The 2653 unused code for Fortezza ciphersuites has been removed. 2654 2655 For consistency with EDH, ephemeral ECDH is now called "EECDH" 2656 (not "ECDHE"). For consistency with the code for DH 2657 certificates, use of ECDH certificates is now considered ECDH 2658 authentication, not RSA or ECDSA authentication (the latter is 2659 merely the CA's signing algorithm and not actively used in the 2660 protocol). 2661 2662 The temporary ciphersuite alias "ECCdraft" is no longer 2663 available, and ECC ciphersuites are no longer excluded from "ALL" 2664 and "DEFAULT". The following aliases now exist for RFC 4492 2665 ciphersuites, most of these by analogy with the DH case: 2666 2667 kECDHr - ECDH cert, signed with RSA 2668 kECDHe - ECDH cert, signed with ECDSA 2669 kECDH - ECDH cert (signed with either RSA or ECDSA) 2670 kEECDH - ephemeral ECDH 2671 ECDH - ECDH cert or ephemeral ECDH 2672 2673 aECDH - ECDH cert 2674 aECDSA - ECDSA cert 2675 ECDSA - ECDSA cert 2676 2677 AECDH - anonymous ECDH 2678 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH") 2679 2680 [Bodo Moeller] 2681 2682 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported. 2683 Use correct micalg parameters depending on digest(s) in signed message. 2684 [Steve Henson] 2685 2686 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process 2687 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code. 2688 [Steve Henson] 2689 2690 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit 2691 an engine to register a method. Add ENGINE lookups for methods and 2692 functional reference processing. 2693 [Steve Henson] 2694 2695 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of 2696 EVP_{Sign,Verify}* which allow an application to customise the signature 2697 process. 2698 [Steve Henson] 2699 2700 *) New -resign option to smime utility. This adds one or more signers 2701 to an existing PKCS#7 signedData structure. Also -md option to use an 2702 alternative message digest algorithm for signing. 2703 [Steve Henson] 2704 2705 *) Tidy up PKCS#7 routines and add new functions to make it easier to 2706 create PKCS7 structures containing multiple signers. Update smime 2707 application to support multiple signers. 2708 [Steve Henson] 2709 2710 *) New -macalg option to pkcs12 utility to allow setting of an alternative 2711 digest MAC. 2712 [Steve Henson] 2713 2714 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC. 2715 Reorganize PBE internals to lookup from a static table using NIDs, 2716 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl: 2717 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative 2718 PRF which will be automatically used with PBES2. 2719 [Steve Henson] 2720 2721 *) Replace the algorithm specific calls to generate keys in "req" with the 2722 new API. 2723 [Steve Henson] 2724 2725 *) Update PKCS#7 enveloped data routines to use new API. This is now 2726 supported by any public key method supporting the encrypt operation. A 2727 ctrl is added to allow the public key algorithm to examine or modify 2728 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is 2729 a no op. 2730 [Steve Henson] 2731 2732 *) Add a ctrl to asn1 method to allow a public key algorithm to express 2733 a default digest type to use. In most cases this will be SHA1 but some 2734 algorithms (such as GOST) need to specify an alternative digest. The 2735 return value indicates how strong the prefernce is 1 means optional and 2736 2 is mandatory (that is it is the only supported type). Modify 2737 ASN1_item_sign() to accept a NULL digest argument to indicate it should 2738 use the default md. Update openssl utilities to use the default digest 2739 type for signing if it is not explicitly indicated. 2740 [Steve Henson] 2741 2742 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New 2743 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant 2744 signing method from the key type. This effectively removes the link 2745 between digests and public key types. 2746 [Steve Henson] 2747 2748 *) Add an OID cross reference table and utility functions. Its purpose is to 2749 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1, 2750 rsaEncryption. This will allow some of the algorithm specific hackery 2751 needed to use the correct OID to be removed. 2752 [Steve Henson] 2753 2754 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO 2755 structures for PKCS7_sign(). They are now set up by the relevant public 2756 key ASN1 method. 2757 [Steve Henson] 2758 2759 *) Add provisional EC pkey method with support for ECDSA and ECDH. 2760 [Steve Henson] 2761 2762 *) Add support for key derivation (agreement) in the API, DH method and 2763 pkeyutl. 2764 [Steve Henson] 2765 2766 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support 2767 public and private key formats. As a side effect these add additional 2768 command line functionality not previously available: DSA signatures can be 2769 generated and verified using pkeyutl and DH key support and generation in 2770 pkey, genpkey. 2771 [Steve Henson] 2772 2773 *) BeOS support. 2774 [Oliver Tappe <zooey@hirschkaefer.de>] 2775 2776 *) New make target "install_html_docs" installs HTML renditions of the 2777 manual pages. 2778 [Oliver Tappe <zooey@hirschkaefer.de>] 2779 2780 *) New utility "genpkey" this is analagous to "genrsa" etc except it can 2781 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to 2782 support key and parameter generation and add initial key generation 2783 functionality for RSA. 2784 [Steve Henson] 2785 2786 *) Add functions for main EVP_PKEY_method operations. The undocumented 2787 functions EVP_PKEY_{encrypt,decrypt} have been renamed to 2788 EVP_PKEY_{encrypt,decrypt}_old. 2789 [Steve Henson] 2790 2791 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public 2792 key API, doesn't do much yet. 2793 [Steve Henson] 2794 2795 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about 2796 public key algorithms. New option to openssl utility: 2797 "list-public-key-algorithms" to print out info. 2798 [Steve Henson] 2799 2800 *) Implement the Supported Elliptic Curves Extension for 2801 ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 2802 [Douglas Stebila] 2803 2804 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or 2805 EVP_CIPHER structures to avoid later problems in EVP_cleanup(). 2806 [Steve Henson] 2807 2808 *) New utilities pkey and pkeyparam. These are similar to algorithm specific 2809 utilities such as rsa, dsa, dsaparam etc except they process any key 2810 type. 2811 [Steve Henson] 2812 2813 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New 2814 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(), 2815 EVP_PKEY_print_param() to print public key data from an EVP_PKEY 2816 structure. 2817 [Steve Henson] 2818 2819 *) Initial support for pluggable public key ASN1. 2820 De-spaghettify the public key ASN1 handling. Move public and private 2821 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate 2822 algorithm specific handling to a single module within the relevant 2823 algorithm directory. Add functions to allow (near) opaque processing 2824 of public and private key structures. 2825 [Steve Henson] 2826 2827 *) Implement the Supported Point Formats Extension for 2828 ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 2829 [Douglas Stebila] 2830 2831 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members 2832 for the psk identity [hint] and the psk callback functions to the 2833 SSL_SESSION, SSL and SSL_CTX structure. 2834 2835 New ciphersuites: 2836 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA, 2837 PSK-AES256-CBC-SHA 2838 2839 New functions: 2840 SSL_CTX_use_psk_identity_hint 2841 SSL_get_psk_identity_hint 2842 SSL_get_psk_identity 2843 SSL_use_psk_identity_hint 2844 2845 [Mika Kousa and Pasi Eronen of Nokia Corporation] 2846 2847 *) Add RFC 3161 compliant time stamp request creation, response generation 2848 and response verification functionality. 2849 [Zolt��n Gl��zik <zglozik@opentsa.org>, The OpenTSA Project] 2850 2851 *) Add initial support for TLS extensions, specifically for the server_name 2852 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 2853 have new members for a host name. The SSL data structure has an 2854 additional member SSL_CTX *initial_ctx so that new sessions can be 2855 stored in that context to allow for session resumption, even after the 2856 SSL has been switched to a new SSL_CTX in reaction to a client's 2857 server_name extension. 2858 2859 New functions (subject to change): 2860 2861 SSL_get_servername() 2862 SSL_get_servername_type() 2863 SSL_set_SSL_CTX() 2864 2865 New CTRL codes and macros (subject to change): 2866 2867 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 2868 - SSL_CTX_set_tlsext_servername_callback() 2869 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 2870 - SSL_CTX_set_tlsext_servername_arg() 2871 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 2872 2873 openssl s_client has a new '-servername ...' option. 2874 2875 openssl s_server has new options '-servername_host ...', '-cert2 ...', 2876 '-key2 ...', '-servername_fatal' (subject to change). This allows 2877 testing the HostName extension for a specific single host name ('-cert' 2878 and '-key' remain fallbacks for handshakes without HostName 2879 negotiation). If the unrecogninzed_name alert has to be sent, this by 2880 default is a warning; it becomes fatal with the '-servername_fatal' 2881 option. 2882 2883 [Peter Sylvester, Remy Allais, Christophe Renou] 2884 2885 *) Whirlpool hash implementation is added. 2886 [Andy Polyakov] 2887 2888 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to 2889 bn(64,32). Because of instruction set limitations it doesn't have 2890 any negative impact on performance. This was done mostly in order 2891 to make it possible to share assembler modules, such as bn_mul_mont 2892 implementations, between 32- and 64-bit builds without hassle. 2893 [Andy Polyakov] 2894 2895 *) Move code previously exiled into file crypto/ec/ec2_smpt.c 2896 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP 2897 macro. 2898 [Bodo Moeller] 2899 2900 *) New candidate for BIGNUM assembler implementation, bn_mul_mont, 2901 dedicated Montgomery multiplication procedure, is introduced. 2902 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher 2903 "64-bit" performance on certain 32-bit targets. 2904 [Andy Polyakov] 2905 2906 *) New option SSL_OP_NO_COMP to disable use of compression selectively 2907 in SSL structures. New SSL ctrl to set maximum send fragment size. 2908 Save memory by seeting the I/O buffer sizes dynamically instead of 2909 using the maximum available value. 2910 [Steve Henson] 2911 2912 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code 2913 in addition to the text details. 2914 [Bodo Moeller] 2915 2916 *) Very, very preliminary EXPERIMENTAL support for printing of general 2917 ASN1 structures. This currently produces rather ugly output and doesn't 2918 handle several customised structures at all. 2919 [Steve Henson] 2920 2921 *) Integrated support for PVK file format and some related formats such 2922 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support 2923 these in the 'rsa' and 'dsa' utilities. 2924 [Steve Henson] 2925 2926 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line. 2927 [Steve Henson] 2928 2929 *) Remove the ancient ASN1_METHOD code. This was only ever used in one 2930 place for the (very old) "NETSCAPE" format certificates which are now 2931 handled using new ASN1 code equivalents. 2932 [Steve Henson] 2933 2934 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD 2935 pointer and make the SSL_METHOD parameter in SSL_CTX_new, 2936 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'. 2937 [Nils Larsch] 2938 2939 *) Modify CRL distribution points extension code to print out previously 2940 unsupported fields. Enhance extension setting code to allow setting of 2941 all fields. 2942 [Steve Henson] 2943 2944 *) Add print and set support for Issuing Distribution Point CRL extension. 2945 [Steve Henson] 2946 2947 *) Change 'Configure' script to enable Camellia by default. 2948 [NTT] 2949 2950 Changes between 0.9.8m and 0.9.8n [24 Mar 2010] 2951 2952 *) When rejecting SSL/TLS records due to an incorrect version number, never 2953 update s->server with a new major version number. As of 2954 - OpenSSL 0.9.8m if 'short' is a 16-bit type, 2955 - OpenSSL 0.9.8f if 'short' is longer than 16 bits, 2956 the previous behavior could result in a read attempt at NULL when 2957 receiving specific incorrect SSL/TLS records once record payload 2958 protection is active. (CVE-2010-0740) 2959 [Bodo Moeller, Adam Langley <agl@chromium.org>] 2960 2961 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL 2962 could be crashed if the relevant tables were not present (e.g. chrooted). 2963 [Tomas Hoger <thoger@redhat.com>] 2964 2965 Changes between 0.9.8l and 0.9.8m [25 Feb 2010] 2966 2967 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245) 2968 [Martin Olsson, Neel Mehta] 2969 2970 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to 2971 accommodate for stack sorting, always a write lock!). 2972 [Bodo Moeller] 2973 2974 *) On some versions of WIN32 Heap32Next is very slow. This can cause 2975 excessive delays in the RAND_poll(): over a minute. As a workaround 2976 include a time check in the inner Heap32Next loop too. 2977 [Steve Henson] 2978 2979 *) The code that handled flushing of data in SSL/TLS originally used the 2980 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused 2981 the problem outlined in PR#1949. The fix suggested there however can 2982 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions 2983 of Apache). So instead simplify the code to flush unconditionally. 2984 This should be fine since flushing with no data to flush is a no op. 2985 [Steve Henson] 2986 2987 *) Handle TLS versions 2.0 and later properly and correctly use the 2988 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way 2989 off ancient servers have a habit of sticking around for a while... 2990 [Steve Henson] 2991 2992 *) Modify compression code so it frees up structures without using the 2993 ex_data callbacks. This works around a problem where some applications 2994 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when 2995 restarting) then use compression (e.g. SSL with compression) later. 2996 This results in significant per-connection memory leaks and 2997 has caused some security issues including CVE-2008-1678 and 2998 CVE-2009-4355. 2999 [Steve Henson] 3000 3001 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't 3002 change when encrypting or decrypting. 3003 [Bodo Moeller] 3004 3005 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to 3006 connect and renegotiate with servers which do not support RI. 3007 Until RI is more widely deployed this option is enabled by default. 3008 [Steve Henson] 3009 3010 *) Add "missing" ssl ctrls to clear options and mode. 3011 [Steve Henson] 3012 3013 *) If client attempts to renegotiate and doesn't support RI respond with 3014 a no_renegotiation alert as required by RFC5746. Some renegotiating 3015 TLS clients will continue a connection gracefully when they receive 3016 the alert. Unfortunately OpenSSL mishandled this alert and would hang 3017 waiting for a server hello which it will never receive. Now we treat a 3018 received no_renegotiation alert as a fatal error. This is because 3019 applications requesting a renegotiation might well expect it to succeed 3020 and would have no code in place to handle the server denying it so the 3021 only safe thing to do is to terminate the connection. 3022 [Steve Henson] 3023 3024 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if 3025 peer supports secure renegotiation and 0 otherwise. Print out peer 3026 renegotiation support in s_client/s_server. 3027 [Steve Henson] 3028 3029 *) Replace the highly broken and deprecated SPKAC certification method with 3030 the updated NID creation version. This should correctly handle UTF8. 3031 [Steve Henson] 3032 3033 *) Implement RFC5746. Re-enable renegotiation but require the extension 3034 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION 3035 turns out to be a bad idea. It has been replaced by 3036 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with 3037 SSL_CTX_set_options(). This is really not recommended unless you 3038 know what you are doing. 3039 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson] 3040 3041 *) Fixes to stateless session resumption handling. Use initial_ctx when 3042 issuing and attempting to decrypt tickets in case it has changed during 3043 servername handling. Use a non-zero length session ID when attempting 3044 stateless session resumption: this makes it possible to determine if 3045 a resumption has occurred immediately after receiving server hello 3046 (several places in OpenSSL subtly assume this) instead of later in 3047 the handshake. 3048 [Steve Henson] 3049 3050 *) The functions ENGINE_ctrl(), OPENSSL_isservice(), 3051 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error 3052 fixes for a few places where the return code is not checked 3053 correctly. 3054 [Julia Lawall <julia@diku.dk>] 3055 3056 *) Add --strict-warnings option to Configure script to include devteam 3057 warnings in other configurations. 3058 [Steve Henson] 3059 3060 *) Add support for --libdir option and LIBDIR variable in makefiles. This 3061 makes it possible to install openssl libraries in locations which 3062 have names other than "lib", for example "/usr/lib64" which some 3063 systems need. 3064 [Steve Henson, based on patch from Jeremy Utley] 3065 3066 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of 3067 X690 8.9.12 and can produce some misleading textual output of OIDs. 3068 [Steve Henson, reported by Dan Kaminsky] 3069 3070 *) Delete MD2 from algorithm tables. This follows the recommendation in 3071 several standards that it is not used in new applications due to 3072 several cryptographic weaknesses. For binary compatibility reasons 3073 the MD2 API is still compiled in by default. 3074 [Steve Henson] 3075 3076 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved 3077 and restored. 3078 [Steve Henson] 3079 3080 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and 3081 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name 3082 clash. 3083 [Guenter <lists@gknw.net>] 3084 3085 *) Fix the server certificate chain building code to use X509_verify_cert(), 3086 it used to have an ad-hoc builder which was unable to cope with anything 3087 other than a simple chain. 3088 [David Woodhouse <dwmw2@infradead.org>, Steve Henson] 3089 3090 *) Don't check self signed certificate signatures in X509_verify_cert() 3091 by default (a flag can override this): it just wastes time without 3092 adding any security. As a useful side effect self signed root CAs 3093 with non-FIPS digests are now usable in FIPS mode. 3094 [Steve Henson] 3095 3096 *) In dtls1_process_out_of_seq_message() the check if the current message 3097 is already buffered was missing. For every new message was memory 3098 allocated, allowing an attacker to perform an denial of service attack 3099 with sending out of seq handshake messages until there is no memory 3100 left. Additionally every future messege was buffered, even if the 3101 sequence number made no sense and would be part of another handshake. 3102 So only messages with sequence numbers less than 10 in advance will be 3103 buffered. (CVE-2009-1378) 3104 [Robin Seggelmann, discovered by Daniel Mentz] 3105 3106 *) Records are buffered if they arrive with a future epoch to be 3107 processed after finishing the corresponding handshake. There is 3108 currently no limitation to this buffer allowing an attacker to perform 3109 a DOS attack with sending records with future epochs until there is no 3110 memory left. This patch adds the pqueue_size() function to detemine 3111 the size of a buffer and limits the record buffer to 100 entries. 3112 (CVE-2009-1377) 3113 [Robin Seggelmann, discovered by Daniel Mentz] 3114 3115 *) Keep a copy of frag->msg_header.frag_len so it can be used after the 3116 parent structure is freed. (CVE-2009-1379) 3117 [Daniel Mentz] 3118 3119 *) Handle non-blocking I/O properly in SSL_shutdown() call. 3120 [Darryl Miles <darryl-mailinglists@netbauds.net>] 3121 3122 *) Add 2.5.4.* OIDs 3123 [Ilya O. <vrghost@gmail.com>] 3124 3125 Changes between 0.9.8k and 0.9.8l [5 Nov 2009] 3126 3127 *) Disable renegotiation completely - this fixes a severe security 3128 problem (CVE-2009-3555) at the cost of breaking all 3129 renegotiation. Renegotiation can be re-enabled by setting 3130 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at 3131 run-time. This is really not recommended unless you know what 3132 you're doing. 3133 [Ben Laurie] 3134 3135 Changes between 0.9.8j and 0.9.8k [25 Mar 2009] 3136 3137 *) Don't set val to NULL when freeing up structures, it is freed up by 3138 underlying code. If sizeof(void *) > sizeof(long) this can result in 3139 zeroing past the valid field. (CVE-2009-0789) 3140 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>] 3141 3142 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not 3143 checked correctly. This would allow some invalid signed attributes to 3144 appear to verify correctly. (CVE-2009-0591) 3145 [Ivan Nestlerode <inestlerode@us.ibm.com>] 3146 3147 *) Reject UniversalString and BMPString types with invalid lengths. This 3148 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have 3149 a legal length. (CVE-2009-0590) 3150 [Steve Henson] 3151 3152 *) Set S/MIME signing as the default purpose rather than setting it 3153 unconditionally. This allows applications to override it at the store 3154 level. 3155 [Steve Henson] 3156 3157 *) Permit restricted recursion of ASN1 strings. This is needed in practice 3158 to handle some structures. 3159 [Steve Henson] 3160 3161 *) Improve efficiency of mem_gets: don't search whole buffer each time 3162 for a '\n' 3163 [Jeremy Shapiro <jnshapir@us.ibm.com>] 3164 3165 *) New -hex option for openssl rand. 3166 [Matthieu Herrb] 3167 3168 *) Print out UTF8String and NumericString when parsing ASN1. 3169 [Steve Henson] 3170 3171 *) Support NumericString type for name components. 3172 [Steve Henson] 3173 3174 *) Allow CC in the environment to override the automatically chosen 3175 compiler. Note that nothing is done to ensure flags work with the 3176 chosen compiler. 3177 [Ben Laurie] 3178 3179 Changes between 0.9.8i and 0.9.8j [07 Jan 2009] 3180 3181 *) Properly check EVP_VerifyFinal() and similar return values 3182 (CVE-2008-5077). 3183 [Ben Laurie, Bodo Moeller, Google Security Team] 3184 3185 *) Enable TLS extensions by default. 3186 [Ben Laurie] 3187 3188 *) Allow the CHIL engine to be loaded, whether the application is 3189 multithreaded or not. (This does not release the developer from the 3190 obligation to set up the dynamic locking callbacks.) 3191 [Sander Temme <sander@temme.net>] 3192 3193 *) Use correct exit code if there is an error in dgst command. 3194 [Steve Henson; problem pointed out by Roland Dirlewanger] 3195 3196 *) Tweak Configure so that you need to say "experimental-jpake" to enable 3197 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. 3198 [Bodo Moeller] 3199 3200 *) Add experimental JPAKE support, including demo authentication in 3201 s_client and s_server. 3202 [Ben Laurie] 3203 3204 *) Set the comparison function in v3_addr_canonize(). 3205 [Rob Austein <sra@hactrn.net>] 3206 3207 *) Add support for XMPP STARTTLS in s_client. 3208 [Philip Paeps <philip@freebsd.org>] 3209 3210 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior 3211 to ensure that even with this option, only ciphersuites in the 3212 server's preference list will be accepted. (Note that the option 3213 applies only when resuming a session, so the earlier behavior was 3214 just about the algorithm choice for symmetric cryptography.) 3215 [Bodo Moeller] 3216 3217 Changes between 0.9.8h and 0.9.8i [15 Sep 2008] 3218 3219 *) Fix NULL pointer dereference if a DTLS server received 3220 ChangeCipherSpec as first record (CVE-2009-1386). 3221 [PR #1679] 3222 3223 *) Fix a state transitition in s3_srvr.c and d1_srvr.c 3224 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...). 3225 [Nagendra Modadugu] 3226 3227 *) The fix in 0.9.8c that supposedly got rid of unsafe 3228 double-checked locking was incomplete for RSA blinding, 3229 addressing just one layer of what turns out to have been 3230 doubly unsafe triple-checked locking. 3231 3232 So now fix this for real by retiring the MONT_HELPER macro 3233 in crypto/rsa/rsa_eay.c. 3234 3235 [Bodo Moeller; problem pointed out by Marius Schilder] 3236 3237 *) Various precautionary measures: 3238 3239 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). 3240 3241 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). 3242 (NB: This would require knowledge of the secret session ticket key 3243 to exploit, in which case you'd be SOL either way.) 3244 3245 - Change bn_nist.c so that it will properly handle input BIGNUMs 3246 outside the expected range. 3247 3248 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG 3249 builds. 3250 3251 [Neel Mehta, Bodo Moeller] 3252 3253 *) Allow engines to be "soft loaded" - i.e. optionally don't die if 3254 the load fails. Useful for distros. 3255 [Ben Laurie and the FreeBSD team] 3256 3257 *) Add support for Local Machine Keyset attribute in PKCS#12 files. 3258 [Steve Henson] 3259 3260 *) Fix BN_GF2m_mod_arr() top-bit cleanup code. 3261 [Huang Ying] 3262 3263 *) Expand ENGINE to support engine supplied SSL client certificate functions. 3264 3265 This work was sponsored by Logica. 3266 [Steve Henson] 3267 3268 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows 3269 keystores. Support for SSL/TLS client authentication too. 3270 Not compiled unless enable-capieng specified to Configure. 3271 3272 This work was sponsored by Logica. 3273 [Steve Henson] 3274 3275 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using 3276 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain 3277 attribute creation routines such as certifcate requests and PKCS#12 3278 files. 3279 [Steve Henson] 3280 3281 Changes between 0.9.8g and 0.9.8h [28 May 2008] 3282 3283 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS 3284 handshake which could lead to a cilent crash as found using the 3285 Codenomicon TLS test suite (CVE-2008-1672) 3286 [Steve Henson, Mark Cox] 3287 3288 *) Fix double free in TLS server name extensions which could lead to 3289 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891) 3290 [Joe Orton] 3291 3292 *) Clear error queue in SSL_CTX_use_certificate_chain_file() 3293 3294 Clear the error queue to ensure that error entries left from 3295 older function calls do not interfere with the correct operation. 3296 [Lutz Jaenicke, Erik de Castro Lopo] 3297 3298 *) Remove root CA certificates of commercial CAs: 3299 3300 The OpenSSL project does not recommend any specific CA and does not 3301 have any policy with respect to including or excluding any CA. 3302 Therefore it does not make any sense to ship an arbitrary selection 3303 of root CA certificates with the OpenSSL software. 3304 [Lutz Jaenicke] 3305 3306 *) RSA OAEP patches to fix two separate invalid memory reads. 3307 The first one involves inputs when 'lzero' is greater than 3308 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes 3309 before the beginning of from). The second one involves inputs where 3310 the 'db' section contains nothing but zeroes (there is a one-byte 3311 invalid read after the end of 'db'). 3312 [Ivan Nestlerode <inestlerode@us.ibm.com>] 3313 3314 *) Partial backport from 0.9.9-dev: 3315 3316 Introduce bn_mul_mont (dedicated Montgomery multiplication 3317 procedure) as a candidate for BIGNUM assembler implementation. 3318 While 0.9.9-dev uses assembler for various architectures, only 3319 x86_64 is available by default here in the 0.9.8 branch, and 3320 32-bit x86 is available through a compile-time setting. 3321 3322 To try the 32-bit x86 assembler implementation, use Configure 3323 option "enable-montasm" (which exists only for this backport). 3324 3325 As "enable-montasm" for 32-bit x86 disclaims code stability 3326 anyway, in this constellation we activate additional code 3327 backported from 0.9.9-dev for further performance improvements, 3328 namely BN_from_montgomery_word. (To enable this otherwise, 3329 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".) 3330 3331 [Andy Polyakov (backport partially by Bodo Moeller)] 3332 3333 *) Add TLS session ticket callback. This allows an application to set 3334 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed 3335 values. This is useful for key rollover for example where several key 3336 sets may exist with different names. 3337 [Steve Henson] 3338 3339 *) Reverse ENGINE-internal logic for caching default ENGINE handles. 3340 This was broken until now in 0.9.8 releases, such that the only way 3341 a registered ENGINE could be used (assuming it initialises 3342 successfully on the host) was to explicitly set it as the default 3343 for the relevant algorithms. This is in contradiction with 0.9.7 3344 behaviour and the documentation. With this fix, when an ENGINE is 3345 registered into a given algorithm's table of implementations, the 3346 'uptodate' flag is reset so that auto-discovery will be used next 3347 time a new context for that algorithm attempts to select an 3348 implementation. 3349 [Ian Lister (tweaked by Geoff Thorpe)] 3350 3351 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 3352 implemention in the following ways: 3353 3354 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be 3355 hard coded. 3356 3357 Lack of BER streaming support means one pass streaming processing is 3358 only supported if data is detached: setting the streaming flag is 3359 ignored for embedded content. 3360 3361 CMS support is disabled by default and must be explicitly enabled 3362 with the enable-cms configuration option. 3363 [Steve Henson] 3364 3365 *) Update the GMP engine glue to do direct copies between BIGNUM and 3366 mpz_t when openssl and GMP use the same limb size. Otherwise the 3367 existing "conversion via a text string export" trick is still used. 3368 [Paul Sheer <paulsheer@gmail.com>] 3369 3370 *) Zlib compression BIO. This is a filter BIO which compressed and 3371 uncompresses any data passed through it. 3372 [Steve Henson] 3373 3374 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement 3375 RFC3394 compatible AES key wrapping. 3376 [Steve Henson] 3377 3378 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): 3379 sets string data without copying. X509_ALGOR_set0() and 3380 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) 3381 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data 3382 from an X509_ATTRIBUTE structure optionally checking it occurs only 3383 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied 3384 data. 3385 [Steve Henson] 3386 3387 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() 3388 to get the expected BN_FLG_CONSTTIME behavior. 3389 [Bodo Moeller (Google)] 3390 3391 *) Netware support: 3392 3393 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets 3394 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) 3395 - added some more tests to do_tests.pl 3396 - fixed RunningProcess usage so that it works with newer LIBC NDKs too 3397 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency 3398 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, 3399 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc 3400 - various changes to netware.pl to enable gcc-cross builds on Win32 3401 platform 3402 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) 3403 - various changes to fix missing prototype warnings 3404 - fixed x86nasm.pl to create correct asm files for NASM COFF output 3405 - added AES, WHIRLPOOL and CPUID assembler code to build files 3406 - added missing AES assembler make rules to mk1mf.pl 3407 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply 3408 [Guenter Knauf <eflash@gmx.net>] 3409 3410 *) Implement certificate status request TLS extension defined in RFC3546. 3411 A client can set the appropriate parameters and receive the encoded 3412 OCSP response via a callback. A server can query the supplied parameters 3413 and set the encoded OCSP response in the callback. Add simplified examples 3414 to s_client and s_server. 3415 [Steve Henson] 3416 3417 Changes between 0.9.8f and 0.9.8g [19 Oct 2007] 3418 3419 *) Fix various bugs: 3420 + Binary incompatibility of ssl_ctx_st structure 3421 + DTLS interoperation with non-compliant servers 3422 + Don't call get_session_cb() without proposed session 3423 + Fix ia64 assembler code 3424 [Andy Polyakov, Steve Henson] 3425 3426 Changes between 0.9.8e and 0.9.8f [11 Oct 2007] 3427 3428 *) DTLS Handshake overhaul. There were longstanding issues with 3429 OpenSSL DTLS implementation, which were making it impossible for 3430 RFC 4347 compliant client to communicate with OpenSSL server. 3431 Unfortunately just fixing these incompatibilities would "cut off" 3432 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e 3433 server keeps tolerating non RFC compliant syntax. The opposite is 3434 not true, 0.9.8f client can not communicate with earlier server. 3435 This update even addresses CVE-2007-4995. 3436 [Andy Polyakov] 3437 3438 *) Changes to avoid need for function casts in OpenSSL: some compilers 3439 (gcc 4.2 and later) reject their use. 3440 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, 3441 Steve Henson] 3442 3443 *) Add RFC4507 support to OpenSSL. This includes the corrections in 3444 RFC4507bis. The encrypted ticket format is an encrypted encoded 3445 SSL_SESSION structure, that way new session features are automatically 3446 supported. 3447 3448 If a client application caches session in an SSL_SESSION structure 3449 support is transparent because tickets are now stored in the encoded 3450 SSL_SESSION. 3451 3452 The SSL_CTX structure automatically generates keys for ticket 3453 protection in servers so again support should be possible 3454 with no application modification. 3455 3456 If a client or server wishes to disable RFC4507 support then the option 3457 SSL_OP_NO_TICKET can be set. 3458 3459 Add a TLS extension debugging callback to allow the contents of any client 3460 or server extensions to be examined. 3461 3462 This work was sponsored by Google. 3463 [Steve Henson] 3464 3465 *) Add initial support for TLS extensions, specifically for the server_name 3466 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 3467 have new members for a host name. The SSL data structure has an 3468 additional member SSL_CTX *initial_ctx so that new sessions can be 3469 stored in that context to allow for session resumption, even after the 3470 SSL has been switched to a new SSL_CTX in reaction to a client's 3471 server_name extension. 3472 3473 New functions (subject to change): 3474 3475 SSL_get_servername() 3476 SSL_get_servername_type() 3477 SSL_set_SSL_CTX() 3478 3479 New CTRL codes and macros (subject to change): 3480 3481 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 3482 - SSL_CTX_set_tlsext_servername_callback() 3483 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 3484 - SSL_CTX_set_tlsext_servername_arg() 3485 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 3486 3487 openssl s_client has a new '-servername ...' option. 3488 3489 openssl s_server has new options '-servername_host ...', '-cert2 ...', 3490 '-key2 ...', '-servername_fatal' (subject to change). This allows 3491 testing the HostName extension for a specific single host name ('-cert' 3492 and '-key' remain fallbacks for handshakes without HostName 3493 negotiation). If the unrecogninzed_name alert has to be sent, this by 3494 default is a warning; it becomes fatal with the '-servername_fatal' 3495 option. 3496 3497 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson] 3498 3499 *) Add AES and SSE2 assembly language support to VC++ build. 3500 [Steve Henson] 3501 3502 *) Mitigate attack on final subtraction in Montgomery reduction. 3503 [Andy Polyakov] 3504 3505 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 3506 (which previously caused an internal error). 3507 [Bodo Moeller] 3508 3509 *) Squeeze another 10% out of IGE mode when in != out. 3510 [Ben Laurie] 3511 3512 *) AES IGE mode speedup. 3513 [Dean Gaudet (Google)] 3514 3515 *) Add the Korean symmetric 128-bit cipher SEED (see 3516 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and 3517 add SEED ciphersuites from RFC 4162: 3518 3519 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" 3520 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" 3521 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" 3522 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" 3523 3524 To minimize changes between patchlevels in the OpenSSL 0.9.8 3525 series, SEED remains excluded from compilation unless OpenSSL 3526 is configured with 'enable-seed'. 3527 [KISA, Bodo Moeller] 3528 3529 *) Mitigate branch prediction attacks, which can be practical if a 3530 single processor is shared, allowing a spy process to extract 3531 information. For detailed background information, see 3532 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron, 3533 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL 3534 and Necessary Software Countermeasures"). The core of the change 3535 are new versions BN_div_no_branch() and 3536 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), 3537 respectively, which are slower, but avoid the security-relevant 3538 conditional branches. These are automatically called by BN_div() 3539 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one 3540 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to 3541 remove a conditional branch. 3542 3543 BN_FLG_CONSTTIME is the new name for the previous 3544 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just 3545 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag 3546 in the exponent causes BN_mod_exp_mont() to use the alternative 3547 implementation in BN_mod_exp_mont_consttime().) The old name 3548 remains as a deprecated alias. 3549 3550 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general 3551 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses 3552 constant-time implementations for more than just exponentiation. 3553 Here too the old name is kept as a deprecated alias. 3554 3555 BN_BLINDING_new() will now use BN_dup() for the modulus so that 3556 the BN_BLINDING structure gets an independent copy of the 3557 modulus. This means that the previous "BIGNUM *m" argument to 3558 BN_BLINDING_new() and to BN_BLINDING_create_param() now 3559 essentially becomes "const BIGNUM *m", although we can't actually 3560 change this in the header file before 0.9.9. It allows 3561 RSA_setup_blinding() to use BN_with_flags() on the modulus to 3562 enable BN_FLG_CONSTTIME. 3563 3564 [Matthew D Wood (Intel Corp)] 3565 3566 *) In the SSL/TLS server implementation, be strict about session ID 3567 context matching (which matters if an application uses a single 3568 external cache for different purposes). Previously, 3569 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was 3570 set. This did ensure strict client verification, but meant that, 3571 with applications using a single external cache for quite 3572 different requirements, clients could circumvent ciphersuite 3573 restrictions for a given session ID context by starting a session 3574 in a different context. 3575 [Bodo Moeller] 3576 3577 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 3578 a ciphersuite string such as "DEFAULT:RSA" cannot enable 3579 authentication-only ciphersuites. 3580 [Bodo Moeller] 3581 3582 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was 3583 not complete and could lead to a possible single byte overflow 3584 (CVE-2007-5135) [Ben Laurie] 3585 3586 Changes between 0.9.8d and 0.9.8e [23 Feb 2007] 3587 3588 *) Since AES128 and AES256 (and similarly Camellia128 and 3589 Camellia256) share a single mask bit in the logic of 3590 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 3591 kludge to work properly if AES128 is available and AES256 isn't 3592 (or if Camellia128 is available and Camellia256 isn't). 3593 [Victor Duchovni] 3594 3595 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c 3596 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): 3597 When a point or a seed is encoded in a BIT STRING, we need to 3598 prevent the removal of trailing zero bits to get the proper DER 3599 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case 3600 of a NamedBitList, for which trailing 0 bits need to be removed.) 3601 [Bodo Moeller] 3602 3603 *) Have SSL/TLS server implementation tolerate "mismatched" record 3604 protocol version while receiving ClientHello even if the 3605 ClientHello is fragmented. (The server can't insist on the 3606 particular protocol version it has chosen before the ServerHello 3607 message has informed the client about his choice.) 3608 [Bodo Moeller] 3609 3610 *) Add RFC 3779 support. 3611 [Rob Austein for ARIN, Ben Laurie] 3612 3613 *) Load error codes if they are not already present instead of using a 3614 static variable. This allows them to be cleanly unloaded and reloaded. 3615 Improve header file function name parsing. 3616 [Steve Henson] 3617 3618 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO 3619 or CAPABILITY handshake as required by RFCs. 3620 [Goetz Babin-Ebell] 3621 3622 Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 3623 3624 *) Introduce limits to prevent malicious keys being able to 3625 cause a denial of service. (CVE-2006-2940) 3626 [Steve Henson, Bodo Moeller] 3627 3628 *) Fix ASN.1 parsing of certain invalid structures that can result 3629 in a denial of service. (CVE-2006-2937) [Steve Henson] 3630 3631 *) Fix buffer overflow in SSL_get_shared_ciphers() function. 3632 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 3633 3634 *) Fix SSL client code which could crash if connecting to a 3635 malicious SSLv2 server. (CVE-2006-4343) 3636 [Tavis Ormandy and Will Drewry, Google Security Team] 3637 3638 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 3639 match only those. Before that, "AES256-SHA" would be interpreted 3640 as a pattern and match "AES128-SHA" too (since AES128-SHA got 3641 the same strength classification in 0.9.7h) as we currently only 3642 have a single AES bit in the ciphersuite description bitmap. 3643 That change, however, also applied to ciphersuite strings such as 3644 "RC4-MD5" that intentionally matched multiple ciphersuites -- 3645 namely, SSL 2.0 ciphersuites in addition to the more common ones 3646 from SSL 3.0/TLS 1.0. 3647 3648 So we change the selection algorithm again: Naming an explicit 3649 ciphersuite selects this one ciphersuite, and any other similar 3650 ciphersuite (same bitmap) from *other* protocol versions. 3651 Thus, "RC4-MD5" again will properly select both the SSL 2.0 3652 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 3653 3654 Since SSL 2.0 does not have any ciphersuites for which the 3655 128/256 bit distinction would be relevant, this works for now. 3656 The proper fix will be to use different bits for AES128 and 3657 AES256, which would have avoided the problems from the beginning; 3658 however, bits are scarce, so we can only do this in a new release 3659 (not just a patchlevel) when we can change the SSL_CIPHER 3660 definition to split the single 'unsigned long mask' bitmap into 3661 multiple values to extend the available space. 3662 3663 [Bodo Moeller] 3664 3665 Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 3666 3667 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 3668 (CVE-2006-4339) [Ben Laurie and Google Security Team] 3669 3670 *) Add AES IGE and biIGE modes. 3671 [Ben Laurie] 3672 3673 *) Change the Unix randomness entropy gathering to use poll() when 3674 possible instead of select(), since the latter has some 3675 undesirable limitations. 3676 [Darryl Miles via Richard Levitte and Bodo Moeller] 3677 3678 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special 3679 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 3680 cannot be implicitly activated as part of, e.g., the "AES" alias. 3681 However, please upgrade to OpenSSL 0.9.9[-dev] for 3682 non-experimental use of the ECC ciphersuites to get TLS extension 3683 support, which is required for curve and point format negotiation 3684 to avoid potential handshake problems. 3685 [Bodo Moeller] 3686 3687 *) Disable rogue ciphersuites: 3688 3689 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 3690 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 3691 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 3692 3693 The latter two were purportedly from 3694 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 3695 appear there. 3696 3697 Also deactivate the remaining ciphersuites from 3698 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 3699 unofficial, and the ID has long expired. 3700 [Bodo Moeller] 3701 3702 *) Fix RSA blinding Heisenbug (problems sometimes occured on 3703 dual-core machines) and other potential thread-safety issues. 3704 [Bodo Moeller] 3705 3706 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 3707 versions), which is now available for royalty-free use 3708 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). 3709 Also, add Camellia TLS ciphersuites from RFC 4132. 3710 3711 To minimize changes between patchlevels in the OpenSSL 0.9.8 3712 series, Camellia remains excluded from compilation unless OpenSSL 3713 is configured with 'enable-camellia'. 3714 [NTT] 3715 3716 *) Disable the padding bug check when compression is in use. The padding 3717 bug check assumes the first packet is of even length, this is not 3718 necessarily true if compresssion is enabled and can result in false 3719 positives causing handshake failure. The actual bug test is ancient 3720 code so it is hoped that implementations will either have fixed it by 3721 now or any which still have the bug do not support compression. 3722 [Steve Henson] 3723 3724 Changes between 0.9.8a and 0.9.8b [04 May 2006] 3725 3726 *) When applying a cipher rule check to see if string match is an explicit 3727 cipher suite and only match that one cipher suite if it is. 3728 [Steve Henson] 3729 3730 *) Link in manifests for VC++ if needed. 3731 [Austin Ziegler <halostatue@gmail.com>] 3732 3733 *) Update support for ECC-based TLS ciphersuites according to 3734 draft-ietf-tls-ecc-12.txt with proposed changes (but without 3735 TLS extensions, which are supported starting with the 0.9.9 3736 branch, not in the OpenSSL 0.9.8 branch). 3737 [Douglas Stebila] 3738 3739 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 3740 opaque EVP_CIPHER_CTX handling. 3741 [Steve Henson] 3742 3743 *) Fixes and enhancements to zlib compression code. We now only use 3744 "zlib1.dll" and use the default __cdecl calling convention on Win32 3745 to conform with the standards mentioned here: 3746 http://www.zlib.net/DLL_FAQ.txt 3747 Static zlib linking now works on Windows and the new --with-zlib-include 3748 --with-zlib-lib options to Configure can be used to supply the location 3749 of the headers and library. Gracefully handle case where zlib library 3750 can't be loaded. 3751 [Steve Henson] 3752 3753 *) Several fixes and enhancements to the OID generation code. The old code 3754 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 3755 handle numbers larger than ULONG_MAX, truncated printing and had a 3756 non standard OBJ_obj2txt() behaviour. 3757 [Steve Henson] 3758 3759 *) Add support for building of engines under engine/ as shared libraries 3760 under VC++ build system. 3761 [Steve Henson] 3762 3763 *) Corrected the numerous bugs in the Win32 path splitter in DSO. 3764 Hopefully, we will not see any false combination of paths any more. 3765 [Richard Levitte] 3766 3767 Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 3768 3769 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 3770 (part of SSL_OP_ALL). This option used to disable the 3771 countermeasure against man-in-the-middle protocol-version 3772 rollback in the SSL 2.0 server implementation, which is a bad 3773 idea. (CVE-2005-2969) 3774 3775 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 3776 for Information Security, National Institute of Advanced Industrial 3777 Science and Technology [AIST], Japan)] 3778 3779 *) Add two function to clear and return the verify parameter flags. 3780 [Steve Henson] 3781 3782 *) Keep cipherlists sorted in the source instead of sorting them at 3783 runtime, thus removing the need for a lock. 3784 [Nils Larsch] 3785 3786 *) Avoid some small subgroup attacks in Diffie-Hellman. 3787 [Nick Mathewson and Ben Laurie] 3788 3789 *) Add functions for well-known primes. 3790 [Nick Mathewson] 3791 3792 *) Extended Windows CE support. 3793 [Satoshi Nakamura and Andy Polyakov] 3794 3795 *) Initialize SSL_METHOD structures at compile time instead of during 3796 runtime, thus removing the need for a lock. 3797 [Steve Henson] 3798 3799 *) Make PKCS7_decrypt() work even if no certificate is supplied by 3800 attempting to decrypt each encrypted key in turn. Add support to 3801 smime utility. 3802 [Steve Henson] 3803 3804 Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 3805 3806 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 3807 OpenSSL 0.9.8.] 3808 3809 *) Add libcrypto.pc and libssl.pc for those who feel they need them. 3810 [Richard Levitte] 3811 3812 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 3813 key into the same file any more. 3814 [Richard Levitte] 3815 3816 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 3817 [Andy Polyakov] 3818 3819 *) Add -utf8 command line and config file option to 'ca'. 3820 [Stefan <stf@udoma.org] 3821 3822 *) Removed the macro des_crypt(), as it seems to conflict with some 3823 libraries. Use DES_crypt(). 3824 [Richard Levitte] 3825 3826 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 3827 involves renaming the source and generated shared-libs for 3828 both. The engines will accept the corrected or legacy ids 3829 ('ncipher' and '4758_cca' respectively) when binding. NB, 3830 this only applies when building 'shared'. 3831 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 3832 3833 *) Add attribute functions to EVP_PKEY structure. Modify 3834 PKCS12_create() to recognize a CSP name attribute and 3835 use it. Make -CSP option work again in pkcs12 utility. 3836 [Steve Henson] 3837 3838 *) Add new functionality to the bn blinding code: 3839 - automatic re-creation of the BN_BLINDING parameters after 3840 a fixed number of uses (currently 32) 3841 - add new function for parameter creation 3842 - introduce flags to control the update behaviour of the 3843 BN_BLINDING parameters 3844 - hide BN_BLINDING structure 3845 Add a second BN_BLINDING slot to the RSA structure to improve 3846 performance when a single RSA object is shared among several 3847 threads. 3848 [Nils Larsch] 3849 3850 *) Add support for DTLS. 3851 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 3852 3853 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 3854 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 3855 [Walter Goulet] 3856 3857 *) Remove buggy and incompletet DH cert support from 3858 ssl/ssl_rsa.c and ssl/s3_both.c 3859 [Nils Larsch] 3860 3861 *) Use SHA-1 instead of MD5 as the default digest algorithm for 3862 the apps/openssl applications. 3863 [Nils Larsch] 3864 3865 *) Compile clean with "-Wall -Wmissing-prototypes 3866 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 3867 DEBUG_SAFESTACK must also be set. 3868 [Ben Laurie] 3869 3870 *) Change ./Configure so that certain algorithms can be disabled by default. 3871 The new counterpiece to "no-xxx" is "enable-xxx". 3872 3873 The patented RC5 and MDC2 algorithms will now be disabled unless 3874 "enable-rc5" and "enable-mdc2", respectively, are specified. 3875 3876 (IDEA remains enabled despite being patented. This is because IDEA 3877 is frequently required for interoperability, and there is no license 3878 fee for non-commercial use. As before, "no-idea" can be used to 3879 avoid this algorithm.) 3880 3881 [Bodo Moeller] 3882 3883 *) Add processing of proxy certificates (see RFC 3820). This work was 3884 sponsored by KTH (The Royal Institute of Technology in Stockholm) and 3885 EGEE (Enabling Grids for E-science in Europe). 3886 [Richard Levitte] 3887 3888 *) RC4 performance overhaul on modern architectures/implementations, such 3889 as Intel P4, IA-64 and AMD64. 3890 [Andy Polyakov] 3891 3892 *) New utility extract-section.pl. This can be used specify an alternative 3893 section number in a pod file instead of having to treat each file as 3894 a separate case in Makefile. This can be done by adding two lines to the 3895 pod file: 3896 3897 =for comment openssl_section:XXX 3898 3899 The blank line is mandatory. 3900 3901 [Steve Henson] 3902 3903 *) New arguments -certform, -keyform and -pass for s_client and s_server 3904 to allow alternative format key and certificate files and passphrase 3905 sources. 3906 [Steve Henson] 3907 3908 *) New structure X509_VERIFY_PARAM which combines current verify parameters, 3909 update associated structures and add various utility functions. 3910 3911 Add new policy related verify parameters, include policy checking in 3912 standard verify code. Enhance 'smime' application with extra parameters 3913 to support policy checking and print out. 3914 [Steve Henson] 3915 3916 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 3917 Nehemiah processors. These extensions support AES encryption in hardware 3918 as well as RNG (though RNG support is currently disabled). 3919 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 3920 3921 *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 3922 [Geoff Thorpe] 3923 3924 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 3925 [Andy Polyakov and a number of other people] 3926 3927 *) Improved PowerPC platform support. Most notably BIGNUM assembler 3928 implementation contributed by IBM. 3929 [Suresh Chari, Peter Waltenberg, Andy Polyakov] 3930 3931 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 3932 exponent rather than 'unsigned long'. There is a corresponding change to 3933 the new 'rsa_keygen' element of the RSA_METHOD structure. 3934 [Jelte Jansen, Geoff Thorpe] 3935 3936 *) Functionality for creating the initial serial number file is now 3937 moved from CA.pl to the 'ca' utility with a new option -create_serial. 3938 3939 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 3940 number file to 1, which is bound to cause problems. To avoid 3941 the problems while respecting compatibility between different 0.9.7 3942 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 3943 CA.pl for serial number initialization. With the new release 0.9.8, 3944 we can fix the problem directly in the 'ca' utility.) 3945 [Steve Henson] 3946 3947 *) Reduced header interdepencies by declaring more opaque objects in 3948 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 3949 give fewer recursive includes, which could break lazy source code - so 3950 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 3951 developers should define this symbol when building and using openssl to 3952 ensure they track the recommended behaviour, interfaces, [etc], but 3953 backwards-compatible behaviour prevails when this isn't defined. 3954 [Geoff Thorpe] 3955 3956 *) New function X509_POLICY_NODE_print() which prints out policy nodes. 3957 [Steve Henson] 3958 3959 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 3960 This will generate a random key of the appropriate length based on the 3961 cipher context. The EVP_CIPHER can provide its own random key generation 3962 routine to support keys of a specific form. This is used in the des and 3963 3des routines to generate a key of the correct parity. Update S/MIME 3964 code to use new functions and hence generate correct parity DES keys. 3965 Add EVP_CHECK_DES_KEY #define to return an error if the key is not 3966 valid (weak or incorrect parity). 3967 [Steve Henson] 3968 3969 *) Add a local set of CRLs that can be used by X509_verify_cert() as well 3970 as looking them up. This is useful when the verified structure may contain 3971 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 3972 present unless the new PKCS7_NO_CRL flag is asserted. 3973 [Steve Henson] 3974 3975 *) Extend ASN1 oid configuration module. It now additionally accepts the 3976 syntax: 3977 3978 shortName = some long name, 1.2.3.4 3979 [Steve Henson] 3980 3981 *) Reimplemented the BN_CTX implementation. There is now no more static 3982 limitation on the number of variables it can handle nor the depth of the 3983 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 3984 information can now expand as required, and rather than having a single 3985 static array of bignums, BN_CTX now uses a linked-list of such arrays 3986 allowing it to expand on demand whilst maintaining the usefulness of 3987 BN_CTX's "bundling". 3988 [Geoff Thorpe] 3989 3990 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 3991 to allow all RSA operations to function using a single BN_CTX. 3992 [Geoff Thorpe] 3993 3994 *) Preliminary support for certificate policy evaluation and checking. This 3995 is initially intended to pass the tests outlined in "Conformance Testing 3996 of Relying Party Client Certificate Path Processing Logic" v1.07. 3997 [Steve Henson] 3998 3999 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 4000 remained unused and not that useful. A variety of other little bignum 4001 tweaks and fixes have also been made continuing on from the audit (see 4002 below). 4003 [Geoff Thorpe] 4004 4005 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 4006 associated ASN1, EVP and SSL functions and old ASN1 macros. 4007 [Richard Levitte] 4008 4009 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 4010 and this should never fail. So the return value from the use of 4011 BN_set_word() (which can fail due to needless expansion) is now deprecated; 4012 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 4013 [Geoff Thorpe] 4014 4015 *) BN_CTX_get() should return zero-valued bignums, providing the same 4016 initialised value as BN_new(). 4017 [Geoff Thorpe, suggested by Ulf M��ller] 4018 4019 *) Support for inhibitAnyPolicy certificate extension. 4020 [Steve Henson] 4021 4022 *) An audit of the BIGNUM code is underway, for which debugging code is 4023 enabled when BN_DEBUG is defined. This makes stricter enforcements on what 4024 is considered valid when processing BIGNUMs, and causes execution to 4025 assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 4026 further steps are taken to deliberately pollute unused data in BIGNUM 4027 structures to try and expose faulty code further on. For now, openssl will 4028 (in its default mode of operation) continue to tolerate the inconsistent 4029 forms that it has tolerated in the past, but authors and packagers should 4030 consider trying openssl and their own applications when compiled with 4031 these debugging symbols defined. It will help highlight potential bugs in 4032 their own code, and will improve the test coverage for OpenSSL itself. At 4033 some point, these tighter rules will become openssl's default to improve 4034 maintainability, though the assert()s and other overheads will remain only 4035 in debugging configurations. See bn.h for more details. 4036 [Geoff Thorpe, Nils Larsch, Ulf M��ller] 4037 4038 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 4039 that can only be obtained through BN_CTX_new() (which implicitly 4040 initialises it). The presence of this function only made it possible 4041 to overwrite an existing structure (and cause memory leaks). 4042 [Geoff Thorpe] 4043 4044 *) Because of the callback-based approach for implementing LHASH as a 4045 template type, lh_insert() adds opaque objects to hash-tables and 4046 lh_doall() or lh_doall_arg() are typically used with a destructor callback 4047 to clean up those corresponding objects before destroying the hash table 4048 (and losing the object pointers). So some over-zealous constifications in 4049 LHASH have been relaxed so that lh_insert() does not take (nor store) the 4050 objects as "const" and the lh_doall[_arg] callback wrappers are not 4051 prototyped to have "const" restrictions on the object pointers they are 4052 given (and so aren't required to cast them away any more). 4053 [Geoff Thorpe] 4054 4055 *) The tmdiff.h API was so ugly and minimal that our own timing utility 4056 (speed) prefers to use its own implementation. The two implementations 4057 haven't been consolidated as yet (volunteers?) but the tmdiff API has had 4058 its object type properly exposed (MS_TM) instead of casting to/from "char 4059 *". This may still change yet if someone realises MS_TM and "ms_time_***" 4060 aren't necessarily the greatest nomenclatures - but this is what was used 4061 internally to the implementation so I've used that for now. 4062 [Geoff Thorpe] 4063 4064 *) Ensure that deprecated functions do not get compiled when 4065 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 4066 the self-tests were still using deprecated key-generation functions so 4067 these have been updated also. 4068 [Geoff Thorpe] 4069 4070 *) Reorganise PKCS#7 code to separate the digest location functionality 4071 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 4072 New function PKCS7_set_digest() to set the digest type for PKCS#7 4073 digestedData type. Add additional code to correctly generate the 4074 digestedData type and add support for this type in PKCS7 initialization 4075 functions. 4076 [Steve Henson] 4077 4078 *) New function PKCS7_set0_type_other() this initializes a PKCS7 4079 structure of type "other". 4080 [Steve Henson] 4081 4082 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 4083 sure the loop does correctly stop and breaking ("division by zero") 4084 modulus operations are not performed. The (pre-generated) prime 4085 table crypto/bn/bn_prime.h was already correct, but it could not be 4086 re-generated on some platforms because of the "division by zero" 4087 situation in the script. 4088 [Ralf S. Engelschall] 4089 4090 *) Update support for ECC-based TLS ciphersuites according to 4091 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 4092 SHA-1 now is only used for "small" curves (where the 4093 representation of a field element takes up to 24 bytes); for 4094 larger curves, the field element resulting from ECDH is directly 4095 used as premaster secret. 4096 [Douglas Stebila (Sun Microsystems Laboratories)] 4097 4098 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 4099 curve secp160r1 to the tests. 4100 [Douglas Stebila (Sun Microsystems Laboratories)] 4101 4102 *) Add the possibility to load symbols globally with DSO. 4103 [G��tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 4104 4105 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 4106 control of the error stack. 4107 [Richard Levitte] 4108 4109 *) Add support for STORE in ENGINE. 4110 [Richard Levitte] 4111 4112 *) Add the STORE type. The intention is to provide a common interface 4113 to certificate and key stores, be they simple file-based stores, or 4114 HSM-type store, or LDAP stores, or... 4115 NOTE: The code is currently UNTESTED and isn't really used anywhere. 4116 [Richard Levitte] 4117 4118 *) Add a generic structure called OPENSSL_ITEM. This can be used to 4119 pass a list of arguments to any function as well as provide a way 4120 for a function to pass data back to the caller. 4121 [Richard Levitte] 4122 4123 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 4124 works like BUF_strdup() but can be used to duplicate a portion of 4125 a string. The copy gets NUL-terminated. BUF_memdup() duplicates 4126 a memory area. 4127 [Richard Levitte] 4128 4129 *) Add the function sk_find_ex() which works like sk_find(), but will 4130 return an index to an element even if an exact match couldn't be 4131 found. The index is guaranteed to point at the element where the 4132 searched-for key would be inserted to preserve sorting order. 4133 [Richard Levitte] 4134 4135 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 4136 takes an extra flags argument for optional functionality. Currently, 4137 the following flags are defined: 4138 4139 OBJ_BSEARCH_VALUE_ON_NOMATCH 4140 This one gets OBJ_bsearch_ex() to return a pointer to the first 4141 element where the comparing function returns a negative or zero 4142 number. 4143 4144 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 4145 This one gets OBJ_bsearch_ex() to return a pointer to the first 4146 element where the comparing function returns zero. This is useful 4147 if there are more than one element where the comparing function 4148 returns zero. 4149 [Richard Levitte] 4150 4151 *) Make it possible to create self-signed certificates with 'openssl ca' 4152 in such a way that the self-signed certificate becomes part of the 4153 CA database and uses the same mechanisms for serial number generation 4154 as all other certificate signing. The new flag '-selfsign' enables 4155 this functionality. Adapt CA.sh and CA.pl.in. 4156 [Richard Levitte] 4157 4158 *) Add functionality to check the public key of a certificate request 4159 against a given private. This is useful to check that a certificate 4160 request can be signed by that key (self-signing). 4161 [Richard Levitte] 4162 4163 *) Make it possible to have multiple active certificates with the same 4164 subject in the CA index file. This is done only if the keyword 4165 'unique_subject' is set to 'no' in the main CA section (default 4166 if 'CA_default') of the configuration file. The value is saved 4167 with the database itself in a separate index attribute file, 4168 named like the index file with '.attr' appended to the name. 4169 [Richard Levitte] 4170 4171 *) Generate muti valued AVAs using '+' notation in config files for 4172 req and dirName. 4173 [Steve Henson] 4174 4175 *) Support for nameConstraints certificate extension. 4176 [Steve Henson] 4177 4178 *) Support for policyConstraints certificate extension. 4179 [Steve Henson] 4180 4181 *) Support for policyMappings certificate extension. 4182 [Steve Henson] 4183 4184 *) Make sure the default DSA_METHOD implementation only uses its 4185 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 4186 and change its own handlers to be NULL so as to remove unnecessary 4187 indirection. This lets alternative implementations fallback to the 4188 default implementation more easily. 4189 [Geoff Thorpe] 4190 4191 *) Support for directoryName in GeneralName related extensions 4192 in config files. 4193 [Steve Henson] 4194 4195 *) Make it possible to link applications using Makefile.shared. 4196 Make that possible even when linking against static libraries! 4197 [Richard Levitte] 4198 4199 *) Support for single pass processing for S/MIME signing. This now 4200 means that S/MIME signing can be done from a pipe, in addition 4201 cleartext signing (multipart/signed type) is effectively streaming 4202 and the signed data does not need to be all held in memory. 4203 4204 This is done with a new flag PKCS7_STREAM. When this flag is set 4205 PKCS7_sign() only initializes the PKCS7 structure and the actual signing 4206 is done after the data is output (and digests calculated) in 4207 SMIME_write_PKCS7(). 4208 [Steve Henson] 4209 4210 *) Add full support for -rpath/-R, both in shared libraries and 4211 applications, at least on the platforms where it's known how 4212 to do it. 4213 [Richard Levitte] 4214 4215 *) In crypto/ec/ec_mult.c, implement fast point multiplication with 4216 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 4217 will now compute a table of multiples of the generator that 4218 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 4219 faster (notably in the case of a single point multiplication, 4220 scalar * generator). 4221 [Nils Larsch, Bodo Moeller] 4222 4223 *) IPv6 support for certificate extensions. The various extensions 4224 which use the IP:a.b.c.d can now take IPv6 addresses using the 4225 formats of RFC1884 2.2 . IPv6 addresses are now also displayed 4226 correctly. 4227 [Steve Henson] 4228 4229 *) Added an ENGINE that implements RSA by performing private key 4230 exponentiations with the GMP library. The conversions to and from 4231 GMP's mpz_t format aren't optimised nor are any montgomery forms 4232 cached, and on x86 it appears OpenSSL's own performance has caught up. 4233 However there are likely to be other architectures where GMP could 4234 provide a boost. This ENGINE is not built in by default, but it can be 4235 specified at Configure time and should be accompanied by the necessary 4236 linker additions, eg; 4237 ./config -DOPENSSL_USE_GMP -lgmp 4238 [Geoff Thorpe] 4239 4240 *) "openssl engine" will not display ENGINE/DSO load failure errors when 4241 testing availability of engines with "-t" - the old behaviour is 4242 produced by increasing the feature's verbosity with "-tt". 4243 [Geoff Thorpe] 4244 4245 *) ECDSA routines: under certain error conditions uninitialized BN objects 4246 could be freed. Solution: make sure initialization is performed early 4247 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 4248 via PR#459) 4249 [Lutz Jaenicke] 4250 4251 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 4252 and DH_METHOD (eg. by ENGINE implementations) to override the normal 4253 software implementations. For DSA and DH, parameter generation can 4254 also be overriden by providing the appropriate method callbacks. 4255 [Geoff Thorpe] 4256 4257 *) Change the "progress" mechanism used in key-generation and 4258 primality testing to functions that take a new BN_GENCB pointer in 4259 place of callback/argument pairs. The new API functions have "_ex" 4260 postfixes and the older functions are reimplemented as wrappers for 4261 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 4262 declarations of the old functions to help (graceful) attempts to 4263 migrate to the new functions. Also, the new key-generation API 4264 functions operate on a caller-supplied key-structure and return 4265 success/failure rather than returning a key or NULL - this is to 4266 help make "keygen" another member function of RSA_METHOD etc. 4267 4268 Example for using the new callback interface: 4269 4270 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 4271 void *my_arg = ...; 4272 BN_GENCB my_cb; 4273 4274 BN_GENCB_set(&my_cb, my_callback, my_arg); 4275 4276 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 4277 /* For the meaning of a, b in calls to my_callback(), see the 4278 * documentation of the function that calls the callback. 4279 * cb will point to my_cb; my_arg can be retrieved as cb->arg. 4280 * my_callback should return 1 if it wants BN_is_prime_ex() 4281 * to continue, or 0 to stop. 4282 */ 4283 4284 [Geoff Thorpe] 4285 4286 *) Change the ZLIB compression method to be stateful, and make it 4287 available to TLS with the number defined in 4288 draft-ietf-tls-compression-04.txt. 4289 [Richard Levitte] 4290 4291 *) Add the ASN.1 structures and functions for CertificatePair, which 4292 is defined as follows (according to X.509_4thEditionDraftV6.pdf): 4293 4294 CertificatePair ::= SEQUENCE { 4295 forward [0] Certificate OPTIONAL, 4296 reverse [1] Certificate OPTIONAL, 4297 -- at least one of the pair shall be present -- } 4298 4299 Also implement the PEM functions to read and write certificate 4300 pairs, and defined the PEM tag as "CERTIFICATE PAIR". 4301 4302 This needed to be defined, mostly for the sake of the LDAP 4303 attribute crossCertificatePair, but may prove useful elsewhere as 4304 well. 4305 [Richard Levitte] 4306 4307 *) Make it possible to inhibit symlinking of shared libraries in 4308 Makefile.shared, for Cygwin's sake. 4309 [Richard Levitte] 4310 4311 *) Extend the BIGNUM API by creating a function 4312 void BN_set_negative(BIGNUM *a, int neg); 4313 and a macro that behave like 4314 int BN_is_negative(const BIGNUM *a); 4315 4316 to avoid the need to access 'a->neg' directly in applications. 4317 [Nils Larsch] 4318 4319 *) Implement fast modular reduction for pseudo-Mersenne primes 4320 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 4321 EC_GROUP_new_curve_GFp() will now automatically use this 4322 if applicable. 4323 [Nils Larsch <nla@trustcenter.de>] 4324 4325 *) Add new lock type (CRYPTO_LOCK_BN). 4326 [Bodo Moeller] 4327 4328 *) Change the ENGINE framework to automatically load engines 4329 dynamically from specific directories unless they could be 4330 found to already be built in or loaded. Move all the 4331 current engines except for the cryptodev one to a new 4332 directory engines/. 4333 The engines in engines/ are built as shared libraries if 4334 the "shared" options was given to ./Configure or ./config. 4335 Otherwise, they are inserted in libcrypto.a. 4336 /usr/local/ssl/engines is the default directory for dynamic 4337 engines, but that can be overriden at configure time through 4338 the usual use of --prefix and/or --openssldir, and at run 4339 time with the environment variable OPENSSL_ENGINES. 4340 [Geoff Thorpe and Richard Levitte] 4341 4342 *) Add Makefile.shared, a helper makefile to build shared 4343 libraries. Addapt Makefile.org. 4344 [Richard Levitte] 4345 4346 *) Add version info to Win32 DLLs. 4347 [Peter 'Luna' Runestig" <peter@runestig.com>] 4348 4349 *) Add new 'medium level' PKCS#12 API. Certificates and keys 4350 can be added using this API to created arbitrary PKCS#12 4351 files while avoiding the low level API. 4352 4353 New options to PKCS12_create(), key or cert can be NULL and 4354 will then be omitted from the output file. The encryption 4355 algorithm NIDs can be set to -1 for no encryption, the mac 4356 iteration count can be set to 0 to omit the mac. 4357 4358 Enhance pkcs12 utility by making the -nokeys and -nocerts 4359 options work when creating a PKCS#12 file. New option -nomac 4360 to omit the mac, NONE can be set for an encryption algorithm. 4361 New code is modified to use the enhanced PKCS12_create() 4362 instead of the low level API. 4363 [Steve Henson] 4364 4365 *) Extend ASN1 encoder to support indefinite length constructed 4366 encoding. This can output sequences tags and octet strings in 4367 this form. Modify pk7_asn1.c to support indefinite length 4368 encoding. This is experimental and needs additional code to 4369 be useful, such as an ASN1 bio and some enhanced streaming 4370 PKCS#7 code. 4371 4372 Extend template encode functionality so that tagging is passed 4373 down to the template encoder. 4374 [Steve Henson] 4375 4376 *) Let 'openssl req' fail if an argument to '-newkey' is not 4377 recognized instead of using RSA as a default. 4378 [Bodo Moeller] 4379 4380 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 4381 As these are not official, they are not included in "ALL"; 4382 the "ECCdraft" ciphersuite group alias can be used to select them. 4383 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 4384 4385 *) Add ECDH engine support. 4386 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 4387 4388 *) Add ECDH in new directory crypto/ecdh/. 4389 [Douglas Stebila (Sun Microsystems Laboratories)] 4390 4391 *) Let BN_rand_range() abort with an error after 100 iterations 4392 without success (which indicates a broken PRNG). 4393 [Bodo Moeller] 4394 4395 *) Change BN_mod_sqrt() so that it verifies that the input value 4396 is really the square of the return value. (Previously, 4397 BN_mod_sqrt would show GIGO behaviour.) 4398 [Bodo Moeller] 4399 4400 *) Add named elliptic curves over binary fields from X9.62, SECG, 4401 and WAP/WTLS; add OIDs that were still missing. 4402 4403 [Sheueling Chang Shantz and Douglas Stebila 4404 (Sun Microsystems Laboratories)] 4405 4406 *) Extend the EC library for elliptic curves over binary fields 4407 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 4408 New EC_METHOD: 4409 4410 EC_GF2m_simple_method 4411 4412 New API functions: 4413 4414 EC_GROUP_new_curve_GF2m 4415 EC_GROUP_set_curve_GF2m 4416 EC_GROUP_get_curve_GF2m 4417 EC_POINT_set_affine_coordinates_GF2m 4418 EC_POINT_get_affine_coordinates_GF2m 4419 EC_POINT_set_compressed_coordinates_GF2m 4420 4421 Point compression for binary fields is disabled by default for 4422 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 4423 enable it). 4424 4425 As binary polynomials are represented as BIGNUMs, various members 4426 of the EC_GROUP and EC_POINT data structures can be shared 4427 between the implementations for prime fields and binary fields; 4428 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 4429 are essentially identical to their ..._GFp counterparts. 4430 (For simplicity, the '..._GFp' prefix has been dropped from 4431 various internal method names.) 4432 4433 An internal 'field_div' method (similar to 'field_mul' and 4434 'field_sqr') has been added; this is used only for binary fields. 4435 4436 [Sheueling Chang Shantz and Douglas Stebila 4437 (Sun Microsystems Laboratories)] 4438 4439 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 4440 through methods ('mul', 'precompute_mult'). 4441 4442 The generic implementations (now internally called 'ec_wNAF_mul' 4443 and 'ec_wNAF_precomputed_mult') remain the default if these 4444 methods are undefined. 4445 4446 [Sheueling Chang Shantz and Douglas Stebila 4447 (Sun Microsystems Laboratories)] 4448 4449 *) New function EC_GROUP_get_degree, which is defined through 4450 EC_METHOD. For curves over prime fields, this returns the bit 4451 length of the modulus. 4452 4453 [Sheueling Chang Shantz and Douglas Stebila 4454 (Sun Microsystems Laboratories)] 4455 4456 *) New functions EC_GROUP_dup, EC_POINT_dup. 4457 (These simply call ..._new and ..._copy). 4458 4459 [Sheueling Chang Shantz and Douglas Stebila 4460 (Sun Microsystems Laboratories)] 4461 4462 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 4463 Polynomials are represented as BIGNUMs (where the sign bit is not 4464 used) in the following functions [macros]: 4465 4466 BN_GF2m_add 4467 BN_GF2m_sub [= BN_GF2m_add] 4468 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 4469 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 4470 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 4471 BN_GF2m_mod_inv 4472 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 4473 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 4474 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 4475 BN_GF2m_cmp [= BN_ucmp] 4476 4477 (Note that only the 'mod' functions are actually for fields GF(2^m). 4478 BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 4479 4480 For some functions, an the irreducible polynomial defining a 4481 field can be given as an 'unsigned int[]' with strictly 4482 decreasing elements giving the indices of those bits that are set; 4483 i.e., p[] represents the polynomial 4484 f(t) = t^p[0] + t^p[1] + ... + t^p[k] 4485 where 4486 p[0] > p[1] > ... > p[k] = 0. 4487 This applies to the following functions: 4488 4489 BN_GF2m_mod_arr 4490 BN_GF2m_mod_mul_arr 4491 BN_GF2m_mod_sqr_arr 4492 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 4493 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 4494 BN_GF2m_mod_exp_arr 4495 BN_GF2m_mod_sqrt_arr 4496 BN_GF2m_mod_solve_quad_arr 4497 BN_GF2m_poly2arr 4498 BN_GF2m_arr2poly 4499 4500 Conversion can be performed by the following functions: 4501 4502 BN_GF2m_poly2arr 4503 BN_GF2m_arr2poly 4504 4505 bntest.c has additional tests for binary polynomial arithmetic. 4506 4507 Two implementations for BN_GF2m_mod_div() are available. 4508 The default algorithm simply uses BN_GF2m_mod_inv() and 4509 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 4510 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 4511 copyright notice in crypto/bn/bn_gf2m.c before enabling it). 4512 4513 [Sheueling Chang Shantz and Douglas Stebila 4514 (Sun Microsystems Laboratories)] 4515 4516 *) Add new error code 'ERR_R_DISABLED' that can be used when some 4517 functionality is disabled at compile-time. 4518 [Douglas Stebila <douglas.stebila@sun.com>] 4519 4520 *) Change default behaviour of 'openssl asn1parse' so that more 4521 information is visible when viewing, e.g., a certificate: 4522 4523 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 4524 mode the content of non-printable OCTET STRINGs is output in a 4525 style similar to INTEGERs, but with '[HEX DUMP]' prepended to 4526 avoid the appearance of a printable string. 4527 [Nils Larsch <nla@trustcenter.de>] 4528 4529 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 4530 functions 4531 EC_GROUP_set_asn1_flag() 4532 EC_GROUP_get_asn1_flag() 4533 EC_GROUP_set_point_conversion_form() 4534 EC_GROUP_get_point_conversion_form() 4535 These control ASN1 encoding details: 4536 - Curves (i.e., groups) are encoded explicitly unless asn1_flag 4537 has been set to OPENSSL_EC_NAMED_CURVE. 4538 - Points are encoded in uncompressed form by default; options for 4539 asn1_for are as for point2oct, namely 4540 POINT_CONVERSION_COMPRESSED 4541 POINT_CONVERSION_UNCOMPRESSED 4542 POINT_CONVERSION_HYBRID 4543 4544 Also add 'seed' and 'seed_len' members to EC_GROUP with access 4545 functions 4546 EC_GROUP_set_seed() 4547 EC_GROUP_get0_seed() 4548 EC_GROUP_get_seed_len() 4549 This is used only for ASN1 purposes (so far). 4550 [Nils Larsch <nla@trustcenter.de>] 4551 4552 *) Add 'field_type' member to EC_METHOD, which holds the NID 4553 of the appropriate field type OID. The new function 4554 EC_METHOD_get_field_type() returns this value. 4555 [Nils Larsch <nla@trustcenter.de>] 4556 4557 *) Add functions 4558 EC_POINT_point2bn() 4559 EC_POINT_bn2point() 4560 EC_POINT_point2hex() 4561 EC_POINT_hex2point() 4562 providing useful interfaces to EC_POINT_point2oct() and 4563 EC_POINT_oct2point(). 4564 [Nils Larsch <nla@trustcenter.de>] 4565 4566 *) Change internals of the EC library so that the functions 4567 EC_GROUP_set_generator() 4568 EC_GROUP_get_generator() 4569 EC_GROUP_get_order() 4570 EC_GROUP_get_cofactor() 4571 are implemented directly in crypto/ec/ec_lib.c and not dispatched 4572 to methods, which would lead to unnecessary code duplication when 4573 adding different types of curves. 4574 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 4575 4576 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 4577 arithmetic, and such that modified wNAFs are generated 4578 (which avoid length expansion in many cases). 4579 [Bodo Moeller] 4580 4581 *) Add a function EC_GROUP_check_discriminant() (defined via 4582 EC_METHOD) that verifies that the curve discriminant is non-zero. 4583 4584 Add a function EC_GROUP_check() that makes some sanity tests 4585 on a EC_GROUP, its generator and order. This includes 4586 EC_GROUP_check_discriminant(). 4587 [Nils Larsch <nla@trustcenter.de>] 4588 4589 *) Add ECDSA in new directory crypto/ecdsa/. 4590 4591 Add applications 'openssl ecparam' and 'openssl ecdsa' 4592 (these are based on 'openssl dsaparam' and 'openssl dsa'). 4593 4594 ECDSA support is also included in various other files across the 4595 library. Most notably, 4596 - 'openssl req' now has a '-newkey ecdsa:file' option; 4597 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 4598 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 4599 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 4600 them suitable for ECDSA where domain parameters must be 4601 extracted before the specific public key; 4602 - ECDSA engine support has been added. 4603 [Nils Larsch <nla@trustcenter.de>] 4604 4605 *) Include some named elliptic curves, and add OIDs from X9.62, 4606 SECG, and WAP/WTLS. Each curve can be obtained from the new 4607 function 4608 EC_GROUP_new_by_curve_name(), 4609 and the list of available named curves can be obtained with 4610 EC_get_builtin_curves(). 4611 Also add a 'curve_name' member to EC_GROUP objects, which can be 4612 accessed via 4613 EC_GROUP_set_curve_name() 4614 EC_GROUP_get_curve_name() 4615 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 4616 4617 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 4618 was actually never needed) and in BN_mul(). The removal in BN_mul() 4619 required a small change in bn_mul_part_recursive() and the addition 4620 of the functions bn_cmp_part_words(), bn_sub_part_words() and 4621 bn_add_part_words(), which do the same thing as bn_cmp_words(), 4622 bn_sub_words() and bn_add_words() except they take arrays with 4623 differing sizes. 4624 [Richard Levitte] 4625 4626 Changes between 0.9.7l and 0.9.7m [23 Feb 2007] 4627 4628 *) Cleanse PEM buffers before freeing them since they may contain 4629 sensitive data. 4630 [Benjamin Bennett <ben@psc.edu>] 4631 4632 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 4633 a ciphersuite string such as "DEFAULT:RSA" cannot enable 4634 authentication-only ciphersuites. 4635 [Bodo Moeller] 4636 4637 *) Since AES128 and AES256 share a single mask bit in the logic of 4638 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 4639 kludge to work properly if AES128 is available and AES256 isn't. 4640 [Victor Duchovni] 4641 4642 *) Expand security boundary to match 1.1.1 module. 4643 [Steve Henson] 4644 4645 *) Remove redundant features: hash file source, editing of test vectors 4646 modify fipsld to use external fips_premain.c signature. 4647 [Steve Henson] 4648 4649 *) New perl script mkfipsscr.pl to create shell scripts or batch files to 4650 run algorithm test programs. 4651 [Steve Henson] 4652 4653 *) Make algorithm test programs more tolerant of whitespace. 4654 [Steve Henson] 4655 4656 *) Have SSL/TLS server implementation tolerate "mismatched" record 4657 protocol version while receiving ClientHello even if the 4658 ClientHello is fragmented. (The server can't insist on the 4659 particular protocol version it has chosen before the ServerHello 4660 message has informed the client about his choice.) 4661 [Bodo Moeller] 4662 4663 *) Load error codes if they are not already present instead of using a 4664 static variable. This allows them to be cleanly unloaded and reloaded. 4665 [Steve Henson] 4666 4667 Changes between 0.9.7k and 0.9.7l [28 Sep 2006] 4668 4669 *) Introduce limits to prevent malicious keys being able to 4670 cause a denial of service. (CVE-2006-2940) 4671 [Steve Henson, Bodo Moeller] 4672 4673 *) Fix ASN.1 parsing of certain invalid structures that can result 4674 in a denial of service. (CVE-2006-2937) [Steve Henson] 4675 4676 *) Fix buffer overflow in SSL_get_shared_ciphers() function. 4677 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 4678 4679 *) Fix SSL client code which could crash if connecting to a 4680 malicious SSLv2 server. (CVE-2006-4343) 4681 [Tavis Ormandy and Will Drewry, Google Security Team] 4682 4683 *) Change ciphersuite string processing so that an explicit 4684 ciphersuite selects this one ciphersuite (so that "AES256-SHA" 4685 will no longer include "AES128-SHA"), and any other similar 4686 ciphersuite (same bitmap) from *other* protocol versions (so that 4687 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 4688 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 4689 changes from 0.9.8b and 0.9.8d. 4690 [Bodo Moeller] 4691 4692 Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 4693 4694 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 4695 (CVE-2006-4339) [Ben Laurie and Google Security Team] 4696 4697 *) Change the Unix randomness entropy gathering to use poll() when 4698 possible instead of select(), since the latter has some 4699 undesirable limitations. 4700 [Darryl Miles via Richard Levitte and Bodo Moeller] 4701 4702 *) Disable rogue ciphersuites: 4703 4704 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 4705 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 4706 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 4707 4708 The latter two were purportedly from 4709 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 4710 appear there. 4711 4712 Also deactive the remaining ciphersuites from 4713 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 4714 unofficial, and the ID has long expired. 4715 [Bodo Moeller] 4716 4717 *) Fix RSA blinding Heisenbug (problems sometimes occured on 4718 dual-core machines) and other potential thread-safety issues. 4719 [Bodo Moeller] 4720 4721 Changes between 0.9.7i and 0.9.7j [04 May 2006] 4722 4723 *) Adapt fipsld and the build system to link against the validated FIPS 4724 module in FIPS mode. 4725 [Steve Henson] 4726 4727 *) Fixes for VC++ 2005 build under Windows. 4728 [Steve Henson] 4729 4730 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 4731 from a Windows bash shell such as MSYS. It is autodetected from the 4732 "config" script when run from a VC++ environment. Modify standard VC++ 4733 build to use fipscanister.o from the GNU make build. 4734 [Steve Henson] 4735 4736 Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 4737 4738 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 4739 The value now differs depending on if you build for FIPS or not. 4740 BEWARE! A program linked with a shared FIPSed libcrypto can't be 4741 safely run with a non-FIPSed libcrypto, as it may crash because of 4742 the difference induced by this change. 4743 [Andy Polyakov] 4744 4745 Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 4746 4747 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 4748 (part of SSL_OP_ALL). This option used to disable the 4749 countermeasure against man-in-the-middle protocol-version 4750 rollback in the SSL 2.0 server implementation, which is a bad 4751 idea. (CVE-2005-2969) 4752 4753 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 4754 for Information Security, National Institute of Advanced Industrial 4755 Science and Technology [AIST], Japan)] 4756 4757 *) Minimal support for X9.31 signatures and PSS padding modes. This is 4758 mainly for FIPS compliance and not fully integrated at this stage. 4759 [Steve Henson] 4760 4761 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 4762 the exponentiation using a fixed-length exponent. (Otherwise, 4763 the information leaked through timing could expose the secret key 4764 after many signatures; cf. Bleichenbacher's attack on DSA with 4765 biased k.) 4766 [Bodo Moeller] 4767 4768 *) Make a new fixed-window mod_exp implementation the default for 4769 RSA, DSA, and DH private-key operations so that the sequence of 4770 squares and multiplies and the memory access pattern are 4771 independent of the particular secret key. This will mitigate 4772 cache-timing and potential related attacks. 4773 4774 BN_mod_exp_mont_consttime() is the new exponentiation implementation, 4775 and this is automatically used by BN_mod_exp_mont() if the new flag 4776 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 4777 will use this BN flag for private exponents unless the flag 4778 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 4779 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 4780 4781 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 4782 4783 *) Change the client implementation for SSLv23_method() and 4784 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 4785 Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 4786 (Previously, the SSL 2.0 backwards compatible Client Hello 4787 message format would be used even with SSL_OP_NO_SSLv2.) 4788 [Bodo Moeller] 4789 4790 *) Add support for smime-type MIME parameter in S/MIME messages which some 4791 clients need. 4792 [Steve Henson] 4793 4794 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 4795 a threadsafe manner. Modify rsa code to use new function and add calls 4796 to dsa and dh code (which had race conditions before). 4797 [Steve Henson] 4798 4799 *) Include the fixed error library code in the C error file definitions 4800 instead of fixing them up at runtime. This keeps the error code 4801 structures constant. 4802 [Steve Henson] 4803 4804 Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 4805 4806 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 4807 OpenSSL 0.9.8.] 4808 4809 *) Fixes for newer kerberos headers. NB: the casts are needed because 4810 the 'length' field is signed on one version and unsigned on another 4811 with no (?) obvious way to tell the difference, without these VC++ 4812 complains. Also the "definition" of FAR (blank) is no longer included 4813 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 4814 some needed definitions. 4815 [Steve Henson] 4816 4817 *) Undo Cygwin change. 4818 [Ulf M��ller] 4819 4820 *) Added support for proxy certificates according to RFC 3820. 4821 Because they may be a security thread to unaware applications, 4822 they must be explicitely allowed in run-time. See 4823 docs/HOWTO/proxy_certificates.txt for further information. 4824 [Richard Levitte] 4825 4826 Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 4827 4828 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 4829 server and client random values. Previously 4830 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 4831 less random data when sizeof(time_t) > 4 (some 64 bit platforms). 4832 4833 This change has negligible security impact because: 4834 4835 1. Server and client random values still have 24 bytes of pseudo random 4836 data. 4837 4838 2. Server and client random values are sent in the clear in the initial 4839 handshake. 4840 4841 3. The master secret is derived using the premaster secret (48 bytes in 4842 size for static RSA ciphersuites) as well as client server and random 4843 values. 4844 4845 The OpenSSL team would like to thank the UK NISCC for bringing this issue 4846 to our attention. 4847 4848 [Stephen Henson, reported by UK NISCC] 4849 4850 *) Use Windows randomness collection on Cygwin. 4851 [Ulf M��ller] 4852 4853 *) Fix hang in EGD/PRNGD query when communication socket is closed 4854 prematurely by EGD/PRNGD. 4855 [Darren Tucker <dtucker@zip.com.au> via Lutz J��nicke, resolves #1014] 4856 4857 *) Prompt for pass phrases when appropriate for PKCS12 input format. 4858 [Steve Henson] 4859 4860 *) Back-port of selected performance improvements from development 4861 branch, as well as improved support for PowerPC platforms. 4862 [Andy Polyakov] 4863 4864 *) Add lots of checks for memory allocation failure, error codes to indicate 4865 failure and freeing up memory if a failure occurs. 4866 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 4867 4868 *) Add new -passin argument to dgst. 4869 [Steve Henson] 4870 4871 *) Perform some character comparisons of different types in X509_NAME_cmp: 4872 this is needed for some certificates that reencode DNs into UTF8Strings 4873 (in violation of RFC3280) and can't or wont issue name rollover 4874 certificates. 4875 [Steve Henson] 4876 4877 *) Make an explicit check during certificate validation to see that 4878 the CA setting in each certificate on the chain is correct. As a 4879 side effect always do the following basic checks on extensions, 4880 not just when there's an associated purpose to the check: 4881 4882 - if there is an unhandled critical extension (unless the user 4883 has chosen to ignore this fault) 4884 - if the path length has been exceeded (if one is set at all) 4885 - that certain extensions fit the associated purpose (if one has 4886 been given) 4887 [Richard Levitte] 4888 4889 Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 4890 4891 *) Avoid a race condition when CRLs are checked in a multi threaded 4892 environment. This would happen due to the reordering of the revoked 4893 entries during signature checking and serial number lookup. Now the 4894 encoding is cached and the serial number sort performed under a lock. 4895 Add new STACK function sk_is_sorted(). 4896 [Steve Henson] 4897 4898 *) Add Delta CRL to the extension code. 4899 [Steve Henson] 4900 4901 *) Various fixes to s3_pkt.c so alerts are sent properly. 4902 [David Holmes <d.holmes@f5.com>] 4903 4904 *) Reduce the chances of duplicate issuer name and serial numbers (in 4905 violation of RFC3280) using the OpenSSL certificate creation utilities. 4906 This is done by creating a random 64 bit value for the initial serial 4907 number when a serial number file is created or when a self signed 4908 certificate is created using 'openssl req -x509'. The initial serial 4909 number file is created using 'openssl x509 -next_serial' in CA.pl 4910 rather than being initialized to 1. 4911 [Steve Henson] 4912 4913 Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 4914 4915 *) Fix null-pointer assignment in do_change_cipher_spec() revealed 4916 by using the Codenomicon TLS Test Tool (CVE-2004-0079) 4917 [Joe Orton, Steve Henson] 4918 4919 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 4920 (CVE-2004-0112) 4921 [Joe Orton, Steve Henson] 4922 4923 *) Make it possible to have multiple active certificates with the same 4924 subject in the CA index file. This is done only if the keyword 4925 'unique_subject' is set to 'no' in the main CA section (default 4926 if 'CA_default') of the configuration file. The value is saved 4927 with the database itself in a separate index attribute file, 4928 named like the index file with '.attr' appended to the name. 4929 [Richard Levitte] 4930 4931 *) X509 verify fixes. Disable broken certificate workarounds when 4932 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 4933 keyUsage extension present. Don't accept CRLs with unhandled critical 4934 extensions: since verify currently doesn't process CRL extensions this 4935 rejects a CRL with *any* critical extensions. Add new verify error codes 4936 for these cases. 4937 [Steve Henson] 4938 4939 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 4940 A clarification of RFC2560 will require the use of OCTET STRINGs and 4941 some implementations cannot handle the current raw format. Since OpenSSL 4942 copies and compares OCSP nonces as opaque blobs without any attempt at 4943 parsing them this should not create any compatibility issues. 4944 [Steve Henson] 4945 4946 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 4947 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 4948 this HMAC (and other) operations are several times slower than OpenSSL 4949 < 0.9.7. 4950 [Steve Henson] 4951 4952 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 4953 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 4954 4955 *) Use the correct content when signing type "other". 4956 [Steve Henson] 4957 4958 Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 4959 4960 *) Fix various bugs revealed by running the NISCC test suite: 4961 4962 Stop out of bounds reads in the ASN1 code when presented with 4963 invalid tags (CVE-2003-0543 and CVE-2003-0544). 4964 4965 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 4966 4967 If verify callback ignores invalid public key errors don't try to check 4968 certificate signature with the NULL public key. 4969 4970 [Steve Henson] 4971 4972 *) New -ignore_err option in ocsp application to stop the server 4973 exiting on the first error in a request. 4974 [Steve Henson] 4975 4976 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 4977 if the server requested one: as stated in TLS 1.0 and SSL 3.0 4978 specifications. 4979 [Steve Henson] 4980 4981 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 4982 extra data after the compression methods not only for TLS 1.0 4983 but also for SSL 3.0 (as required by the specification). 4984 [Bodo Moeller; problem pointed out by Matthias Loepfe] 4985 4986 *) Change X509_certificate_type() to mark the key as exported/exportable 4987 when it's 512 *bits* long, not 512 bytes. 4988 [Richard Levitte] 4989 4990 *) Change AES_cbc_encrypt() so it outputs exact multiple of 4991 blocks during encryption. 4992 [Richard Levitte] 4993 4994 *) Various fixes to base64 BIO and non blocking I/O. On write 4995 flushes were not handled properly if the BIO retried. On read 4996 data was not being buffered properly and had various logic bugs. 4997 This also affects blocking I/O when the data being decoded is a 4998 certain size. 4999 [Steve Henson] 5000 5001 *) Various S/MIME bugfixes and compatibility changes: 5002 output correct application/pkcs7 MIME type if 5003 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 5004 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 5005 of files as .eml work). Correctly handle very long lines in MIME 5006 parser. 5007 [Steve Henson] 5008 5009 Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 5010 5011 *) Countermeasure against the Klima-Pokorny-Rosa extension of 5012 Bleichbacher's attack on PKCS #1 v1.5 padding: treat 5013 a protocol version number mismatch like a decryption error 5014 in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 5015 [Bodo Moeller] 5016 5017 *) Turn on RSA blinding by default in the default implementation 5018 to avoid a timing attack. Applications that don't want it can call 5019 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 5020 They would be ill-advised to do so in most cases. 5021 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 5022 5023 *) Change RSA blinding code so that it works when the PRNG is not 5024 seeded (in this case, the secret RSA exponent is abused as 5025 an unpredictable seed -- if it is not unpredictable, there 5026 is no point in blinding anyway). Make RSA blinding thread-safe 5027 by remembering the creator's thread ID in rsa->blinding and 5028 having all other threads use local one-time blinding factors 5029 (this requires more computation than sharing rsa->blinding, but 5030 avoids excessive locking; and if an RSA object is not shared 5031 between threads, blinding will still be very fast). 5032 [Bodo Moeller] 5033 5034 *) Fixed a typo bug that would cause ENGINE_set_default() to set an 5035 ENGINE as defaults for all supported algorithms irrespective of 5036 the 'flags' parameter. 'flags' is now honoured, so applications 5037 should make sure they are passing it correctly. 5038 [Geoff Thorpe] 5039 5040 *) Target "mingw" now allows native Windows code to be generated in 5041 the Cygwin environment as well as with the MinGW compiler. 5042 [Ulf Moeller] 5043 5044 Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 5045 5046 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 5047 via timing by performing a MAC computation even if incorrrect 5048 block cipher padding has been found. This is a countermeasure 5049 against active attacks where the attacker has to distinguish 5050 between bad padding and a MAC verification error. (CVE-2003-0078) 5051 5052 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 5053 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 5054 Martin Vuagnoux (EPFL, Ilion)] 5055 5056 *) Make the no-err option work as intended. The intention with no-err 5057 is not to have the whole error stack handling routines removed from 5058 libcrypto, it's only intended to remove all the function name and 5059 reason texts, thereby removing some of the footprint that may not 5060 be interesting if those errors aren't displayed anyway. 5061 5062 NOTE: it's still possible for any application or module to have it's 5063 own set of error texts inserted. The routines are there, just not 5064 used by default when no-err is given. 5065 [Richard Levitte] 5066 5067 *) Add support for FreeBSD on IA64. 5068 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 5069 5070 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 5071 Kerberos function mit_des_cbc_cksum(). Before this change, 5072 the value returned by DES_cbc_cksum() was like the one from 5073 mit_des_cbc_cksum(), except the bytes were swapped. 5074 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 5075 5076 *) Allow an application to disable the automatic SSL chain building. 5077 Before this a rather primitive chain build was always performed in 5078 ssl3_output_cert_chain(): an application had no way to send the 5079 correct chain if the automatic operation produced an incorrect result. 5080 5081 Now the chain builder is disabled if either: 5082 5083 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 5084 5085 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 5086 5087 The reasoning behind this is that an application would not want the 5088 auto chain building to take place if extra chain certificates are 5089 present and it might also want a means of sending no additional 5090 certificates (for example the chain has two certificates and the 5091 root is omitted). 5092 [Steve Henson] 5093 5094 *) Add the possibility to build without the ENGINE framework. 5095 [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 5096 5097 *) Under Win32 gmtime() can return NULL: check return value in 5098 OPENSSL_gmtime(). Add error code for case where gmtime() fails. 5099 [Steve Henson] 5100 5101 *) DSA routines: under certain error conditions uninitialized BN objects 5102 could be freed. Solution: make sure initialization is performed early 5103 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 5104 Nils Larsch <nla@trustcenter.de> via PR#459) 5105 [Lutz Jaenicke] 5106 5107 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 5108 checked on reconnect on the client side, therefore session resumption 5109 could still fail with a "ssl session id is different" error. This 5110 behaviour is masked when SSL_OP_ALL is used due to 5111 SSL_OP_MICROSOFT_SESS_ID_BUG being set. 5112 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 5113 followup to PR #377. 5114 [Lutz Jaenicke] 5115 5116 *) IA-32 assembler support enhancements: unified ELF targets, support 5117 for SCO/Caldera platforms, fix for Cygwin shared build. 5118 [Andy Polyakov] 5119 5120 *) Add support for FreeBSD on sparc64. As a consequence, support for 5121 FreeBSD on non-x86 processors is separate from x86 processors on 5122 the config script, much like the NetBSD support. 5123 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 5124 5125 Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 5126 5127 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 5128 OpenSSL 0.9.7.] 5129 5130 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 5131 code (06) was taken as the first octet of the session ID and the last 5132 octet was ignored consequently. As a result SSLv2 client side session 5133 caching could not have worked due to the session ID mismatch between 5134 client and server. 5135 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 5136 PR #377. 5137 [Lutz Jaenicke] 5138 5139 *) Change the declaration of needed Kerberos libraries to use EX_LIBS 5140 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 5141 removed entirely. 5142 [Richard Levitte] 5143 5144 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 5145 seems that in spite of existing for more than a year, many application 5146 author have done nothing to provide the necessary callbacks, which 5147 means that this particular engine will not work properly anywhere. 5148 This is a very unfortunate situation which forces us, in the name 5149 of usability, to give the hw_ncipher.c a static lock, which is part 5150 of libcrypto. 5151 NOTE: This is for the 0.9.7 series ONLY. This hack will never 5152 appear in 0.9.8 or later. We EXPECT application authors to have 5153 dealt properly with this when 0.9.8 is released (unless we actually 5154 make such changes in the libcrypto locking code that changes will 5155 have to be made anyway). 5156 [Richard Levitte] 5157 5158 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 5159 octets have been read, EOF or an error occurs. Without this change 5160 some truncated ASN1 structures will not produce an error. 5161 [Steve Henson] 5162 5163 *) Disable Heimdal support, since it hasn't been fully implemented. 5164 Still give the possibility to force the use of Heimdal, but with 5165 warnings and a request that patches get sent to openssl-dev. 5166 [Richard Levitte] 5167 5168 *) Add the VC-CE target, introduce the WINCE sysname, and add 5169 INSTALL.WCE and appropriate conditionals to make it build. 5170 [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 5171 5172 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 5173 cygssl-x.y.z.dll, where x, y and z are the major, minor and 5174 edit numbers of the version. 5175 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 5176 5177 *) Introduce safe string copy and catenation functions 5178 (BUF_strlcpy() and BUF_strlcat()). 5179 [Ben Laurie (CHATS) and Richard Levitte] 5180 5181 *) Avoid using fixed-size buffers for one-line DNs. 5182 [Ben Laurie (CHATS)] 5183 5184 *) Add BUF_MEM_grow_clean() to avoid information leakage when 5185 resizing buffers containing secrets, and use where appropriate. 5186 [Ben Laurie (CHATS)] 5187 5188 *) Avoid using fixed size buffers for configuration file location. 5189 [Ben Laurie (CHATS)] 5190 5191 *) Avoid filename truncation for various CA files. 5192 [Ben Laurie (CHATS)] 5193 5194 *) Use sizeof in preference to magic numbers. 5195 [Ben Laurie (CHATS)] 5196 5197 *) Avoid filename truncation in cert requests. 5198 [Ben Laurie (CHATS)] 5199 5200 *) Add assertions to check for (supposedly impossible) buffer 5201 overflows. 5202 [Ben Laurie (CHATS)] 5203 5204 *) Don't cache truncated DNS entries in the local cache (this could 5205 potentially lead to a spoofing attack). 5206 [Ben Laurie (CHATS)] 5207 5208 *) Fix various buffers to be large enough for hex/decimal 5209 representations in a platform independent manner. 5210 [Ben Laurie (CHATS)] 5211 5212 *) Add CRYPTO_realloc_clean() to avoid information leakage when 5213 resizing buffers containing secrets, and use where appropriate. 5214 [Ben Laurie (CHATS)] 5215 5216 *) Add BIO_indent() to avoid much slightly worrying code to do 5217 indents. 5218 [Ben Laurie (CHATS)] 5219 5220 *) Convert sprintf()/BIO_puts() to BIO_printf(). 5221 [Ben Laurie (CHATS)] 5222 5223 *) buffer_gets() could terminate with the buffer only half 5224 full. Fixed. 5225 [Ben Laurie (CHATS)] 5226 5227 *) Add assertions to prevent user-supplied crypto functions from 5228 overflowing internal buffers by having large block sizes, etc. 5229 [Ben Laurie (CHATS)] 5230 5231 *) New OPENSSL_assert() macro (similar to assert(), but enabled 5232 unconditionally). 5233 [Ben Laurie (CHATS)] 5234 5235 *) Eliminate unused copy of key in RC4. 5236 [Ben Laurie (CHATS)] 5237 5238 *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 5239 [Ben Laurie (CHATS)] 5240 5241 *) Fix off-by-one error in EGD path. 5242 [Ben Laurie (CHATS)] 5243 5244 *) If RANDFILE path is too long, ignore instead of truncating. 5245 [Ben Laurie (CHATS)] 5246 5247 *) Eliminate unused and incorrectly sized X.509 structure 5248 CBCParameter. 5249 [Ben Laurie (CHATS)] 5250 5251 *) Eliminate unused and dangerous function knumber(). 5252 [Ben Laurie (CHATS)] 5253 5254 *) Eliminate unused and dangerous structure, KSSL_ERR. 5255 [Ben Laurie (CHATS)] 5256 5257 *) Protect against overlong session ID context length in an encoded 5258 session object. Since these are local, this does not appear to be 5259 exploitable. 5260 [Ben Laurie (CHATS)] 5261 5262 *) Change from security patch (see 0.9.6e below) that did not affect 5263 the 0.9.6 release series: 5264 5265 Remote buffer overflow in SSL3 protocol - an attacker could 5266 supply an oversized master key in Kerberos-enabled versions. 5267 (CVE-2002-0657) 5268 [Ben Laurie (CHATS)] 5269 5270 *) Change the SSL kerb5 codes to match RFC 2712. 5271 [Richard Levitte] 5272 5273 *) Make -nameopt work fully for req and add -reqopt switch. 5274 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 5275 5276 *) The "block size" for block ciphers in CFB and OFB mode should be 1. 5277 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 5278 5279 *) Make sure tests can be performed even if the corresponding algorithms 5280 have been removed entirely. This was also the last step to make 5281 OpenSSL compilable with DJGPP under all reasonable conditions. 5282 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 5283 5284 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 5285 to allow version independent disabling of normally unselected ciphers, 5286 which may be activated as a side-effect of selecting a single cipher. 5287 5288 (E.g., cipher list string "RSA" enables ciphersuites that are left 5289 out of "ALL" because they do not provide symmetric encryption. 5290 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 5291 [Lutz Jaenicke, Bodo Moeller] 5292 5293 *) Add appropriate support for separate platform-dependent build 5294 directories. The recommended way to make a platform-dependent 5295 build directory is the following (tested on Linux), maybe with 5296 some local tweaks: 5297 5298 # Place yourself outside of the OpenSSL source tree. In 5299 # this example, the environment variable OPENSSL_SOURCE 5300 # is assumed to contain the absolute OpenSSL source directory. 5301 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 5302 cd objtree/"`uname -s`-`uname -r`-`uname -m`" 5303 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 5304 mkdir -p `dirname $F` 5305 ln -s $OPENSSL_SOURCE/$F $F 5306 done 5307 5308 To be absolutely sure not to disturb the source tree, a "make clean" 5309 is a good thing. If it isn't successfull, don't worry about it, 5310 it probably means the source directory is very clean. 5311 [Richard Levitte] 5312 5313 *) Make sure any ENGINE control commands make local copies of string 5314 pointers passed to them whenever necessary. Otherwise it is possible 5315 the caller may have overwritten (or deallocated) the original string 5316 data when a later ENGINE operation tries to use the stored values. 5317 [G��tz Babin-Ebell <babinebell@trustcenter.de>] 5318 5319 *) Improve diagnostics in file reading and command-line digests. 5320 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 5321 5322 *) Add AES modes CFB and OFB to the object database. Correct an 5323 error in AES-CFB decryption. 5324 [Richard Levitte] 5325 5326 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 5327 allows existing EVP_CIPHER_CTX structures to be reused after 5328 calling EVP_*Final(). This behaviour is used by encryption 5329 BIOs and some applications. This has the side effect that 5330 applications must explicitly clean up cipher contexts with 5331 EVP_CIPHER_CTX_cleanup() or they will leak memory. 5332 [Steve Henson] 5333 5334 *) Check the values of dna and dnb in bn_mul_recursive before calling 5335 bn_mul_comba (a non zero value means the a or b arrays do not contain 5336 n2 elements) and fallback to bn_mul_normal if either is not zero. 5337 [Steve Henson] 5338 5339 *) Fix escaping of non-ASCII characters when using the -subj option 5340 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 5341 [Lutz Jaenicke] 5342 5343 *) Make object definitions compliant to LDAP (RFC2256): SN is the short 5344 form for "surname", serialNumber has no short form. 5345 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 5346 therefore remove "mail" short name for "internet 7". 5347 The OID for unique identifiers in X509 certificates is 5348 x500UniqueIdentifier, not uniqueIdentifier. 5349 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 5350 [Lutz Jaenicke] 5351 5352 *) Add an "init" command to the ENGINE config module and auto initialize 5353 ENGINEs. Without any "init" command the ENGINE will be initialized 5354 after all ctrl commands have been executed on it. If init=1 the 5355 ENGINE is initailized at that point (ctrls before that point are run 5356 on the uninitialized ENGINE and after on the initialized one). If 5357 init=0 then the ENGINE will not be iniatialized at all. 5358 [Steve Henson] 5359 5360 *) Fix the 'app_verify_callback' interface so that the user-defined 5361 argument is actually passed to the callback: In the 5362 SSL_CTX_set_cert_verify_callback() prototype, the callback 5363 declaration has been changed from 5364 int (*cb)() 5365 into 5366 int (*cb)(X509_STORE_CTX *,void *); 5367 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 5368 i=s->ctx->app_verify_callback(&ctx) 5369 has been changed into 5370 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 5371 5372 To update applications using SSL_CTX_set_cert_verify_callback(), 5373 a dummy argument can be added to their callback functions. 5374 [D. K. Smetters <smetters@parc.xerox.com>] 5375 5376 *) Added the '4758cca' ENGINE to support IBM 4758 cards. 5377 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 5378 5379 *) Add and OPENSSL_LOAD_CONF define which will cause 5380 OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 5381 This allows older applications to transparently support certain 5382 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 5383 Two new functions OPENSSL_add_all_algorithms_noconf() which will never 5384 load the config file and OPENSSL_add_all_algorithms_conf() which will 5385 always load it have also been added. 5386 [Steve Henson] 5387 5388 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 5389 Adjust NIDs and EVP layer. 5390 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 5391 5392 *) Config modules support in openssl utility. 5393 5394 Most commands now load modules from the config file, 5395 though in a few (such as version) this isn't done 5396 because it couldn't be used for anything. 5397 5398 In the case of ca and req the config file used is 5399 the same as the utility itself: that is the -config 5400 command line option can be used to specify an 5401 alternative file. 5402 [Steve Henson] 5403 5404 *) Move default behaviour from OPENSSL_config(). If appname is NULL 5405 use "openssl_conf" if filename is NULL use default openssl config file. 5406 [Steve Henson] 5407 5408 *) Add an argument to OPENSSL_config() to allow the use of an alternative 5409 config section name. Add a new flag to tolerate a missing config file 5410 and move code to CONF_modules_load_file(). 5411 [Steve Henson] 5412 5413 *) Support for crypto accelerator cards from Accelerated Encryption 5414 Processing, www.aep.ie. (Use engine 'aep') 5415 The support was copied from 0.9.6c [engine] and adapted/corrected 5416 to work with the new engine framework. 5417 [AEP Inc. and Richard Levitte] 5418 5419 *) Support for SureWare crypto accelerator cards from Baltimore 5420 Technologies. (Use engine 'sureware') 5421 The support was copied from 0.9.6c [engine] and adapted 5422 to work with the new engine framework. 5423 [Richard Levitte] 5424 5425 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 5426 make the newer ENGINE framework commands for the CHIL engine work. 5427 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 5428 5429 *) Make it possible to produce shared libraries on ReliantUNIX. 5430 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 5431 5432 *) Add the configuration target debug-linux-ppro. 5433 Make 'openssl rsa' use the general key loading routines 5434 implemented in apps.c, and make those routines able to 5435 handle the key format FORMAT_NETSCAPE and the variant 5436 FORMAT_IISSGC. 5437 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 5438 5439 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 5440 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 5441 5442 *) Add -keyform to rsautl, and document -engine. 5443 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 5444 5445 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 5446 BIO_R_NO_SUCH_FILE error code rather than the generic 5447 ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 5448 [Ben Laurie] 5449 5450 *) Add new functions 5451 ERR_peek_last_error 5452 ERR_peek_last_error_line 5453 ERR_peek_last_error_line_data. 5454 These are similar to 5455 ERR_peek_error 5456 ERR_peek_error_line 5457 ERR_peek_error_line_data, 5458 but report on the latest error recorded rather than the first one 5459 still in the error queue. 5460 [Ben Laurie, Bodo Moeller] 5461 5462 *) default_algorithms option in ENGINE config module. This allows things 5463 like: 5464 default_algorithms = ALL 5465 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 5466 [Steve Henson] 5467 5468 *) Prelminary ENGINE config module. 5469 [Steve Henson] 5470 5471 *) New experimental application configuration code. 5472 [Steve Henson] 5473 5474 *) Change the AES code to follow the same name structure as all other 5475 symmetric ciphers, and behave the same way. Move everything to 5476 the directory crypto/aes, thereby obsoleting crypto/rijndael. 5477 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 5478 5479 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 5480 [Ben Laurie and Theo de Raadt] 5481 5482 *) Add option to output public keys in req command. 5483 [Massimiliano Pala madwolf@openca.org] 5484 5485 *) Use wNAFs in EC_POINTs_mul() for improved efficiency 5486 (up to about 10% better than before for P-192 and P-224). 5487 [Bodo Moeller] 5488 5489 *) New functions/macros 5490 5491 SSL_CTX_set_msg_callback(ctx, cb) 5492 SSL_CTX_set_msg_callback_arg(ctx, arg) 5493 SSL_set_msg_callback(ssl, cb) 5494 SSL_set_msg_callback_arg(ssl, arg) 5495 5496 to request calling a callback function 5497 5498 void cb(int write_p, int version, int content_type, 5499 const void *buf, size_t len, SSL *ssl, void *arg) 5500 5501 whenever a protocol message has been completely received 5502 (write_p == 0) or sent (write_p == 1). Here 'version' is the 5503 protocol version according to which the SSL library interprets 5504 the current protocol message (SSL2_VERSION, SSL3_VERSION, or 5505 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 5506 the content type as defined in the SSL 3.0/TLS 1.0 protocol 5507 specification (change_cipher_spec(20), alert(21), handshake(22)). 5508 'buf' and 'len' point to the actual message, 'ssl' to the 5509 SSL object, and 'arg' is the application-defined value set by 5510 SSL[_CTX]_set_msg_callback_arg(). 5511 5512 'openssl s_client' and 'openssl s_server' have new '-msg' options 5513 to enable a callback that displays all protocol messages. 5514 [Bodo Moeller] 5515 5516 *) Change the shared library support so shared libraries are built as 5517 soon as the corresponding static library is finished, and thereby get 5518 openssl and the test programs linked against the shared library. 5519 This still only happens when the keyword "shard" has been given to 5520 the configuration scripts. 5521 5522 NOTE: shared library support is still an experimental thing, and 5523 backward binary compatibility is still not guaranteed. 5524 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 5525 5526 *) Add support for Subject Information Access extension. 5527 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 5528 5529 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 5530 additional bytes when new memory had to be allocated, not just 5531 when reusing an existing buffer. 5532 [Bodo Moeller] 5533 5534 *) New command line and configuration option 'utf8' for the req command. 5535 This allows field values to be specified as UTF8 strings. 5536 [Steve Henson] 5537 5538 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 5539 runs for the former and machine-readable output for the latter. 5540 [Ben Laurie] 5541 5542 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 5543 of the e-mail address in the DN (i.e., it will go into a certificate 5544 extension only). The new configuration file option 'email_in_dn = no' 5545 has the same effect. 5546 [Massimiliano Pala madwolf@openca.org] 5547 5548 *) Change all functions with names starting with des_ to be starting 5549 with DES_ instead. Add wrappers that are compatible with libdes, 5550 but are named _ossl_old_des_*. Finally, add macros that map the 5551 des_* symbols to the corresponding _ossl_old_des_* if libdes 5552 compatibility is desired. If OpenSSL 0.9.6c compatibility is 5553 desired, the des_* symbols will be mapped to DES_*, with one 5554 exception. 5555 5556 Since we provide two compatibility mappings, the user needs to 5557 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 5558 compatibility is desired. The default (i.e., when that macro 5559 isn't defined) is OpenSSL 0.9.6c compatibility. 5560 5561 There are also macros that enable and disable the support of old 5562 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 5563 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 5564 are defined, the default will apply: to support the old des routines. 5565 5566 In either case, one must include openssl/des.h to get the correct 5567 definitions. Do not try to just include openssl/des_old.h, that 5568 won't work. 5569 5570 NOTE: This is a major break of an old API into a new one. Software 5571 authors are encouraged to switch to the DES_ style functions. Some 5572 time in the future, des_old.h and the libdes compatibility functions 5573 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 5574 default), and then completely removed. 5575 [Richard Levitte] 5576 5577 *) Test for certificates which contain unsupported critical extensions. 5578 If such a certificate is found during a verify operation it is 5579 rejected by default: this behaviour can be overridden by either 5580 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 5581 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 5582 X509_supported_extension() has also been added which returns 1 if a 5583 particular extension is supported. 5584 [Steve Henson] 5585 5586 *) Modify the behaviour of EVP cipher functions in similar way to digests 5587 to retain compatibility with existing code. 5588 [Steve Henson] 5589 5590 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 5591 compatibility with existing code. In particular the 'ctx' parameter does 5592 not have to be to be initialized before the call to EVP_DigestInit() and 5593 it is tidied up after a call to EVP_DigestFinal(). New function 5594 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 5595 EVP_MD_CTX_copy() changed to not require the destination to be 5596 initialized valid and new function EVP_MD_CTX_copy_ex() added which 5597 requires the destination to be valid. 5598 5599 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 5600 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 5601 [Steve Henson] 5602 5603 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 5604 so that complete 'Handshake' protocol structures are kept in memory 5605 instead of overwriting 'msg_type' and 'length' with 'body' data. 5606 [Bodo Moeller] 5607 5608 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 5609 [Massimo Santin via Richard Levitte] 5610 5611 *) Major restructuring to the underlying ENGINE code. This includes 5612 reduction of linker bloat, separation of pure "ENGINE" manipulation 5613 (initialisation, etc) from functionality dealing with implementations 5614 of specific crypto iterfaces. This change also introduces integrated 5615 support for symmetric ciphers and digest implementations - so ENGINEs 5616 can now accelerate these by providing EVP_CIPHER and EVP_MD 5617 implementations of their own. This is detailed in crypto/engine/README 5618 as it couldn't be adequately described here. However, there are a few 5619 API changes worth noting - some RSA, DSA, DH, and RAND functions that 5620 were changed in the original introduction of ENGINE code have now 5621 reverted back - the hooking from this code to ENGINE is now a good 5622 deal more passive and at run-time, operations deal directly with 5623 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 5624 dereferencing through an ENGINE pointer any more. Also, the ENGINE 5625 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 5626 they were not being used by the framework as there is no concept of a 5627 BIGNUM_METHOD and they could not be generalised to the new 5628 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 5629 ENGINE_cpy() has been removed as it cannot be consistently defined in 5630 the new code. 5631 [Geoff Thorpe] 5632 5633 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 5634 [Steve Henson] 5635 5636 *) Change mkdef.pl to sort symbols that get the same entry number, 5637 and make sure the automatically generated functions ERR_load_* 5638 become part of libeay.num as well. 5639 [Richard Levitte] 5640 5641 *) New function SSL_renegotiate_pending(). This returns true once 5642 renegotiation has been requested (either SSL_renegotiate() call 5643 or HelloRequest/ClientHello receveived from the peer) and becomes 5644 false once a handshake has been completed. 5645 (For servers, SSL_renegotiate() followed by SSL_do_handshake() 5646 sends a HelloRequest, but does not ensure that a handshake takes 5647 place. SSL_renegotiate_pending() is useful for checking if the 5648 client has followed the request.) 5649 [Bodo Moeller] 5650 5651 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 5652 By default, clients may request session resumption even during 5653 renegotiation (if session ID contexts permit); with this option, 5654 session resumption is possible only in the first handshake. 5655 5656 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 5657 more bits available for options that should not be part of 5658 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 5659 [Bodo Moeller] 5660 5661 *) Add some demos for certificate and certificate request creation. 5662 [Steve Henson] 5663 5664 *) Make maximum certificate chain size accepted from the peer application 5665 settable (SSL*_get/set_max_cert_list()), as proposed by 5666 "Douglas E. Engert" <deengert@anl.gov>. 5667 [Lutz Jaenicke] 5668 5669 *) Add support for shared libraries for Unixware-7 5670 (Boyd Lynn Gerber <gerberb@zenez.com>). 5671 [Lutz Jaenicke] 5672 5673 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 5674 be done prior to destruction. Use this to unload error strings from 5675 ENGINEs that load their own error strings. NB: This adds two new API 5676 functions to "get" and "set" this destroy handler in an ENGINE. 5677 [Geoff Thorpe] 5678 5679 *) Alter all existing ENGINE implementations (except "openssl" and 5680 "openbsd") to dynamically instantiate their own error strings. This 5681 makes them more flexible to be built both as statically-linked ENGINEs 5682 and self-contained shared-libraries loadable via the "dynamic" ENGINE. 5683 Also, add stub code to each that makes building them as self-contained 5684 shared-libraries easier (see README.ENGINE). 5685 [Geoff Thorpe] 5686 5687 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 5688 implementations into applications that are completely implemented in 5689 self-contained shared-libraries. The "dynamic" ENGINE exposes control 5690 commands that can be used to configure what shared-library to load and 5691 to control aspects of the way it is handled. Also, made an update to 5692 the README.ENGINE file that brings its information up-to-date and 5693 provides some information and instructions on the "dynamic" ENGINE 5694 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 5695 [Geoff Thorpe] 5696 5697 *) Make it possible to unload ranges of ERR strings with a new 5698 "ERR_unload_strings" function. 5699 [Geoff Thorpe] 5700 5701 *) Add a copy() function to EVP_MD. 5702 [Ben Laurie] 5703 5704 *) Make EVP_MD routines take a context pointer instead of just the 5705 md_data void pointer. 5706 [Ben Laurie] 5707 5708 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 5709 that the digest can only process a single chunk of data 5710 (typically because it is provided by a piece of 5711 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 5712 is only going to provide a single chunk of data, and hence the 5713 framework needn't accumulate the data for oneshot drivers. 5714 [Ben Laurie] 5715 5716 *) As with "ERR", make it possible to replace the underlying "ex_data" 5717 functions. This change also alters the storage and management of global 5718 ex_data state - it's now all inside ex_data.c and all "class" code (eg. 5719 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 5720 index counters. The API functions that use this state have been changed 5721 to take a "class_index" rather than pointers to the class's local STACK 5722 and counter, and there is now an API function to dynamically create new 5723 classes. This centralisation allows us to (a) plug a lot of the 5724 thread-safety problems that existed, and (b) makes it possible to clean 5725 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 5726 such data would previously have always leaked in application code and 5727 workarounds were in place to make the memory debugging turn a blind eye 5728 to it. Application code that doesn't use this new function will still 5729 leak as before, but their memory debugging output will announce it now 5730 rather than letting it slide. 5731 5732 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 5733 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 5734 has a return value to indicate success or failure. 5735 [Geoff Thorpe] 5736 5737 *) Make it possible to replace the underlying "ERR" functions such that the 5738 global state (2 LHASH tables and 2 locks) is only used by the "default" 5739 implementation. This change also adds two functions to "get" and "set" 5740 the implementation prior to it being automatically set the first time 5741 any other ERR function takes place. Ie. an application can call "get", 5742 pass the return value to a module it has just loaded, and that module 5743 can call its own "set" function using that value. This means the 5744 module's "ERR" operations will use (and modify) the error state in the 5745 application and not in its own statically linked copy of OpenSSL code. 5746 [Geoff Thorpe] 5747 5748 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 5749 reference counts. This performs normal REF_PRINT/REF_CHECK macros on 5750 the operation, and provides a more encapsulated way for external code 5751 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 5752 to use these functions rather than manually incrementing the counts. 5753 5754 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 5755 [Geoff Thorpe] 5756 5757 *) Add EVP test program. 5758 [Ben Laurie] 5759 5760 *) Add symmetric cipher support to ENGINE. Expect the API to change! 5761 [Ben Laurie] 5762 5763 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 5764 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 5765 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 5766 These allow a CRL to be built without having to access X509_CRL fields 5767 directly. Modify 'ca' application to use new functions. 5768 [Steve Henson] 5769 5770 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 5771 bug workarounds. Rollback attack detection is a security feature. 5772 The problem will only arise on OpenSSL servers when TLSv1 is not 5773 available (sslv3_server_method() or SSL_OP_NO_TLSv1). 5774 Software authors not wanting to support TLSv1 will have special reasons 5775 for their choice and can explicitly enable this option. 5776 [Bodo Moeller, Lutz Jaenicke] 5777 5778 *) Rationalise EVP so it can be extended: don't include a union of 5779 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 5780 (similar to those existing for EVP_CIPHER_CTX). 5781 Usage example: 5782 5783 EVP_MD_CTX md; 5784 5785 EVP_MD_CTX_init(&md); /* new function call */ 5786 EVP_DigestInit(&md, EVP_sha1()); 5787 EVP_DigestUpdate(&md, in, len); 5788 EVP_DigestFinal(&md, out, NULL); 5789 EVP_MD_CTX_cleanup(&md); /* new function call */ 5790 5791 [Ben Laurie] 5792 5793 *) Make DES key schedule conform to the usual scheme, as well as 5794 correcting its structure. This means that calls to DES functions 5795 now have to pass a pointer to a des_key_schedule instead of a 5796 plain des_key_schedule (which was actually always a pointer 5797 anyway): E.g., 5798 5799 des_key_schedule ks; 5800 5801 des_set_key_checked(..., &ks); 5802 des_ncbc_encrypt(..., &ks, ...); 5803 5804 (Note that a later change renames 'des_...' into 'DES_...'.) 5805 [Ben Laurie] 5806 5807 *) Initial reduction of linker bloat: the use of some functions, such as 5808 PEM causes large amounts of unused functions to be linked in due to 5809 poor organisation. For example pem_all.c contains every PEM function 5810 which has a knock on effect of linking in large amounts of (unused) 5811 ASN1 code. Grouping together similar functions and splitting unrelated 5812 functions prevents this. 5813 [Steve Henson] 5814 5815 *) Cleanup of EVP macros. 5816 [Ben Laurie] 5817 5818 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 5819 correct _ecb suffix. 5820 [Ben Laurie] 5821 5822 *) Add initial OCSP responder support to ocsp application. The 5823 revocation information is handled using the text based index 5824 use by the ca application. The responder can either handle 5825 requests generated internally, supplied in files (for example 5826 via a CGI script) or using an internal minimal server. 5827 [Steve Henson] 5828 5829 *) Add configuration choices to get zlib compression for TLS. 5830 [Richard Levitte] 5831 5832 *) Changes to Kerberos SSL for RFC 2712 compliance: 5833 1. Implemented real KerberosWrapper, instead of just using 5834 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 5835 2. Implemented optional authenticator field of KerberosWrapper. 5836 5837 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 5838 and authenticator structs; see crypto/krb5/. 5839 5840 Generalized Kerberos calls to support multiple Kerberos libraries. 5841 [Vern Staats <staatsvr@asc.hpc.mil>, 5842 Jeffrey Altman <jaltman@columbia.edu> 5843 via Richard Levitte] 5844 5845 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 5846 already does with RSA. testdsa.h now has 'priv_key/pub_key' 5847 values for each of the key sizes rather than having just 5848 parameters (and 'speed' generating keys each time). 5849 [Geoff Thorpe] 5850 5851 *) Speed up EVP routines. 5852 Before: 5853encrypt 5854type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 5855des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 5856des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 5857des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 5858decrypt 5859des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 5860des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 5861des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 5862 After: 5863encrypt 5864des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 5865decrypt 5866des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 5867 [Ben Laurie] 5868 5869 *) Added the OS2-EMX target. 5870 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 5871 5872 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 5873 to support NCONF routines in extension code. New function CONF_set_nconf() 5874 to allow functions which take an NCONF to also handle the old LHASH 5875 structure: this means that the old CONF compatible routines can be 5876 retained (in particular wrt extensions) without having to duplicate the 5877 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 5878 [Steve Henson] 5879 5880 *) Enhance the general user interface with mechanisms for inner control 5881 and with possibilities to have yes/no kind of prompts. 5882 [Richard Levitte] 5883 5884 *) Change all calls to low level digest routines in the library and 5885 applications to use EVP. Add missing calls to HMAC_cleanup() and 5886 don't assume HMAC_CTX can be copied using memcpy(). 5887 [Verdon Walker <VWalker@novell.com>, Steve Henson] 5888 5889 *) Add the possibility to control engines through control names but with 5890 arbitrary arguments instead of just a string. 5891 Change the key loaders to take a UI_METHOD instead of a callback 5892 function pointer. NOTE: this breaks binary compatibility with earlier 5893 versions of OpenSSL [engine]. 5894 Adapt the nCipher code for these new conditions and add a card insertion 5895 callback. 5896 [Richard Levitte] 5897 5898 *) Enhance the general user interface with mechanisms to better support 5899 dialog box interfaces, application-defined prompts, the possibility 5900 to use defaults (for example default passwords from somewhere else) 5901 and interrupts/cancellations. 5902 [Richard Levitte] 5903 5904 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 5905 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 5906 [Steve Henson] 5907 5908 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 5909 tidy up some unnecessarily weird code in 'sk_new()'). 5910 [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 5911 5912 *) Change the key loading routines for ENGINEs to use the same kind 5913 callback (pem_password_cb) as all other routines that need this 5914 kind of callback. 5915 [Richard Levitte] 5916 5917 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 5918 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 5919 than this minimum value is recommended. 5920 [Lutz Jaenicke] 5921 5922 *) New random seeder for OpenVMS, using the system process statistics 5923 that are easily reachable. 5924 [Richard Levitte] 5925 5926 *) Windows apparently can't transparently handle global 5927 variables defined in DLLs. Initialisations such as: 5928 5929 const ASN1_ITEM *it = &ASN1_INTEGER_it; 5930 5931 wont compile. This is used by the any applications that need to 5932 declare their own ASN1 modules. This was fixed by adding the option 5933 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 5934 needed for static libraries under Win32. 5935 [Steve Henson] 5936 5937 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 5938 setting of purpose and trust fields. New X509_STORE trust and 5939 purpose functions and tidy up setting in other SSL functions. 5940 [Steve Henson] 5941 5942 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 5943 structure. These are inherited by X509_STORE_CTX when it is 5944 initialised. This allows various defaults to be set in the 5945 X509_STORE structure (such as flags for CRL checking and custom 5946 purpose or trust settings) for functions which only use X509_STORE_CTX 5947 internally such as S/MIME. 5948 5949 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 5950 trust settings if they are not set in X509_STORE. This allows X509_STORE 5951 purposes and trust (in S/MIME for example) to override any set by default. 5952 5953 Add command line options for CRL checking to smime, s_client and s_server 5954 applications. 5955 [Steve Henson] 5956 5957 *) Initial CRL based revocation checking. If the CRL checking flag(s) 5958 are set then the CRL is looked up in the X509_STORE structure and 5959 its validity and signature checked, then if the certificate is found 5960 in the CRL the verify fails with a revoked error. 5961 5962 Various new CRL related callbacks added to X509_STORE_CTX structure. 5963 5964 Command line options added to 'verify' application to support this. 5965 5966 This needs some additional work, such as being able to handle multiple 5967 CRLs with different times, extension based lookup (rather than just 5968 by subject name) and ultimately more complete V2 CRL extension 5969 handling. 5970 [Steve Henson] 5971 5972 *) Add a general user interface API (crypto/ui/). This is designed 5973 to replace things like des_read_password and friends (backward 5974 compatibility functions using this new API are provided). 5975 The purpose is to remove prompting functions from the DES code 5976 section as well as provide for prompting through dialog boxes in 5977 a window system and the like. 5978 [Richard Levitte] 5979 5980 *) Add "ex_data" support to ENGINE so implementations can add state at a 5981 per-structure level rather than having to store it globally. 5982 [Geoff] 5983 5984 *) Make it possible for ENGINE structures to be copied when retrieved by 5985 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 5986 This causes the "original" ENGINE structure to act like a template, 5987 analogous to the RSA vs. RSA_METHOD type of separation. Because of this 5988 operational state can be localised to each ENGINE structure, despite the 5989 fact they all share the same "methods". New ENGINE structures returned in 5990 this case have no functional references and the return value is the single 5991 structural reference. This matches the single structural reference returned 5992 by ENGINE_by_id() normally, when it is incremented on the pre-existing 5993 ENGINE structure. 5994 [Geoff] 5995 5996 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 5997 needs to match any other type at all we need to manually clear the 5998 tag cache. 5999 [Steve Henson] 6000 6001 *) Changes to the "openssl engine" utility to include; 6002 - verbosity levels ('-v', '-vv', and '-vvv') that provide information 6003 about an ENGINE's available control commands. 6004 - executing control commands from command line arguments using the 6005 '-pre' and '-post' switches. '-post' is only used if '-t' is 6006 specified and the ENGINE is successfully initialised. The syntax for 6007 the individual commands are colon-separated, for example; 6008 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 6009 [Geoff] 6010 6011 *) New dynamic control command support for ENGINEs. ENGINEs can now 6012 declare their own commands (numbers), names (strings), descriptions, 6013 and input types for run-time discovery by calling applications. A 6014 subset of these commands are implicitly classed as "executable" 6015 depending on their input type, and only these can be invoked through 6016 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 6017 can be based on user input, config files, etc). The distinction is 6018 that "executable" commands cannot return anything other than a boolean 6019 result and can only support numeric or string input, whereas some 6020 discoverable commands may only be for direct use through 6021 ENGINE_ctrl(), eg. supporting the exchange of binary data, function 6022 pointers, or other custom uses. The "executable" commands are to 6023 support parameterisations of ENGINE behaviour that can be 6024 unambiguously defined by ENGINEs and used consistently across any 6025 OpenSSL-based application. Commands have been added to all the 6026 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 6027 control over shared-library paths without source code alterations. 6028 [Geoff] 6029 6030 *) Changed all ENGINE implementations to dynamically allocate their 6031 ENGINEs rather than declaring them statically. Apart from this being 6032 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 6033 this also allows the implementations to compile without using the 6034 internal engine_int.h header. 6035 [Geoff] 6036 6037 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 6038 'const' value. Any code that should be able to modify a RAND_METHOD 6039 should already have non-const pointers to it (ie. they should only 6040 modify their own ones). 6041 [Geoff] 6042 6043 *) Made a variety of little tweaks to the ENGINE code. 6044 - "atalla" and "ubsec" string definitions were moved from header files 6045 to C code. "nuron" string definitions were placed in variables 6046 rather than hard-coded - allowing parameterisation of these values 6047 later on via ctrl() commands. 6048 - Removed unused "#if 0"'d code. 6049 - Fixed engine list iteration code so it uses ENGINE_free() to release 6050 structural references. 6051 - Constified the RAND_METHOD element of ENGINE structures. 6052 - Constified various get/set functions as appropriate and added 6053 missing functions (including a catch-all ENGINE_cpy that duplicates 6054 all ENGINE values onto a new ENGINE except reference counts/state). 6055 - Removed NULL parameter checks in get/set functions. Setting a method 6056 or function to NULL is a way of cancelling out a previously set 6057 value. Passing a NULL ENGINE parameter is just plain stupid anyway 6058 and doesn't justify the extra error symbols and code. 6059 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 6060 flags from engine_int.h to engine.h. 6061 - Changed prototypes for ENGINE handler functions (init(), finish(), 6062 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 6063 [Geoff] 6064 6065 *) Implement binary inversion algorithm for BN_mod_inverse in addition 6066 to the algorithm using long division. The binary algorithm can be 6067 used only if the modulus is odd. On 32-bit systems, it is faster 6068 only for relatively small moduli (roughly 20-30% for 128-bit moduli, 6069 roughly 5-15% for 256-bit moduli), so we use it only for moduli 6070 up to 450 bits. In 64-bit environments, the binary algorithm 6071 appears to be advantageous for much longer moduli; here we use it 6072 for moduli up to 2048 bits. 6073 [Bodo Moeller] 6074 6075 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 6076 could not support the combine flag in choice fields. 6077 [Steve Henson] 6078 6079 *) Add a 'copy_extensions' option to the 'ca' utility. This copies 6080 extensions from a certificate request to the certificate. 6081 [Steve Henson] 6082 6083 *) Allow multiple 'certopt' and 'nameopt' options to be separated 6084 by commas. Add 'namopt' and 'certopt' options to the 'ca' config 6085 file: this allows the display of the certificate about to be 6086 signed to be customised, to allow certain fields to be included 6087 or excluded and extension details. The old system didn't display 6088 multicharacter strings properly, omitted fields not in the policy 6089 and couldn't display additional details such as extensions. 6090 [Steve Henson] 6091 6092 *) Function EC_POINTs_mul for multiple scalar multiplication 6093 of an arbitrary number of elliptic curve points 6094 \sum scalars[i]*points[i], 6095 optionally including the generator defined for the EC_GROUP: 6096 scalar*generator + \sum scalars[i]*points[i]. 6097 6098 EC_POINT_mul is a simple wrapper function for the typical case 6099 that the point list has just one item (besides the optional 6100 generator). 6101 [Bodo Moeller] 6102 6103 *) First EC_METHODs for curves over GF(p): 6104 6105 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 6106 operations and provides various method functions that can also 6107 operate with faster implementations of modular arithmetic. 6108 6109 EC_GFp_mont_method() reuses most functions that are part of 6110 EC_GFp_simple_method, but uses Montgomery arithmetic. 6111 6112 [Bodo Moeller; point addition and point doubling 6113 implementation directly derived from source code provided by 6114 Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 6115 6116 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 6117 crypto/ec/ec_lib.c): 6118 6119 Curves are EC_GROUP objects (with an optional group generator) 6120 based on EC_METHODs that are built into the library. 6121 6122 Points are EC_POINT objects based on EC_GROUP objects. 6123 6124 Most of the framework would be able to handle curves over arbitrary 6125 finite fields, but as there are no obvious types for fields other 6126 than GF(p), some functions are limited to that for now. 6127 [Bodo Moeller] 6128 6129 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 6130 that the file contains a complete HTTP response. 6131 [Richard Levitte] 6132 6133 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 6134 change the def and num file printf format specifier from "%-40sXXX" 6135 to "%-39s XXX". The latter will always guarantee a space after the 6136 field while the former will cause them to run together if the field 6137 is 40 of more characters long. 6138 [Steve Henson] 6139 6140 *) Constify the cipher and digest 'method' functions and structures 6141 and modify related functions to take constant EVP_MD and EVP_CIPHER 6142 pointers. 6143 [Steve Henson] 6144 6145 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 6146 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 6147 [Bodo Moeller] 6148 6149 *) Modify EVP_Digest*() routines so they now return values. Although the 6150 internal software routines can never fail additional hardware versions 6151 might. 6152 [Steve Henson] 6153 6154 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 6155 6156 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 6157 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 6158 6159 ASN1 error codes 6160 ERR_R_NESTED_ASN1_ERROR 6161 ... 6162 ERR_R_MISSING_ASN1_EOS 6163 were 4 .. 9, conflicting with 6164 ERR_LIB_RSA (= ERR_R_RSA_LIB) 6165 ... 6166 ERR_LIB_PEM (= ERR_R_PEM_LIB). 6167 They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 6168 6169 Add new error code 'ERR_R_INTERNAL_ERROR'. 6170 [Bodo Moeller] 6171 6172 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 6173 suffices. 6174 [Bodo Moeller] 6175 6176 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 6177 sets the subject name for a new request or supersedes the 6178 subject name in a given request. Formats that can be parsed are 6179 'CN=Some Name, OU=myOU, C=IT' 6180 and 6181 'CN=Some Name/OU=myOU/C=IT'. 6182 6183 Add options '-batch' and '-verbose' to 'openssl req'. 6184 [Massimiliano Pala <madwolf@hackmasters.net>] 6185 6186 *) Introduce the possibility to access global variables through 6187 functions on platform were that's the best way to handle exporting 6188 global variables in shared libraries. To enable this functionality, 6189 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 6190 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 6191 is normally done by Configure or something similar). 6192 6193 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 6194 in the source file (foo.c) like this: 6195 6196 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 6197 OPENSSL_IMPLEMENT_GLOBAL(double,bar); 6198 6199 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 6200 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 6201 6202 OPENSSL_DECLARE_GLOBAL(int,foo); 6203 #define foo OPENSSL_GLOBAL_REF(foo) 6204 OPENSSL_DECLARE_GLOBAL(double,bar); 6205 #define bar OPENSSL_GLOBAL_REF(bar) 6206 6207 The #defines are very important, and therefore so is including the 6208 header file everywhere where the defined globals are used. 6209 6210 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 6211 of ASN.1 items, but that structure is a bit different. 6212 6213 The largest change is in util/mkdef.pl which has been enhanced with 6214 better and easier to understand logic to choose which symbols should 6215 go into the Windows .def files as well as a number of fixes and code 6216 cleanup (among others, algorithm keywords are now sorted 6217 lexicographically to avoid constant rewrites). 6218 [Richard Levitte] 6219 6220 *) In BN_div() keep a copy of the sign of 'num' before writing the 6221 result to 'rm' because if rm==num the value will be overwritten 6222 and produce the wrong result if 'num' is negative: this caused 6223 problems with BN_mod() and BN_nnmod(). 6224 [Steve Henson] 6225 6226 *) Function OCSP_request_verify(). This checks the signature on an 6227 OCSP request and verifies the signer certificate. The signer 6228 certificate is just checked for a generic purpose and OCSP request 6229 trust settings. 6230 [Steve Henson] 6231 6232 *) Add OCSP_check_validity() function to check the validity of OCSP 6233 responses. OCSP responses are prepared in real time and may only 6234 be a few seconds old. Simply checking that the current time lies 6235 between thisUpdate and nextUpdate max reject otherwise valid responses 6236 caused by either OCSP responder or client clock inaccuracy. Instead 6237 we allow thisUpdate and nextUpdate to fall within a certain period of 6238 the current time. The age of the response can also optionally be 6239 checked. Two new options -validity_period and -status_age added to 6240 ocsp utility. 6241 [Steve Henson] 6242 6243 *) If signature or public key algorithm is unrecognized print out its 6244 OID rather that just UNKNOWN. 6245 [Steve Henson] 6246 6247 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 6248 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 6249 ID to be generated from the issuer certificate alone which can then be 6250 passed to OCSP_id_issuer_cmp(). 6251 [Steve Henson] 6252 6253 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 6254 ASN1 modules to export functions returning ASN1_ITEM pointers 6255 instead of the ASN1_ITEM structures themselves. This adds several 6256 new macros which allow the underlying ASN1 function/structure to 6257 be accessed transparently. As a result code should not use ASN1_ITEM 6258 references directly (such as &X509_it) but instead use the relevant 6259 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 6260 use of the new ASN1 code on platforms where exporting structures 6261 is problematical (for example in shared libraries) but exporting 6262 functions returning pointers to structures is not. 6263 [Steve Henson] 6264 6265 *) Add support for overriding the generation of SSL/TLS session IDs. 6266 These callbacks can be registered either in an SSL_CTX or per SSL. 6267 The purpose of this is to allow applications to control, if they wish, 6268 the arbitrary values chosen for use as session IDs, particularly as it 6269 can be useful for session caching in multiple-server environments. A 6270 command-line switch for testing this (and any client code that wishes 6271 to use such a feature) has been added to "s_server". 6272 [Geoff Thorpe, Lutz Jaenicke] 6273 6274 *) Modify mkdef.pl to recognise and parse preprocessor conditionals 6275 of the form '#if defined(...) || defined(...) || ...' and 6276 '#if !defined(...) && !defined(...) && ...'. This also avoids 6277 the growing number of special cases it was previously handling. 6278 [Richard Levitte] 6279 6280 *) Make all configuration macros available for application by making 6281 sure they are available in opensslconf.h, by giving them names starting 6282 with "OPENSSL_" to avoid conflicts with other packages and by making 6283 sure e_os2.h will cover all platform-specific cases together with 6284 opensslconf.h. 6285 Additionally, it is now possible to define configuration/platform- 6286 specific names (called "system identities"). In the C code, these 6287 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 6288 macro with the name beginning with "OPENSSL_SYS_", which is determined 6289 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 6290 what is available. 6291 [Richard Levitte] 6292 6293 *) New option -set_serial to 'req' and 'x509' this allows the serial 6294 number to use to be specified on the command line. Previously self 6295 signed certificates were hard coded with serial number 0 and the 6296 CA options of 'x509' had to use a serial number in a file which was 6297 auto incremented. 6298 [Steve Henson] 6299 6300 *) New options to 'ca' utility to support V2 CRL entry extensions. 6301 Currently CRL reason, invalidity date and hold instruction are 6302 supported. Add new CRL extensions to V3 code and some new objects. 6303 [Steve Henson] 6304 6305 *) New function EVP_CIPHER_CTX_set_padding() this is used to 6306 disable standard block padding (aka PKCS#5 padding) in the EVP 6307 API, which was previously mandatory. This means that the data is 6308 not padded in any way and so the total length much be a multiple 6309 of the block size, otherwise an error occurs. 6310 [Steve Henson] 6311 6312 *) Initial (incomplete) OCSP SSL support. 6313 [Steve Henson] 6314 6315 *) New function OCSP_parse_url(). This splits up a URL into its host, 6316 port and path components: primarily to parse OCSP URLs. New -url 6317 option to ocsp utility. 6318 [Steve Henson] 6319 6320 *) New nonce behavior. The return value of OCSP_check_nonce() now 6321 reflects the various checks performed. Applications can decide 6322 whether to tolerate certain situations such as an absent nonce 6323 in a response when one was present in a request: the ocsp application 6324 just prints out a warning. New function OCSP_add1_basic_nonce() 6325 this is to allow responders to include a nonce in a response even if 6326 the request is nonce-less. 6327 [Steve Henson] 6328 6329 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 6330 skipped when using openssl x509 multiple times on a single input file, 6331 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 6332 [Bodo Moeller] 6333 6334 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 6335 set string type: to handle setting ASN1_TIME structures. Fix ca 6336 utility to correctly initialize revocation date of CRLs. 6337 [Steve Henson] 6338 6339 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 6340 the clients preferred ciphersuites and rather use its own preferences. 6341 Should help to work around M$ SGC (Server Gated Cryptography) bug in 6342 Internet Explorer by ensuring unchanged hash method during stepup. 6343 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 6344 [Lutz Jaenicke] 6345 6346 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 6347 to aes and add a new 'exist' option to print out symbols that don't 6348 appear to exist. 6349 [Steve Henson] 6350 6351 *) Additional options to ocsp utility to allow flags to be set and 6352 additional certificates supplied. 6353 [Steve Henson] 6354 6355 *) Add the option -VAfile to 'openssl ocsp', so the user can give the 6356 OCSP client a number of certificate to only verify the response 6357 signature against. 6358 [Richard Levitte] 6359 6360 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 6361 handle the new API. Currently only ECB, CBC modes supported. Add new 6362 AES OIDs. 6363 6364 Add TLS AES ciphersuites as described in RFC3268, "Advanced 6365 Encryption Standard (AES) Ciphersuites for Transport Layer 6366 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 6367 not enabled by default and were not part of the "ALL" ciphersuite 6368 alias because they were not yet official; they could be 6369 explicitly requested by specifying the "AESdraft" ciphersuite 6370 group alias. In the final release of OpenSSL 0.9.7, the group 6371 alias is called "AES" and is part of "ALL".) 6372 [Ben Laurie, Steve Henson, Bodo Moeller] 6373 6374 *) New function OCSP_copy_nonce() to copy nonce value (if present) from 6375 request to response. 6376 [Steve Henson] 6377 6378 *) Functions for OCSP responders. OCSP_request_onereq_count(), 6379 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 6380 extract information from a certificate request. OCSP_response_create() 6381 creates a response and optionally adds a basic response structure. 6382 OCSP_basic_add1_status() adds a complete single response to a basic 6383 response and returns the OCSP_SINGLERESP structure just added (to allow 6384 extensions to be included for example). OCSP_basic_add1_cert() adds a 6385 certificate to a basic response and OCSP_basic_sign() signs a basic 6386 response with various flags. New helper functions ASN1_TIME_check() 6387 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 6388 (converts ASN1_TIME to GeneralizedTime). 6389 [Steve Henson] 6390 6391 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 6392 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 6393 structure from a certificate. X509_pubkey_digest() digests the public_key 6394 contents: this is used in various key identifiers. 6395 [Steve Henson] 6396 6397 *) Make sk_sort() tolerate a NULL argument. 6398 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 6399 6400 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 6401 passed by the function are trusted implicitly. If any of them signed the 6402 response then it is assumed to be valid and is not verified. 6403 [Steve Henson] 6404 6405 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 6406 to data. This was previously part of the PKCS7 ASN1 code. This 6407 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 6408 [Steve Henson, reported by Kenneth R. Robinette 6409 <support@securenetterm.com>] 6410 6411 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 6412 routines: without these tracing memory leaks is very painful. 6413 Fix leaks in PKCS12 and PKCS7 routines. 6414 [Steve Henson] 6415 6416 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 6417 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 6418 effectively meant GeneralizedTime would never be used. Now it 6419 is initialised to -1 but X509_time_adj() now has to check the value 6420 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 6421 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 6422 [Steve Henson, reported by Kenneth R. Robinette 6423 <support@securenetterm.com>] 6424 6425 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 6426 result in a zero length in the ASN1_INTEGER structure which was 6427 not consistent with the structure when d2i_ASN1_INTEGER() was used 6428 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 6429 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 6430 where it did not print out a minus for negative ASN1_INTEGER. 6431 [Steve Henson] 6432 6433 *) Add summary printout to ocsp utility. The various functions which 6434 convert status values to strings have been renamed to: 6435 OCSP_response_status_str(), OCSP_cert_status_str() and 6436 OCSP_crl_reason_str() and are no longer static. New options 6437 to verify nonce values and to disable verification. OCSP response 6438 printout format cleaned up. 6439 [Steve Henson] 6440 6441 *) Add additional OCSP certificate checks. These are those specified 6442 in RFC2560. This consists of two separate checks: the CA of the 6443 certificate being checked must either be the OCSP signer certificate 6444 or the issuer of the OCSP signer certificate. In the latter case the 6445 OCSP signer certificate must contain the OCSP signing extended key 6446 usage. This check is performed by attempting to match the OCSP 6447 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 6448 in the OCSP_CERTID structures of the response. 6449 [Steve Henson] 6450 6451 *) Initial OCSP certificate verification added to OCSP_basic_verify() 6452 and related routines. This uses the standard OpenSSL certificate 6453 verify routines to perform initial checks (just CA validity) and 6454 to obtain the certificate chain. Then additional checks will be 6455 performed on the chain. Currently the root CA is checked to see 6456 if it is explicitly trusted for OCSP signing. This is used to set 6457 a root CA as a global signing root: that is any certificate that 6458 chains to that CA is an acceptable OCSP signing certificate. 6459 [Steve Henson] 6460 6461 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 6462 extensions from a separate configuration file. 6463 As when reading extensions from the main configuration file, 6464 the '-extensions ...' option may be used for specifying the 6465 section to use. 6466 [Massimiliano Pala <madwolf@comune.modena.it>] 6467 6468 *) New OCSP utility. Allows OCSP requests to be generated or 6469 read. The request can be sent to a responder and the output 6470 parsed, outputed or printed in text form. Not complete yet: 6471 still needs to check the OCSP response validity. 6472 [Steve Henson] 6473 6474 *) New subcommands for 'openssl ca': 6475 'openssl ca -status <serial>' prints the status of the cert with 6476 the given serial number (according to the index file). 6477 'openssl ca -updatedb' updates the expiry status of certificates 6478 in the index file. 6479 [Massimiliano Pala <madwolf@comune.modena.it>] 6480 6481 *) New '-newreq-nodes' command option to CA.pl. This is like 6482 '-newreq', but calls 'openssl req' with the '-nodes' option 6483 so that the resulting key is not encrypted. 6484 [Damien Miller <djm@mindrot.org>] 6485 6486 *) New configuration for the GNU Hurd. 6487 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 6488 6489 *) Initial code to implement OCSP basic response verify. This 6490 is currently incomplete. Currently just finds the signer's 6491 certificate and verifies the signature on the response. 6492 [Steve Henson] 6493 6494 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 6495 value of OPENSSLDIR. This is available via the new '-d' option 6496 to 'openssl version', and is also included in 'openssl version -a'. 6497 [Bodo Moeller] 6498 6499 *) Allowing defining memory allocation callbacks that will be given 6500 file name and line number information in additional arguments 6501 (a const char* and an int). The basic functionality remains, as 6502 well as the original possibility to just replace malloc(), 6503 realloc() and free() by functions that do not know about these 6504 additional arguments. To register and find out the current 6505 settings for extended allocation functions, the following 6506 functions are provided: 6507 6508 CRYPTO_set_mem_ex_functions 6509 CRYPTO_set_locked_mem_ex_functions 6510 CRYPTO_get_mem_ex_functions 6511 CRYPTO_get_locked_mem_ex_functions 6512 6513 These work the same way as CRYPTO_set_mem_functions and friends. 6514 CRYPTO_get_[locked_]mem_functions now writes 0 where such an 6515 extended allocation function is enabled. 6516 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 6517 a conventional allocation function is enabled. 6518 [Richard Levitte, Bodo Moeller] 6519 6520 *) Finish off removing the remaining LHASH function pointer casts. 6521 There should no longer be any prototype-casting required when using 6522 the LHASH abstraction, and any casts that remain are "bugs". See 6523 the callback types and macros at the head of lhash.h for details 6524 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 6525 [Geoff Thorpe] 6526 6527 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 6528 If /dev/[u]random devices are not available or do not return enough 6529 entropy, EGD style sockets (served by EGD or PRNGD) will automatically 6530 be queried. 6531 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 6532 /etc/entropy will be queried once each in this sequence, quering stops 6533 when enough entropy was collected without querying more sockets. 6534 [Lutz Jaenicke] 6535 6536 *) Change the Unix RAND_poll() variant to be able to poll several 6537 random devices, as specified by DEVRANDOM, until a sufficient amount 6538 of data has been collected. We spend at most 10 ms on each file 6539 (select timeout) and read in non-blocking mode. DEVRANDOM now 6540 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 6541 (previously it was just the string "/dev/urandom"), so on typical 6542 platforms the 10 ms delay will never occur. 6543 Also separate out the Unix variant to its own file, rand_unix.c. 6544 For VMS, there's a currently-empty rand_vms.c. 6545 [Richard Levitte] 6546 6547 *) Move OCSP client related routines to ocsp_cl.c. These 6548 provide utility functions which an application needing 6549 to issue a request to an OCSP responder and analyse the 6550 response will typically need: as opposed to those which an 6551 OCSP responder itself would need which will be added later. 6552 6553 OCSP_request_sign() signs an OCSP request with an API similar 6554 to PKCS7_sign(). OCSP_response_status() returns status of OCSP 6555 response. OCSP_response_get1_basic() extracts basic response 6556 from response. OCSP_resp_find_status(): finds and extracts status 6557 information from an OCSP_CERTID structure (which will be created 6558 when the request structure is built). These are built from lower 6559 level functions which work on OCSP_SINGLERESP structures but 6560 wont normally be used unless the application wishes to examine 6561 extensions in the OCSP response for example. 6562 6563 Replace nonce routines with a pair of functions. 6564 OCSP_request_add1_nonce() adds a nonce value and optionally 6565 generates a random value. OCSP_check_nonce() checks the 6566 validity of the nonce in an OCSP response. 6567 [Steve Henson] 6568 6569 *) Change function OCSP_request_add() to OCSP_request_add0_id(). 6570 This doesn't copy the supplied OCSP_CERTID and avoids the 6571 need to free up the newly created id. Change return type 6572 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 6573 This can then be used to add extensions to the request. 6574 Deleted OCSP_request_new(), since most of its functionality 6575 is now in OCSP_REQUEST_new() (and the case insensitive name 6576 clash) apart from the ability to set the request name which 6577 will be added elsewhere. 6578 [Steve Henson] 6579 6580 *) Update OCSP API. Remove obsolete extensions argument from 6581 various functions. Extensions are now handled using the new 6582 OCSP extension code. New simple OCSP HTTP function which 6583 can be used to send requests and parse the response. 6584 [Steve Henson] 6585 6586 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 6587 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 6588 uses the special reorder version of SET OF to sort the attributes 6589 and reorder them to match the encoded order. This resolves a long 6590 standing problem: a verify on a PKCS7 structure just after signing 6591 it used to fail because the attribute order did not match the 6592 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 6593 it uses the received order. This is necessary to tolerate some broken 6594 software that does not order SET OF. This is handled by encoding 6595 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 6596 to produce the required SET OF. 6597 [Steve Henson] 6598 6599 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 6600 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 6601 files to get correct declarations of the ASN.1 item variables. 6602 [Richard Levitte] 6603 6604 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 6605 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 6606 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 6607 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 6608 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 6609 ASN1_ITEM and no wrapper functions. 6610 [Steve Henson] 6611 6612 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 6613 replace the old function pointer based I/O routines. Change most of 6614 the *_d2i_bio() and *_d2i_fp() functions to use these. 6615 [Steve Henson] 6616 6617 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 6618 lines, recognice more "algorithms" that can be deselected, and make 6619 it complain about algorithm deselection that isn't recognised. 6620 [Richard Levitte] 6621 6622 *) New ASN1 functions to handle dup, sign, verify, digest, pack and 6623 unpack operations in terms of ASN1_ITEM. Modify existing wrappers 6624 to use new functions. Add NO_ASN1_OLD which can be set to remove 6625 some old style ASN1 functions: this can be used to determine if old 6626 code will still work when these eventually go away. 6627 [Steve Henson] 6628 6629 *) New extension functions for OCSP structures, these follow the 6630 same conventions as certificates and CRLs. 6631 [Steve Henson] 6632 6633 *) New function X509V3_add1_i2d(). This automatically encodes and 6634 adds an extension. Its behaviour can be customised with various 6635 flags to append, replace or delete. Various wrappers added for 6636 certifcates and CRLs. 6637 [Steve Henson] 6638 6639 *) Fix to avoid calling the underlying ASN1 print routine when 6640 an extension cannot be parsed. Correct a typo in the 6641 OCSP_SERVICELOC extension. Tidy up print OCSP format. 6642 [Steve Henson] 6643 6644 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 6645 entries for variables. 6646 [Steve Henson] 6647 6648 *) Add functionality to apps/openssl.c for detecting locking 6649 problems: As the program is single-threaded, all we have 6650 to do is register a locking callback using an array for 6651 storing which locks are currently held by the program. 6652 [Bodo Moeller] 6653 6654 *) Use a lock around the call to CRYPTO_get_ex_new_index() in 6655 SSL_get_ex_data_X509_STORE_idx(), which is used in 6656 ssl_verify_cert_chain() and thus can be called at any time 6657 during TLS/SSL handshakes so that thread-safety is essential. 6658 Unfortunately, the ex_data design is not at all suited 6659 for multi-threaded use, so it probably should be abolished. 6660 [Bodo Moeller] 6661 6662 *) Added Broadcom "ubsec" ENGINE to OpenSSL. 6663 [Broadcom, tweaked and integrated by Geoff Thorpe] 6664 6665 *) Move common extension printing code to new function 6666 X509V3_print_extensions(). Reorganise OCSP print routines and 6667 implement some needed OCSP ASN1 functions. Add OCSP extensions. 6668 [Steve Henson] 6669 6670 *) New function X509_signature_print() to remove duplication in some 6671 print routines. 6672 [Steve Henson] 6673 6674 *) Add a special meaning when SET OF and SEQUENCE OF flags are both 6675 set (this was treated exactly the same as SET OF previously). This 6676 is used to reorder the STACK representing the structure to match the 6677 encoding. This will be used to get round a problem where a PKCS7 6678 structure which was signed could not be verified because the STACK 6679 order did not reflect the encoded order. 6680 [Steve Henson] 6681 6682 *) Reimplement the OCSP ASN1 module using the new code. 6683 [Steve Henson] 6684 6685 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 6686 for its ASN1 operations. The old style function pointers still exist 6687 for now but they will eventually go away. 6688 [Steve Henson] 6689 6690 *) Merge in replacement ASN1 code from the ASN1 branch. This almost 6691 completely replaces the old ASN1 functionality with a table driven 6692 encoder and decoder which interprets an ASN1_ITEM structure describing 6693 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 6694 largely maintained. Almost all of the old asn1_mac.h macro based ASN1 6695 has also been converted to the new form. 6696 [Steve Henson] 6697 6698 *) Change BN_mod_exp_recp so that negative moduli are tolerated 6699 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 6700 so that BN_mod_exp_mont and BN_mod_exp_mont_word work 6701 for negative moduli. 6702 [Bodo Moeller] 6703 6704 *) Fix BN_uadd and BN_usub: Always return non-negative results instead 6705 of not touching the result's sign bit. 6706 [Bodo Moeller] 6707 6708 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 6709 set. 6710 [Bodo Moeller] 6711 6712 *) Changed the LHASH code to use prototypes for callbacks, and created 6713 macros to declare and implement thin (optionally static) functions 6714 that provide type-safety and avoid function pointer casting for the 6715 type-specific callbacks. 6716 [Geoff Thorpe] 6717 6718 *) Added Kerberos Cipher Suites to be used with TLS, as written in 6719 RFC 2712. 6720 [Veers Staats <staatsvr@asc.hpc.mil>, 6721 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 6722 6723 *) Reformat the FAQ so the different questions and answers can be divided 6724 in sections depending on the subject. 6725 [Richard Levitte] 6726 6727 *) Have the zlib compression code load ZLIB.DLL dynamically under 6728 Windows. 6729 [Richard Levitte] 6730 6731 *) New function BN_mod_sqrt for computing square roots modulo a prime 6732 (using the probabilistic Tonelli-Shanks algorithm unless 6733 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 6734 be handled deterministically). 6735 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 6736 6737 *) Make BN_mod_inverse faster by explicitly handling small quotients 6738 in the Euclid loop. (Speed gain about 20% for small moduli [256 or 6739 512 bits], about 30% for larger ones [1024 or 2048 bits].) 6740 [Bodo Moeller] 6741 6742 *) New function BN_kronecker. 6743 [Bodo Moeller] 6744 6745 *) Fix BN_gcd so that it works on negative inputs; the result is 6746 positive unless both parameters are zero. 6747 Previously something reasonably close to an infinite loop was 6748 possible because numbers could be growing instead of shrinking 6749 in the implementation of Euclid's algorithm. 6750 [Bodo Moeller] 6751 6752 *) Fix BN_is_word() and BN_is_one() macros to take into account the 6753 sign of the number in question. 6754 6755 Fix BN_is_word(a,w) to work correctly for w == 0. 6756 6757 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 6758 because its test if the absolute value of 'a' equals 'w'. 6759 Note that BN_abs_is_word does *not* handle w == 0 reliably; 6760 it exists mostly for use in the implementations of BN_is_zero(), 6761 BN_is_one(), and BN_is_word(). 6762 [Bodo Moeller] 6763 6764 *) New function BN_swap. 6765 [Bodo Moeller] 6766 6767 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 6768 the exponentiation functions are more likely to produce reasonable 6769 results on negative inputs. 6770 [Bodo Moeller] 6771 6772 *) Change BN_mod_mul so that the result is always non-negative. 6773 Previously, it could be negative if one of the factors was negative; 6774 I don't think anyone really wanted that behaviour. 6775 [Bodo Moeller] 6776 6777 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 6778 (except for exponentiation, which stays in crypto/bn/bn_exp.c, 6779 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 6780 and add new functions: 6781 6782 BN_nnmod 6783 BN_mod_sqr 6784 BN_mod_add 6785 BN_mod_add_quick 6786 BN_mod_sub 6787 BN_mod_sub_quick 6788 BN_mod_lshift1 6789 BN_mod_lshift1_quick 6790 BN_mod_lshift 6791 BN_mod_lshift_quick 6792 6793 These functions always generate non-negative results. 6794 6795 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 6796 such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 6797 6798 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 6799 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 6800 be reduced modulo m. 6801 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 6802 6803#if 0 6804 The following entry accidentily appeared in the CHANGES file 6805 distributed with OpenSSL 0.9.7. The modifications described in 6806 it do *not* apply to OpenSSL 0.9.7. 6807 6808 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 6809 was actually never needed) and in BN_mul(). The removal in BN_mul() 6810 required a small change in bn_mul_part_recursive() and the addition 6811 of the functions bn_cmp_part_words(), bn_sub_part_words() and 6812 bn_add_part_words(), which do the same thing as bn_cmp_words(), 6813 bn_sub_words() and bn_add_words() except they take arrays with 6814 differing sizes. 6815 [Richard Levitte] 6816#endif 6817 6818 *) In 'openssl passwd', verify passwords read from the terminal 6819 unless the '-salt' option is used (which usually means that 6820 verification would just waste user's time since the resulting 6821 hash is going to be compared with some given password hash) 6822 or the new '-noverify' option is used. 6823 6824 This is an incompatible change, but it does not affect 6825 non-interactive use of 'openssl passwd' (passwords on the command 6826 line, '-stdin' option, '-in ...' option) and thus should not 6827 cause any problems. 6828 [Bodo Moeller] 6829 6830 *) Remove all references to RSAref, since there's no more need for it. 6831 [Richard Levitte] 6832 6833 *) Make DSO load along a path given through an environment variable 6834 (SHLIB_PATH) with shl_load(). 6835 [Richard Levitte] 6836 6837 *) Constify the ENGINE code as a result of BIGNUM constification. 6838 Also constify the RSA code and most things related to it. In a 6839 few places, most notable in the depth of the ASN.1 code, ugly 6840 casts back to non-const were required (to be solved at a later 6841 time) 6842 [Richard Levitte] 6843 6844 *) Make it so the openssl application has all engines loaded by default. 6845 [Richard Levitte] 6846 6847 *) Constify the BIGNUM routines a little more. 6848 [Richard Levitte] 6849 6850 *) Add the following functions: 6851 6852 ENGINE_load_cswift() 6853 ENGINE_load_chil() 6854 ENGINE_load_atalla() 6855 ENGINE_load_nuron() 6856 ENGINE_load_builtin_engines() 6857 6858 That way, an application can itself choose if external engines that 6859 are built-in in OpenSSL shall ever be used or not. The benefit is 6860 that applications won't have to be linked with libdl or other dso 6861 libraries unless it's really needed. 6862 6863 Changed 'openssl engine' to load all engines on demand. 6864 Changed the engine header files to avoid the duplication of some 6865 declarations (they differed!). 6866 [Richard Levitte] 6867 6868 *) 'openssl engine' can now list capabilities. 6869 [Richard Levitte] 6870 6871 *) Better error reporting in 'openssl engine'. 6872 [Richard Levitte] 6873 6874 *) Never call load_dh_param(NULL) in s_server. 6875 [Bodo Moeller] 6876 6877 *) Add engine application. It can currently list engines by name and 6878 identity, and test if they are actually available. 6879 [Richard Levitte] 6880 6881 *) Improve RPM specification file by forcing symbolic linking and making 6882 sure the installed documentation is also owned by root.root. 6883 [Damien Miller <djm@mindrot.org>] 6884 6885 *) Give the OpenSSL applications more possibilities to make use of 6886 keys (public as well as private) handled by engines. 6887 [Richard Levitte] 6888 6889 *) Add OCSP code that comes from CertCo. 6890 [Richard Levitte] 6891 6892 *) Add VMS support for the Rijndael code. 6893 [Richard Levitte] 6894 6895 *) Added untested support for Nuron crypto accelerator. 6896 [Ben Laurie] 6897 6898 *) Add support for external cryptographic devices. This code was 6899 previously distributed separately as the "engine" branch. 6900 [Geoff Thorpe, Richard Levitte] 6901 6902 *) Rework the filename-translation in the DSO code. It is now possible to 6903 have far greater control over how a "name" is turned into a filename 6904 depending on the operating environment and any oddities about the 6905 different shared library filenames on each system. 6906 [Geoff Thorpe] 6907 6908 *) Support threads on FreeBSD-elf in Configure. 6909 [Richard Levitte] 6910 6911 *) Fix for SHA1 assembly problem with MASM: it produces 6912 warnings about corrupt line number information when assembling 6913 with debugging information. This is caused by the overlapping 6914 of two sections. 6915 [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 6916 6917 *) NCONF changes. 6918 NCONF_get_number() has no error checking at all. As a replacement, 6919 NCONF_get_number_e() is defined (_e for "error checking") and is 6920 promoted strongly. The old NCONF_get_number is kept around for 6921 binary backward compatibility. 6922 Make it possible for methods to load from something other than a BIO, 6923 by providing a function pointer that is given a name instead of a BIO. 6924 For example, this could be used to load configuration data from an 6925 LDAP server. 6926 [Richard Levitte] 6927 6928 *) Fix for non blocking accept BIOs. Added new I/O special reason 6929 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 6930 with non blocking I/O was not possible because no retry code was 6931 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 6932 this case. 6933 [Steve Henson] 6934 6935 *) Added the beginnings of Rijndael support. 6936 [Ben Laurie] 6937 6938 *) Fix for bug in DirectoryString mask setting. Add support for 6939 X509_NAME_print_ex() in 'req' and X509_print_ex() function 6940 to allow certificate printing to more controllable, additional 6941 'certopt' option to 'x509' to allow new printing options to be 6942 set. 6943 [Steve Henson] 6944 6945 *) Clean old EAY MD5 hack from e_os.h. 6946 [Richard Levitte] 6947 6948 Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 6949 6950 *) Fix null-pointer assignment in do_change_cipher_spec() revealed 6951 by using the Codenomicon TLS Test Tool (CVE-2004-0079) 6952 [Joe Orton, Steve Henson] 6953 6954 Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 6955 6956 *) Fix additional bug revealed by the NISCC test suite: 6957 6958 Stop bug triggering large recursion when presented with 6959 certain ASN.1 tags (CVE-2003-0851) 6960 [Steve Henson] 6961 6962 Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 6963 6964 *) Fix various bugs revealed by running the NISCC test suite: 6965 6966 Stop out of bounds reads in the ASN1 code when presented with 6967 invalid tags (CVE-2003-0543 and CVE-2003-0544). 6968 6969 If verify callback ignores invalid public key errors don't try to check 6970 certificate signature with the NULL public key. 6971 6972 [Steve Henson] 6973 6974 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 6975 if the server requested one: as stated in TLS 1.0 and SSL 3.0 6976 specifications. 6977 [Steve Henson] 6978 6979 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 6980 extra data after the compression methods not only for TLS 1.0 6981 but also for SSL 3.0 (as required by the specification). 6982 [Bodo Moeller; problem pointed out by Matthias Loepfe] 6983 6984 *) Change X509_certificate_type() to mark the key as exported/exportable 6985 when it's 512 *bits* long, not 512 bytes. 6986 [Richard Levitte] 6987 6988 Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 6989 6990 *) Countermeasure against the Klima-Pokorny-Rosa extension of 6991 Bleichbacher's attack on PKCS #1 v1.5 padding: treat 6992 a protocol version number mismatch like a decryption error 6993 in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 6994 [Bodo Moeller] 6995 6996 *) Turn on RSA blinding by default in the default implementation 6997 to avoid a timing attack. Applications that don't want it can call 6998 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 6999 They would be ill-advised to do so in most cases. 7000 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 7001 7002 *) Change RSA blinding code so that it works when the PRNG is not 7003 seeded (in this case, the secret RSA exponent is abused as 7004 an unpredictable seed -- if it is not unpredictable, there 7005 is no point in blinding anyway). Make RSA blinding thread-safe 7006 by remembering the creator's thread ID in rsa->blinding and 7007 having all other threads use local one-time blinding factors 7008 (this requires more computation than sharing rsa->blinding, but 7009 avoids excessive locking; and if an RSA object is not shared 7010 between threads, blinding will still be very fast). 7011 [Bodo Moeller] 7012 7013 Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 7014 7015 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 7016 via timing by performing a MAC computation even if incorrrect 7017 block cipher padding has been found. This is a countermeasure 7018 against active attacks where the attacker has to distinguish 7019 between bad padding and a MAC verification error. (CVE-2003-0078) 7020 7021 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 7022 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 7023 Martin Vuagnoux (EPFL, Ilion)] 7024 7025 Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 7026 7027 *) New function OPENSSL_cleanse(), which is used to cleanse a section of 7028 memory from it's contents. This is done with a counter that will 7029 place alternating values in each byte. This can be used to solve 7030 two issues: 1) the removal of calls to memset() by highly optimizing 7031 compilers, and 2) cleansing with other values than 0, since those can 7032 be read through on certain media, for example a swap space on disk. 7033 [Geoff Thorpe] 7034 7035 *) Bugfix: client side session caching did not work with external caching, 7036 because the session->cipher setting was not restored when reloading 7037 from the external cache. This problem was masked, when 7038 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 7039 (Found by Steve Haslam <steve@araqnid.ddts.net>.) 7040 [Lutz Jaenicke] 7041 7042 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 7043 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 7044 [Zeev Lieber <zeev-l@yahoo.com>] 7045 7046 *) Undo an undocumented change introduced in 0.9.6e which caused 7047 repeated calls to OpenSSL_add_all_ciphers() and 7048 OpenSSL_add_all_digests() to be ignored, even after calling 7049 EVP_cleanup(). 7050 [Richard Levitte] 7051 7052 *) Change the default configuration reader to deal with last line not 7053 being properly terminated. 7054 [Richard Levitte] 7055 7056 *) Change X509_NAME_cmp() so it applies the special rules on handling 7057 DN values that are of type PrintableString, as well as RDNs of type 7058 emailAddress where the value has the type ia5String. 7059 [stefank@valicert.com via Richard Levitte] 7060 7061 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 7062 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 7063 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 7064 the bitwise-OR of the two for use by the majority of applications 7065 wanting this behaviour, and update the docs. The documented 7066 behaviour and actual behaviour were inconsistent and had been 7067 changing anyway, so this is more a bug-fix than a behavioural 7068 change. 7069 [Geoff Thorpe, diagnosed by Nadav Har'El] 7070 7071 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 7072 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 7073 [Bodo Moeller] 7074 7075 *) Fix initialization code race conditions in 7076 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 7077 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 7078 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 7079 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 7080 ssl2_get_cipher_by_char(), 7081 ssl3_get_cipher_by_char(). 7082 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 7083 7084 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 7085 the cached sessions are flushed, as the remove_cb() might use ex_data 7086 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 7087 (see [openssl.org #212]). 7088 [Geoff Thorpe, Lutz Jaenicke] 7089 7090 *) Fix typo in OBJ_txt2obj which incorrectly passed the content 7091 length, instead of the encoding length to d2i_ASN1_OBJECT. 7092 [Steve Henson] 7093 7094 Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 7095 7096 *) [In 0.9.6g-engine release:] 7097 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 7098 [Lynn Gazis <lgazis@rainbow.com>] 7099 7100 Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 7101 7102 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 7103 and get fix the header length calculation. 7104 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 7105 Alon Kantor <alonk@checkpoint.com> (and others), 7106 Steve Henson] 7107 7108 *) Use proper error handling instead of 'assertions' in buffer 7109 overflow checks added in 0.9.6e. This prevents DoS (the 7110 assertions could call abort()). 7111 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 7112 7113 Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 7114 7115 *) Add various sanity checks to asn1_get_length() to reject 7116 the ASN1 length bytes if they exceed sizeof(long), will appear 7117 negative or the content length exceeds the length of the 7118 supplied buffer. 7119 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 7120 7121 *) Fix cipher selection routines: ciphers without encryption had no flags 7122 for the cipher strength set and where therefore not handled correctly 7123 by the selection routines (PR #130). 7124 [Lutz Jaenicke] 7125 7126 *) Fix EVP_dsa_sha macro. 7127 [Nils Larsch] 7128 7129 *) New option 7130 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 7131 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 7132 that was added in OpenSSL 0.9.6d. 7133 7134 As the countermeasure turned out to be incompatible with some 7135 broken SSL implementations, the new option is part of SSL_OP_ALL. 7136 SSL_OP_ALL is usually employed when compatibility with weird SSL 7137 implementations is desired (e.g. '-bugs' option to 's_client' and 7138 's_server'), so the new option is automatically set in many 7139 applications. 7140 [Bodo Moeller] 7141 7142 *) Changes in security patch: 7143 7144 Changes marked "(CHATS)" were sponsored by the Defense Advanced 7145 Research Projects Agency (DARPA) and Air Force Research Laboratory, 7146 Air Force Materiel Command, USAF, under agreement number 7147 F30602-01-2-0537. 7148 7149 *) Add various sanity checks to asn1_get_length() to reject 7150 the ASN1 length bytes if they exceed sizeof(long), will appear 7151 negative or the content length exceeds the length of the 7152 supplied buffer. (CVE-2002-0659) 7153 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 7154 7155 *) Assertions for various potential buffer overflows, not known to 7156 happen in practice. 7157 [Ben Laurie (CHATS)] 7158 7159 *) Various temporary buffers to hold ASCII versions of integers were 7160 too small for 64 bit platforms. (CVE-2002-0655) 7161 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 7162 7163 *) Remote buffer overflow in SSL3 protocol - an attacker could 7164 supply an oversized session ID to a client. (CVE-2002-0656) 7165 [Ben Laurie (CHATS)] 7166 7167 *) Remote buffer overflow in SSL2 protocol - an attacker could 7168 supply an oversized client master key. (CVE-2002-0656) 7169 [Ben Laurie (CHATS)] 7170 7171 Changes between 0.9.6c and 0.9.6d [9 May 2002] 7172 7173 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 7174 encoded as NULL) with id-dsa-with-sha1. 7175 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 7176 7177 *) Check various X509_...() return values in apps/req.c. 7178 [Nils Larsch <nla@trustcenter.de>] 7179 7180 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 7181 an end-of-file condition would erronously be flagged, when the CRLF 7182 was just at the end of a processed block. The bug was discovered when 7183 processing data through a buffering memory BIO handing the data to a 7184 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 7185 <ptsekov@syntrex.com> and Nedelcho Stanev. 7186 [Lutz Jaenicke] 7187 7188 *) Implement a countermeasure against a vulnerability recently found 7189 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 7190 before application data chunks to avoid the use of known IVs 7191 with data potentially chosen by the attacker. 7192 [Bodo Moeller] 7193 7194 *) Fix length checks in ssl3_get_client_hello(). 7195 [Bodo Moeller] 7196 7197 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 7198 to prevent ssl3_read_internal() from incorrectly assuming that 7199 ssl3_read_bytes() found application data while handshake 7200 processing was enabled when in fact s->s3->in_read_app_data was 7201 merely automatically cleared during the initial handshake. 7202 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 7203 7204 *) Fix object definitions for Private and Enterprise: they were not 7205 recognized in their shortname (=lowercase) representation. Extend 7206 obj_dat.pl to issue an error when using undefined keywords instead 7207 of silently ignoring the problem (Svenning Sorensen 7208 <sss@sss.dnsalias.net>). 7209 [Lutz Jaenicke] 7210 7211 *) Fix DH_generate_parameters() so that it works for 'non-standard' 7212 generators, i.e. generators other than 2 and 5. (Previously, the 7213 code did not properly initialise the 'add' and 'rem' values to 7214 BN_generate_prime().) 7215 7216 In the new general case, we do not insist that 'generator' is 7217 actually a primitive root: This requirement is rather pointless; 7218 a generator of the order-q subgroup is just as good, if not 7219 better. 7220 [Bodo Moeller] 7221 7222 *) Map new X509 verification errors to alerts. Discovered and submitted by 7223 Tom Wu <tom@arcot.com>. 7224 [Lutz Jaenicke] 7225 7226 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 7227 returning non-zero before the data has been completely received 7228 when using non-blocking I/O. 7229 [Bodo Moeller; problem pointed out by John Hughes] 7230 7231 *) Some of the ciphers missed the strength entry (SSL_LOW etc). 7232 [Ben Laurie, Lutz Jaenicke] 7233 7234 *) Fix bug in SSL_clear(): bad sessions were not removed (found by 7235 Yoram Zahavi <YoramZ@gilian.com>). 7236 [Lutz Jaenicke] 7237 7238 *) Add information about CygWin 1.3 and on, and preserve proper 7239 configuration for the versions before that. 7240 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 7241 7242 *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 7243 check whether we deal with a copy of a session and do not delete from 7244 the cache in this case. Problem reported by "Izhar Shoshani Levi" 7245 <izhar@checkpoint.com>. 7246 [Lutz Jaenicke] 7247 7248 *) Do not store session data into the internal session cache, if it 7249 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 7250 flag is set). Proposed by Aslam <aslam@funk.com>. 7251 [Lutz Jaenicke] 7252 7253 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 7254 value is 0. 7255 [Richard Levitte] 7256 7257 *) [In 0.9.6d-engine release:] 7258 Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 7259 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 7260 7261 *) Add the configuration target linux-s390x. 7262 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 7263 7264 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 7265 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 7266 variable as an indication that a ClientHello message has been 7267 received. As the flag value will be lost between multiple 7268 invocations of ssl3_accept when using non-blocking I/O, the 7269 function may not be aware that a handshake has actually taken 7270 place, thus preventing a new session from being added to the 7271 session cache. 7272 7273 To avoid this problem, we now set s->new_session to 2 instead of 7274 using a local variable. 7275 [Lutz Jaenicke, Bodo Moeller] 7276 7277 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 7278 if the SSL_R_LENGTH_MISMATCH error is detected. 7279 [Geoff Thorpe, Bodo Moeller] 7280 7281 *) New 'shared_ldflag' column in Configure platform table. 7282 [Richard Levitte] 7283 7284 *) Fix EVP_CIPHER_mode macro. 7285 ["Dan S. Camper" <dan@bti.net>] 7286 7287 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 7288 type, we must throw them away by setting rr->length to 0. 7289 [D P Chang <dpc@qualys.com>] 7290 7291 Changes between 0.9.6b and 0.9.6c [21 dec 2001] 7292 7293 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 7294 <Dominikus.Scherkl@biodata.com>. (The previous implementation 7295 worked incorrectly for those cases where range = 10..._2 and 7296 3*range is two bits longer than range.) 7297 [Bodo Moeller] 7298 7299 *) Only add signing time to PKCS7 structures if it is not already 7300 present. 7301 [Steve Henson] 7302 7303 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 7304 OBJ_ld_ce should be OBJ_id_ce. 7305 Also some ip-pda OIDs in crypto/objects/objects.txt were 7306 incorrect (cf. RFC 3039). 7307 [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 7308 7309 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 7310 returns early because it has nothing to do. 7311 [Andy Schneider <andy.schneider@bjss.co.uk>] 7312 7313 *) [In 0.9.6c-engine release:] 7314 Fix mutex callback return values in crypto/engine/hw_ncipher.c. 7315 [Andy Schneider <andy.schneider@bjss.co.uk>] 7316 7317 *) [In 0.9.6c-engine release:] 7318 Add support for Cryptographic Appliance's keyserver technology. 7319 (Use engine 'keyclient') 7320 [Cryptographic Appliances and Geoff Thorpe] 7321 7322 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 7323 is called via tools/c89.sh because arguments have to be 7324 rearranged (all '-L' options must appear before the first object 7325 modules). 7326 [Richard Shapiro <rshapiro@abinitio.com>] 7327 7328 *) [In 0.9.6c-engine release:] 7329 Add support for Broadcom crypto accelerator cards, backported 7330 from 0.9.7. 7331 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 7332 7333 *) [In 0.9.6c-engine release:] 7334 Add support for SureWare crypto accelerator cards from 7335 Baltimore Technologies. (Use engine 'sureware') 7336 [Baltimore Technologies and Mark Cox] 7337 7338 *) [In 0.9.6c-engine release:] 7339 Add support for crypto accelerator cards from Accelerated 7340 Encryption Processing, www.aep.ie. (Use engine 'aep') 7341 [AEP Inc. and Mark Cox] 7342 7343 *) Add a configuration entry for gcc on UnixWare. 7344 [Gary Benson <gbenson@redhat.com>] 7345 7346 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 7347 messages are stored in a single piece (fixed-length part and 7348 variable-length part combined) and fix various bugs found on the way. 7349 [Bodo Moeller] 7350 7351 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 7352 instead. BIO_gethostbyname() does not know what timeouts are 7353 appropriate, so entries would stay in cache even when they have 7354 become invalid. 7355 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 7356 7357 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 7358 faced with a pathologically small ClientHello fragment that does 7359 not contain client_version: Instead of aborting with an error, 7360 simply choose the highest available protocol version (i.e., 7361 TLS 1.0 unless it is disabled). In practice, ClientHello 7362 messages are never sent like this, but this change gives us 7363 strictly correct behaviour at least for TLS. 7364 [Bodo Moeller] 7365 7366 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 7367 never resets s->method to s->ctx->method when called from within 7368 one of the SSL handshake functions. 7369 [Bodo Moeller; problem pointed out by Niko Baric] 7370 7371 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 7372 (sent using the client's version number) if client_version is 7373 smaller than the protocol version in use. Also change 7374 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 7375 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 7376 the client will at least see that alert. 7377 [Bodo Moeller] 7378 7379 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 7380 correctly. 7381 [Bodo Moeller] 7382 7383 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 7384 client receives HelloRequest while in a handshake. 7385 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 7386 7387 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 7388 should end in 'break', not 'goto end' which circuments various 7389 cleanups done in state SSL_ST_OK. But session related stuff 7390 must be disabled for SSL_ST_OK in the case that we just sent a 7391 HelloRequest. 7392 7393 Also avoid some overhead by not calling ssl_init_wbio_buffer() 7394 before just sending a HelloRequest. 7395 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 7396 7397 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 7398 reveal whether illegal block cipher padding was found or a MAC 7399 verification error occured. (Neither SSLerr() codes nor alerts 7400 are directly visible to potential attackers, but the information 7401 may leak via logfiles.) 7402 7403 Similar changes are not required for the SSL 2.0 implementation 7404 because the number of padding bytes is sent in clear for SSL 2.0, 7405 and the extra bytes are just ignored. However ssl/s2_pkt.c 7406 failed to verify that the purported number of padding bytes is in 7407 the legal range. 7408 [Bodo Moeller] 7409 7410 *) Add OpenUNIX-8 support including shared libraries 7411 (Boyd Lynn Gerber <gerberb@zenez.com>). 7412 [Lutz Jaenicke] 7413 7414 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 7415 'wristwatch attack' using huge encoding parameters (cf. 7416 James H. Manger's CRYPTO 2001 paper). Note that the 7417 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 7418 encoding parameters and hence was not vulnerable. 7419 [Bodo Moeller] 7420 7421 *) BN_sqr() bug fix. 7422 [Ulf M��ller, reported by Jim Ellis <jim.ellis@cavium.com>] 7423 7424 *) Rabin-Miller test analyses assume uniformly distributed witnesses, 7425 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 7426 followed by modular reduction. 7427 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 7428 7429 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 7430 equivalent based on BN_pseudo_rand() instead of BN_rand(). 7431 [Bodo Moeller] 7432 7433 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 7434 This function was broken, as the check for a new client hello message 7435 to handle SGC did not allow these large messages. 7436 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 7437 [Lutz Jaenicke] 7438 7439 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 7440 [Lutz Jaenicke] 7441 7442 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 7443 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 7444 [Lutz Jaenicke] 7445 7446 *) Rework the configuration and shared library support for Tru64 Unix. 7447 The configuration part makes use of modern compiler features and 7448 still retains old compiler behavior for those that run older versions 7449 of the OS. The shared library support part includes a variant that 7450 uses the RPATH feature, and is available through the special 7451 configuration target "alpha-cc-rpath", which will never be selected 7452 automatically. 7453 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 7454 7455 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 7456 with the same message size as in ssl3_get_certificate_request(). 7457 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 7458 messages might inadvertently be reject as too long. 7459 [Petr Lampa <lampa@fee.vutbr.cz>] 7460 7461 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 7462 [Andy Polyakov] 7463 7464 *) Modified SSL library such that the verify_callback that has been set 7465 specificly for an SSL object with SSL_set_verify() is actually being 7466 used. Before the change, a verify_callback set with this function was 7467 ignored and the verify_callback() set in the SSL_CTX at the time of 7468 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 7469 to allow the necessary settings. 7470 [Lutz Jaenicke] 7471 7472 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 7473 explicitly to NULL, as at least on Solaris 8 this seems not always to be 7474 done automatically (in contradiction to the requirements of the C 7475 standard). This made problems when used from OpenSSH. 7476 [Lutz Jaenicke] 7477 7478 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 7479 dh->length and always used 7480 7481 BN_rand_range(priv_key, dh->p). 7482 7483 BN_rand_range() is not necessary for Diffie-Hellman, and this 7484 specific range makes Diffie-Hellman unnecessarily inefficient if 7485 dh->length (recommended exponent length) is much smaller than the 7486 length of dh->p. We could use BN_rand_range() if the order of 7487 the subgroup was stored in the DH structure, but we only have 7488 dh->length. 7489 7490 So switch back to 7491 7492 BN_rand(priv_key, l, ...) 7493 7494 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 7495 otherwise. 7496 [Bodo Moeller] 7497 7498 *) In 7499 7500 RSA_eay_public_encrypt 7501 RSA_eay_private_decrypt 7502 RSA_eay_private_encrypt (signing) 7503 RSA_eay_public_decrypt (signature verification) 7504 7505 (default implementations for RSA_public_encrypt, 7506 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 7507 always reject numbers >= n. 7508 [Bodo Moeller] 7509 7510 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 7511 to synchronize access to 'locking_thread'. This is necessary on 7512 systems where access to 'locking_thread' (an 'unsigned long' 7513 variable) is not atomic. 7514 [Bodo Moeller] 7515 7516 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 7517 *before* setting the 'crypto_lock_rand' flag. The previous code had 7518 a race condition if 0 is a valid thread ID. 7519 [Travis Vitek <vitek@roguewave.com>] 7520 7521 *) Add support for shared libraries under Irix. 7522 [Albert Chin-A-Young <china@thewrittenword.com>] 7523 7524 *) Add configuration option to build on Linux on both big-endian and 7525 little-endian MIPS. 7526 [Ralf Baechle <ralf@uni-koblenz.de>] 7527 7528 *) Add the possibility to create shared libraries on HP-UX. 7529 [Richard Levitte] 7530 7531 Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 7532 7533 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 7534 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 7535 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 7536 PRNG state recovery was possible based on the output of 7537 one PRNG request appropriately sized to gain knowledge on 7538 'md' followed by enough consecutive 1-byte PRNG requests 7539 to traverse all of 'state'. 7540 7541 1. When updating 'md_local' (the current thread's copy of 'md') 7542 during PRNG output generation, hash all of the previous 7543 'md_local' value, not just the half used for PRNG output. 7544 7545 2. Make the number of bytes from 'state' included into the hash 7546 independent from the number of PRNG bytes requested. 7547 7548 The first measure alone would be sufficient to avoid 7549 Markku-Juhani's attack. (Actually it had never occurred 7550 to me that the half of 'md_local' used for chaining was the 7551 half from which PRNG output bytes were taken -- I had always 7552 assumed that the secret half would be used.) The second 7553 measure makes sure that additional data from 'state' is never 7554 mixed into 'md_local' in small portions; this heuristically 7555 further strengthens the PRNG. 7556 [Bodo Moeller] 7557 7558 *) Fix crypto/bn/asm/mips3.s. 7559 [Andy Polyakov] 7560 7561 *) When only the key is given to "enc", the IV is undefined. Print out 7562 an error message in this case. 7563 [Lutz Jaenicke] 7564 7565 *) Handle special case when X509_NAME is empty in X509 printing routines. 7566 [Steve Henson] 7567 7568 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 7569 positive and less than q. 7570 [Bodo Moeller] 7571 7572 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 7573 used: it isn't thread safe and the add_lock_callback should handle 7574 that itself. 7575 [Paul Rose <Paul.Rose@bridge.com>] 7576 7577 *) Verify that incoming data obeys the block size in 7578 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 7579 [Bodo Moeller] 7580 7581 *) Fix OAEP check. 7582 [Ulf M��ller, Bodo M��ller] 7583 7584 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 7585 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 7586 when fixing the server behaviour for backwards-compatible 'client 7587 hello' messages. (Note that the attack is impractical against 7588 SSL 3.0 and TLS 1.0 anyway because length and version checking 7589 means that the probability of guessing a valid ciphertext is 7590 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 7591 paper.) 7592 7593 Before 0.9.5, the countermeasure (hide the error by generating a 7594 random 'decryption result') did not work properly because 7595 ERR_clear_error() was missing, meaning that SSL_get_error() would 7596 detect the supposedly ignored error. 7597 7598 Both problems are now fixed. 7599 [Bodo Moeller] 7600 7601 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 7602 (previously it was 1024). 7603 [Bodo Moeller] 7604 7605 *) Fix for compatibility mode trust settings: ignore trust settings 7606 unless some valid trust or reject settings are present. 7607 [Steve Henson] 7608 7609 *) Fix for blowfish EVP: its a variable length cipher. 7610 [Steve Henson] 7611 7612 *) Fix various bugs related to DSA S/MIME verification. Handle missing 7613 parameters in DSA public key structures and return an error in the 7614 DSA routines if parameters are absent. 7615 [Steve Henson] 7616 7617 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 7618 in the current directory if neither $RANDFILE nor $HOME was set. 7619 RAND_file_name() in 0.9.6a returned NULL in this case. This has 7620 caused some confusion to Windows users who haven't defined $HOME. 7621 Thus RAND_file_name() is changed again: e_os.h can define a 7622 DEFAULT_HOME, which will be used if $HOME is not set. 7623 For Windows, we use "C:"; on other platforms, we still require 7624 environment variables. 7625 7626 *) Move 'if (!initialized) RAND_poll()' into regions protected by 7627 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 7628 having multiple threads call RAND_poll() concurrently. 7629 [Bodo Moeller] 7630 7631 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 7632 combination of a flag and a thread ID variable. 7633 Otherwise while one thread is in ssleay_rand_bytes (which sets the 7634 flag), *other* threads can enter ssleay_add_bytes without obeying 7635 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 7636 that they do not hold after the first thread unsets add_do_not_lock). 7637 [Bodo Moeller] 7638 7639 *) Change bctest again: '-x' expressions are not available in all 7640 versions of 'test'. 7641 [Bodo Moeller] 7642 7643 Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 7644 7645 *) Fix a couple of memory leaks in PKCS7_dataDecode() 7646 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 7647 7648 *) Change Configure and Makefiles to provide EXE_EXT, which will contain 7649 the default extension for executables, if any. Also, make the perl 7650 scripts that use symlink() to test if it really exists and use "cp" 7651 if it doesn't. All this made OpenSSL compilable and installable in 7652 CygWin. 7653 [Richard Levitte] 7654 7655 *) Fix for asn1_GetSequence() for indefinite length constructed data. 7656 If SEQUENCE is length is indefinite just set c->slen to the total 7657 amount of data available. 7658 [Steve Henson, reported by shige@FreeBSD.org] 7659 [This change does not apply to 0.9.7.] 7660 7661 *) Change bctest to avoid here-documents inside command substitution 7662 (workaround for FreeBSD /bin/sh bug). 7663 For compatibility with Ultrix, avoid shell functions (introduced 7664 in the bctest version that searches along $PATH). 7665 [Bodo Moeller] 7666 7667 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 7668 with des_encrypt() defined on some operating systems, like Solaris 7669 and UnixWare. 7670 [Richard Levitte] 7671 7672 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 7673 On the Importance of Eliminating Errors in Cryptographic 7674 Computations, J. Cryptology 14 (2001) 2, 101-119, 7675 http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 7676 [Ulf Moeller] 7677 7678 *) MIPS assembler BIGNUM division bug fix. 7679 [Andy Polyakov] 7680 7681 *) Disabled incorrect Alpha assembler code. 7682 [Richard Levitte] 7683 7684 *) Fix PKCS#7 decode routines so they correctly update the length 7685 after reading an EOC for the EXPLICIT tag. 7686 [Steve Henson] 7687 [This change does not apply to 0.9.7.] 7688 7689 *) Fix bug in PKCS#12 key generation routines. This was triggered 7690 if a 3DES key was generated with a 0 initial byte. Include 7691 PKCS12_BROKEN_KEYGEN compilation option to retain the old 7692 (but broken) behaviour. 7693 [Steve Henson] 7694 7695 *) Enhance bctest to search for a working bc along $PATH and print 7696 it when found. 7697 [Tim Rice <tim@multitalents.net> via Richard Levitte] 7698 7699 *) Fix memory leaks in err.c: free err_data string if necessary; 7700 don't write to the wrong index in ERR_set_error_data. 7701 [Bodo Moeller] 7702 7703 *) Implement ssl23_peek (analogous to ssl23_read), which previously 7704 did not exist. 7705 [Bodo Moeller] 7706 7707 *) Replace rdtsc with _emit statements for VC++ version 5. 7708 [Jeremy Cooper <jeremy@baymoo.org>] 7709 7710 *) Make it possible to reuse SSLv2 sessions. 7711 [Richard Levitte] 7712 7713 *) In copy_email() check for >= 0 as a return value for 7714 X509_NAME_get_index_by_NID() since 0 is a valid index. 7715 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 7716 7717 *) Avoid coredump with unsupported or invalid public keys by checking if 7718 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 7719 PKCS7_verify() fails with non detached data. 7720 [Steve Henson] 7721 7722 *) Don't use getenv in library functions when run as setuid/setgid. 7723 New function OPENSSL_issetugid(). 7724 [Ulf Moeller] 7725 7726 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 7727 due to incorrect handling of multi-threading: 7728 7729 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 7730 7731 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 7732 7733 3. Count how many times MemCheck_off() has been called so that 7734 nested use can be treated correctly. This also avoids 7735 inband-signalling in the previous code (which relied on the 7736 assumption that thread ID 0 is impossible). 7737 [Bodo Moeller] 7738 7739 *) Add "-rand" option also to s_client and s_server. 7740 [Lutz Jaenicke] 7741 7742 *) Fix CPU detection on Irix 6.x. 7743 [Kurt Hockenbury <khockenb@stevens-tech.edu> and 7744 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 7745 7746 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 7747 was empty. 7748 [Steve Henson] 7749 [This change does not apply to 0.9.7.] 7750 7751 *) Use the cached encoding of an X509_NAME structure rather than 7752 copying it. This is apparently the reason for the libsafe "errors" 7753 but the code is actually correct. 7754 [Steve Henson] 7755 7756 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 7757 Bleichenbacher's DSA attack. 7758 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 7759 to be set and top=0 forces the highest bit to be set; top=-1 is new 7760 and leaves the highest bit random. 7761 [Ulf Moeller, Bodo Moeller] 7762 7763 *) In the NCONF_...-based implementations for CONF_... queries 7764 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 7765 a temporary CONF structure with the data component set to NULL 7766 (which gives segmentation faults in lh_retrieve). 7767 Instead, use NULL for the CONF pointer in CONF_get_string and 7768 CONF_get_number (which may use environment variables) and directly 7769 return NULL from CONF_get_section. 7770 [Bodo Moeller] 7771 7772 *) Fix potential buffer overrun for EBCDIC. 7773 [Ulf Moeller] 7774 7775 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 7776 keyUsage if basicConstraints absent for a CA. 7777 [Steve Henson] 7778 7779 *) Make SMIME_write_PKCS7() write mail header values with a format that 7780 is more generally accepted (no spaces before the semicolon), since 7781 some programs can't parse those values properly otherwise. Also make 7782 sure BIO's that break lines after each write do not create invalid 7783 headers. 7784 [Richard Levitte] 7785 7786 *) Make the CRL encoding routines work with empty SEQUENCE OF. The 7787 macros previously used would not encode an empty SEQUENCE OF 7788 and break the signature. 7789 [Steve Henson] 7790 [This change does not apply to 0.9.7.] 7791 7792 *) Zero the premaster secret after deriving the master secret in 7793 DH ciphersuites. 7794 [Steve Henson] 7795 7796 *) Add some EVP_add_digest_alias registrations (as found in 7797 OpenSSL_add_all_digests()) to SSL_library_init() 7798 aka OpenSSL_add_ssl_algorithms(). This provides improved 7799 compatibility with peers using X.509 certificates 7800 with unconventional AlgorithmIdentifier OIDs. 7801 [Bodo Moeller] 7802 7803 *) Fix for Irix with NO_ASM. 7804 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 7805 7806 *) ./config script fixes. 7807 [Ulf Moeller, Richard Levitte] 7808 7809 *) Fix 'openssl passwd -1'. 7810 [Bodo Moeller] 7811 7812 *) Change PKCS12_key_gen_asc() so it can cope with non null 7813 terminated strings whose length is passed in the passlen 7814 parameter, for example from PEM callbacks. This was done 7815 by adding an extra length parameter to asc2uni(). 7816 [Steve Henson, reported by <oddissey@samsung.co.kr>] 7817 7818 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 7819 call failed, free the DSA structure. 7820 [Bodo Moeller] 7821 7822 *) Fix to uni2asc() to cope with zero length Unicode strings. 7823 These are present in some PKCS#12 files. 7824 [Steve Henson] 7825 7826 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 7827 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 7828 when writing a 32767 byte record. 7829 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 7830 7831 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 7832 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 7833 7834 (RSA objects have a reference count access to which is protected 7835 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 7836 so they are meant to be shared between threads.) 7837 [Bodo Moeller, Geoff Thorpe; original patch submitted by 7838 "Reddie, Steven" <Steven.Reddie@ca.com>] 7839 7840 *) Fix a deadlock in CRYPTO_mem_leaks(). 7841 [Bodo Moeller] 7842 7843 *) Use better test patterns in bntest. 7844 [Ulf M��ller] 7845 7846 *) rand_win.c fix for Borland C. 7847 [Ulf M��ller] 7848 7849 *) BN_rshift bugfix for n == 0. 7850 [Bodo Moeller] 7851 7852 *) Add a 'bctest' script that checks for some known 'bc' bugs 7853 so that 'make test' does not abort just because 'bc' is broken. 7854 [Bodo Moeller] 7855 7856 *) Store verify_result within SSL_SESSION also for client side to 7857 avoid potential security hole. (Re-used sessions on the client side 7858 always resulted in verify_result==X509_V_OK, not using the original 7859 result of the server certificate verification.) 7860 [Lutz Jaenicke] 7861 7862 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 7863 SSL3_RT_APPLICATION_DATA, return 0. 7864 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 7865 [Bodo Moeller] 7866 7867 *) Fix SSL_peek: 7868 Both ssl2_peek and ssl3_peek, which were totally broken in earlier 7869 releases, have been re-implemented by renaming the previous 7870 implementations of ssl2_read and ssl3_read to ssl2_read_internal 7871 and ssl3_read_internal, respectively, and adding 'peek' parameters 7872 to them. The new ssl[23]_{read,peek} functions are calls to 7873 ssl[23]_read_internal with the 'peek' flag set appropriately. 7874 A 'peek' parameter has also been added to ssl3_read_bytes, which 7875 does the actual work for ssl3_read_internal. 7876 [Bodo Moeller] 7877 7878 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 7879 the method-specific "init()" handler. Also clean up ex_data after 7880 calling the method-specific "finish()" handler. Previously, this was 7881 happening the other way round. 7882 [Geoff Thorpe] 7883 7884 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 7885 The previous value, 12, was not always sufficient for BN_mod_exp(). 7886 [Bodo Moeller] 7887 7888 *) Make sure that shared libraries get the internal name engine with 7889 the full version number and not just 0. This should mark the 7890 shared libraries as not backward compatible. Of course, this should 7891 be changed again when we can guarantee backward binary compatibility. 7892 [Richard Levitte] 7893 7894 *) Fix typo in get_cert_by_subject() in by_dir.c 7895 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 7896 7897 *) Rework the system to generate shared libraries: 7898 7899 - Make note of the expected extension for the shared libraries and 7900 if there is a need for symbolic links from for example libcrypto.so.0 7901 to libcrypto.so.0.9.7. There is extended info in Configure for 7902 that. 7903 7904 - Make as few rebuilds of the shared libraries as possible. 7905 7906 - Still avoid linking the OpenSSL programs with the shared libraries. 7907 7908 - When installing, install the shared libraries separately from the 7909 static ones. 7910 [Richard Levitte] 7911 7912 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 7913 7914 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 7915 and not in SSL_clear because the latter is also used by the 7916 accept/connect functions; previously, the settings made by 7917 SSL_set_read_ahead would be lost during the handshake. 7918 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 7919 7920 *) Correct util/mkdef.pl to be selective about disabled algorithms. 7921 Previously, it would create entries for disableed algorithms no 7922 matter what. 7923 [Richard Levitte] 7924 7925 *) Added several new manual pages for SSL_* function. 7926 [Lutz Jaenicke] 7927 7928 Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 7929 7930 *) In ssl23_get_client_hello, generate an error message when faced 7931 with an initial SSL 3.0/TLS record that is too small to contain the 7932 first two bytes of the ClientHello message, i.e. client_version. 7933 (Note that this is a pathologic case that probably has never happened 7934 in real life.) The previous approach was to use the version number 7935 from the record header as a substitute; but our protocol choice 7936 should not depend on that one because it is not authenticated 7937 by the Finished messages. 7938 [Bodo Moeller] 7939 7940 *) More robust randomness gathering functions for Windows. 7941 [Jeffrey Altman <jaltman@columbia.edu>] 7942 7943 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 7944 not set then we don't setup the error code for issuer check errors 7945 to avoid possibly overwriting other errors which the callback does 7946 handle. If an application does set the flag then we assume it knows 7947 what it is doing and can handle the new informational codes 7948 appropriately. 7949 [Steve Henson] 7950 7951 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 7952 a general "ANY" type, as such it should be able to decode anything 7953 including tagged types. However it didn't check the class so it would 7954 wrongly interpret tagged types in the same way as their universal 7955 counterpart and unknown types were just rejected. Changed so that the 7956 tagged and unknown types are handled in the same way as a SEQUENCE: 7957 that is the encoding is stored intact. There is also a new type 7958 "V_ASN1_OTHER" which is used when the class is not universal, in this 7959 case we have no idea what the actual type is so we just lump them all 7960 together. 7961 [Steve Henson] 7962 7963 *) On VMS, stdout may very well lead to a file that is written to 7964 in a record-oriented fashion. That means that every write() will 7965 write a separate record, which will be read separately by the 7966 programs trying to read from it. This can be very confusing. 7967 7968 The solution is to put a BIO filter in the way that will buffer 7969 text until a linefeed is reached, and then write everything a 7970 line at a time, so every record written will be an actual line, 7971 not chunks of lines and not (usually doesn't happen, but I've 7972 seen it once) several lines in one record. BIO_f_linebuffer() is 7973 the answer. 7974 7975 Currently, it's a VMS-only method, because that's where it has 7976 been tested well enough. 7977 [Richard Levitte] 7978 7979 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 7980 it can return incorrect results. 7981 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 7982 but it was in 0.9.6-beta[12].) 7983 [Bodo Moeller] 7984 7985 *) Disable the check for content being present when verifying detached 7986 signatures in pk7_smime.c. Some versions of Netscape (wrongly) 7987 include zero length content when signing messages. 7988 [Steve Henson] 7989 7990 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 7991 BIO_ctrl (for BIO pairs). 7992 [Bodo M��ller] 7993 7994 *) Add DSO method for VMS. 7995 [Richard Levitte] 7996 7997 *) Bug fix: Montgomery multiplication could produce results with the 7998 wrong sign. 7999 [Ulf M��ller] 8000 8001 *) Add RPM specification openssl.spec and modify it to build three 8002 packages. The default package contains applications, application 8003 documentation and run-time libraries. The devel package contains 8004 include files, static libraries and function documentation. The 8005 doc package contains the contents of the doc directory. The original 8006 openssl.spec was provided by Damien Miller <djm@mindrot.org>. 8007 [Richard Levitte] 8008 8009 *) Add a large number of documentation files for many SSL routines. 8010 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 8011 8012 *) Add a configuration entry for Sony News 4. 8013 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 8014 8015 *) Don't set the two most significant bits to one when generating a 8016 random number < q in the DSA library. 8017 [Ulf M��ller] 8018 8019 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 8020 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 8021 the underlying transport is blocking) if a handshake took place. 8022 (The default behaviour is needed by applications such as s_client 8023 and s_server that use select() to determine when to use SSL_read; 8024 but for applications that know in advance when to expect data, it 8025 just makes things more complicated.) 8026 [Bodo Moeller] 8027 8028 *) Add RAND_egd_bytes(), which gives control over the number of bytes read 8029 from EGD. 8030 [Ben Laurie] 8031 8032 *) Add a few more EBCDIC conditionals that make `req' and `x509' 8033 work better on such systems. 8034 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 8035 8036 *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 8037 Update PKCS12_parse() so it copies the friendlyName and the 8038 keyid to the certificates aux info. 8039 [Steve Henson] 8040 8041 *) Fix bug in PKCS7_verify() which caused an infinite loop 8042 if there was more than one signature. 8043 [Sven Uszpelkat <su@celocom.de>] 8044 8045 *) Major change in util/mkdef.pl to include extra information 8046 about each symbol, as well as presentig variables as well 8047 as functions. This change means that there's n more need 8048 to rebuild the .num files when some algorithms are excluded. 8049 [Richard Levitte] 8050 8051 *) Allow the verify time to be set by an application, 8052 rather than always using the current time. 8053 [Steve Henson] 8054 8055 *) Phase 2 verify code reorganisation. The certificate 8056 verify code now looks up an issuer certificate by a 8057 number of criteria: subject name, authority key id 8058 and key usage. It also verifies self signed certificates 8059 by the same criteria. The main comparison function is 8060 X509_check_issued() which performs these checks. 8061 8062 Lot of changes were necessary in order to support this 8063 without completely rewriting the lookup code. 8064 8065 Authority and subject key identifier are now cached. 8066 8067 The LHASH 'certs' is X509_STORE has now been replaced 8068 by a STACK_OF(X509_OBJECT). This is mainly because an 8069 LHASH can't store or retrieve multiple objects with 8070 the same hash value. 8071 8072 As a result various functions (which were all internal 8073 use only) have changed to handle the new X509_STORE 8074 structure. This will break anything that messed round 8075 with X509_STORE internally. 8076 8077 The functions X509_STORE_add_cert() now checks for an 8078 exact match, rather than just subject name. 8079 8080 The X509_STORE API doesn't directly support the retrieval 8081 of multiple certificates matching a given criteria, however 8082 this can be worked round by performing a lookup first 8083 (which will fill the cache with candidate certificates) 8084 and then examining the cache for matches. This is probably 8085 the best we can do without throwing out X509_LOOKUP 8086 entirely (maybe later...). 8087 8088 The X509_VERIFY_CTX structure has been enhanced considerably. 8089 8090 All certificate lookup operations now go via a get_issuer() 8091 callback. Although this currently uses an X509_STORE it 8092 can be replaced by custom lookups. This is a simple way 8093 to bypass the X509_STORE hackery necessary to make this 8094 work and makes it possible to use more efficient techniques 8095 in future. A very simple version which uses a simple 8096 STACK for its trusted certificate store is also provided 8097 using X509_STORE_CTX_trusted_stack(). 8098 8099 The verify_cb() and verify() callbacks now have equivalents 8100 in the X509_STORE_CTX structure. 8101 8102 X509_STORE_CTX also has a 'flags' field which can be used 8103 to customise the verify behaviour. 8104 [Steve Henson] 8105 8106 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 8107 excludes S/MIME capabilities. 8108 [Steve Henson] 8109 8110 *) When a certificate request is read in keep a copy of the 8111 original encoding of the signed data and use it when outputing 8112 again. Signatures then use the original encoding rather than 8113 a decoded, encoded version which may cause problems if the 8114 request is improperly encoded. 8115 [Steve Henson] 8116 8117 *) For consistency with other BIO_puts implementations, call 8118 buffer_write(b, ...) directly in buffer_puts instead of calling 8119 BIO_write(b, ...). 8120 8121 In BIO_puts, increment b->num_write as in BIO_write. 8122 [Peter.Sylvester@EdelWeb.fr] 8123 8124 *) Fix BN_mul_word for the case where the word is 0. (We have to use 8125 BN_zero, we may not return a BIGNUM with an array consisting of 8126 words set to zero.) 8127 [Bodo Moeller] 8128 8129 *) Avoid calling abort() from within the library when problems are 8130 detected, except if preprocessor symbols have been defined 8131 (such as REF_CHECK, BN_DEBUG etc.). 8132 [Bodo Moeller] 8133 8134 *) New openssl application 'rsautl'. This utility can be 8135 used for low level RSA operations. DER public key 8136 BIO/fp routines also added. 8137 [Steve Henson] 8138 8139 *) New Configure entry and patches for compiling on QNX 4. 8140 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 8141 8142 *) A demo state-machine implementation was sponsored by 8143 Nuron (http://www.nuron.com/) and is now available in 8144 demos/state_machine. 8145 [Ben Laurie] 8146 8147 *) New options added to the 'dgst' utility for signature 8148 generation and verification. 8149 [Steve Henson] 8150 8151 *) Unrecognized PKCS#7 content types are now handled via a 8152 catch all ASN1_TYPE structure. This allows unsupported 8153 types to be stored as a "blob" and an application can 8154 encode and decode it manually. 8155 [Steve Henson] 8156 8157 *) Fix various signed/unsigned issues to make a_strex.c 8158 compile under VC++. 8159 [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 8160 8161 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 8162 length if passed a buffer. ASN1_INTEGER_to_BN failed 8163 if passed a NULL BN and its argument was negative. 8164 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 8165 8166 *) Modification to PKCS#7 encoding routines to output definite 8167 length encoding. Since currently the whole structures are in 8168 memory there's not real point in using indefinite length 8169 constructed encoding. However if OpenSSL is compiled with 8170 the flag PKCS7_INDEFINITE_ENCODING the old form is used. 8171 [Steve Henson] 8172 8173 *) Added BIO_vprintf() and BIO_vsnprintf(). 8174 [Richard Levitte] 8175 8176 *) Added more prefixes to parse for in the the strings written 8177 through a logging bio, to cover all the levels that are available 8178 through syslog. The prefixes are now: 8179 8180 PANIC, EMERG, EMR => LOG_EMERG 8181 ALERT, ALR => LOG_ALERT 8182 CRIT, CRI => LOG_CRIT 8183 ERROR, ERR => LOG_ERR 8184 WARNING, WARN, WAR => LOG_WARNING 8185 NOTICE, NOTE, NOT => LOG_NOTICE 8186 INFO, INF => LOG_INFO 8187 DEBUG, DBG => LOG_DEBUG 8188 8189 and as before, if none of those prefixes are present at the 8190 beginning of the string, LOG_ERR is chosen. 8191 8192 On Win32, the LOG_* levels are mapped according to this: 8193 8194 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 8195 LOG_WARNING => EVENTLOG_WARNING_TYPE 8196 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 8197 8198 [Richard Levitte] 8199 8200 *) Made it possible to reconfigure with just the configuration 8201 argument "reconf" or "reconfigure". The command line arguments 8202 are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 8203 and are retrieved from there when reconfiguring. 8204 [Richard Levitte] 8205 8206 *) MD4 implemented. 8207 [Assar Westerlund <assar@sics.se>, Richard Levitte] 8208 8209 *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 8210 [Richard Levitte] 8211 8212 *) The obj_dat.pl script was messing up the sorting of object 8213 names. The reason was that it compared the quoted version 8214 of strings as a result "OCSP" > "OCSP Signing" because 8215 " > SPACE. Changed script to store unquoted versions of 8216 names and add quotes on output. It was also omitting some 8217 names from the lookup table if they were given a default 8218 value (that is if SN is missing it is given the same 8219 value as LN and vice versa), these are now added on the 8220 grounds that if an object has a name we should be able to 8221 look it up. Finally added warning output when duplicate 8222 short or long names are found. 8223 [Steve Henson] 8224 8225 *) Changes needed for Tandem NSK. 8226 [Scott Uroff <scott@xypro.com>] 8227 8228 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 8229 RSA_padding_check_SSLv23(), special padding was never detected 8230 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 8231 version rollback attacks was not effective. 8232 8233 In s23_clnt.c, don't use special rollback-attack detection padding 8234 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 8235 client; similarly, in s23_srvr.c, don't do the rollback check if 8236 SSL 2.0 is the only protocol enabled in the server. 8237 [Bodo Moeller] 8238 8239 *) Make it possible to get hexdumps of unprintable data with 'openssl 8240 asn1parse'. By implication, the functions ASN1_parse_dump() and 8241 BIO_dump_indent() are added. 8242 [Richard Levitte] 8243 8244 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 8245 these print out strings and name structures based on various 8246 flags including RFC2253 support and proper handling of 8247 multibyte characters. Added options to the 'x509' utility 8248 to allow the various flags to be set. 8249 [Steve Henson] 8250 8251 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 8252 Also change the functions X509_cmp_current_time() and 8253 X509_gmtime_adj() work with an ASN1_TIME structure, 8254 this will enable certificates using GeneralizedTime in validity 8255 dates to be checked. 8256 [Steve Henson] 8257 8258 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 8259 negative public key encodings) on by default, 8260 NO_NEG_PUBKEY_BUG can be set to disable it. 8261 [Steve Henson] 8262 8263 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 8264 content octets. An i2c_ASN1_OBJECT is unnecessary because 8265 the encoding can be trivially obtained from the structure. 8266 [Steve Henson] 8267 8268 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 8269 not read locks (CRYPTO_r_[un]lock). 8270 [Bodo Moeller] 8271 8272 *) A first attempt at creating official support for shared 8273 libraries through configuration. I've kept it so the 8274 default is static libraries only, and the OpenSSL programs 8275 are always statically linked for now, but there are 8276 preparations for dynamic linking in place. 8277 This has been tested on Linux and Tru64. 8278 [Richard Levitte] 8279 8280 *) Randomness polling function for Win9x, as described in: 8281 Peter Gutmann, Software Generation of Practically Strong 8282 Random Numbers. 8283 [Ulf M��ller] 8284 8285 *) Fix so PRNG is seeded in req if using an already existing 8286 DSA key. 8287 [Steve Henson] 8288 8289 *) New options to smime application. -inform and -outform 8290 allow alternative formats for the S/MIME message including 8291 PEM and DER. The -content option allows the content to be 8292 specified separately. This should allow things like Netscape 8293 form signing output easier to verify. 8294 [Steve Henson] 8295 8296 *) Fix the ASN1 encoding of tags using the 'long form'. 8297 [Steve Henson] 8298 8299 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 8300 STRING types. These convert content octets to and from the 8301 underlying type. The actual tag and length octets are 8302 already assumed to have been read in and checked. These 8303 are needed because all other string types have virtually 8304 identical handling apart from the tag. By having versions 8305 of the ASN1 functions that just operate on content octets 8306 IMPLICIT tagging can be handled properly. It also allows 8307 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 8308 and ASN1_INTEGER are identical apart from the tag. 8309 [Steve Henson] 8310 8311 *) Change the handling of OID objects as follows: 8312 8313 - New object identifiers are inserted in objects.txt, following 8314 the syntax given in objects.README. 8315 - objects.pl is used to process obj_mac.num and create a new 8316 obj_mac.h. 8317 - obj_dat.pl is used to create a new obj_dat.h, using the data in 8318 obj_mac.h. 8319 8320 This is currently kind of a hack, and the perl code in objects.pl 8321 isn't very elegant, but it works as I intended. The simplest way 8322 to check that it worked correctly is to look in obj_dat.h and 8323 check the array nid_objs and make sure the objects haven't moved 8324 around (this is important!). Additions are OK, as well as 8325 consistent name changes. 8326 [Richard Levitte] 8327 8328 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 8329 [Bodo Moeller] 8330 8331 *) Addition of the command line parameter '-rand file' to 'openssl req'. 8332 The given file adds to whatever has already been seeded into the 8333 random pool through the RANDFILE configuration file option or 8334 environment variable, or the default random state file. 8335 [Richard Levitte] 8336 8337 *) mkstack.pl now sorts each macro group into lexical order. 8338 Previously the output order depended on the order the files 8339 appeared in the directory, resulting in needless rewriting 8340 of safestack.h . 8341 [Steve Henson] 8342 8343 *) Patches to make OpenSSL compile under Win32 again. Mostly 8344 work arounds for the VC++ problem that it treats func() as 8345 func(void). Also stripped out the parts of mkdef.pl that 8346 added extra typesafe functions: these no longer exist. 8347 [Steve Henson] 8348 8349 *) Reorganisation of the stack code. The macros are now all 8350 collected in safestack.h . Each macro is defined in terms of 8351 a "stack macro" of the form SKM_<name>(type, a, b). The 8352 DEBUG_SAFESTACK is now handled in terms of function casts, 8353 this has the advantage of retaining type safety without the 8354 use of additional functions. If DEBUG_SAFESTACK is not defined 8355 then the non typesafe macros are used instead. Also modified the 8356 mkstack.pl script to handle the new form. Needs testing to see 8357 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 8358 the default if no major problems. Similar behaviour for ASN1_SET_OF 8359 and PKCS12_STACK_OF. 8360 [Steve Henson] 8361 8362 *) When some versions of IIS use the 'NET' form of private key the 8363 key derivation algorithm is different. Normally MD5(password) is 8364 used as a 128 bit RC4 key. In the modified case 8365 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 8366 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 8367 as the old Netscape_RSA functions except they have an additional 8368 'sgckey' parameter which uses the modified algorithm. Also added 8369 an -sgckey command line option to the rsa utility. Thanks to 8370 Adrian Peck <bertie@ncipher.com> for posting details of the modified 8371 algorithm to openssl-dev. 8372 [Steve Henson] 8373 8374 *) The evp_local.h macros were using 'c.##kname' which resulted in 8375 invalid expansion on some systems (SCO 5.0.5 for example). 8376 Corrected to 'c.kname'. 8377 [Phillip Porch <root@theporch.com>] 8378 8379 *) New X509_get1_email() and X509_REQ_get1_email() functions that return 8380 a STACK of email addresses from a certificate or request, these look 8381 in the subject name and the subject alternative name extensions and 8382 omit any duplicate addresses. 8383 [Steve Henson] 8384 8385 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 8386 This makes DSA verification about 2 % faster. 8387 [Bodo Moeller] 8388 8389 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 8390 (meaning that now 2^5 values will be precomputed, which is only 4 KB 8391 plus overhead for 1024 bit moduli). 8392 This makes exponentiations about 0.5 % faster for 1024 bit 8393 exponents (as measured by "openssl speed rsa2048"). 8394 [Bodo Moeller] 8395 8396 *) Rename memory handling macros to avoid conflicts with other 8397 software: 8398 Malloc => OPENSSL_malloc 8399 Malloc_locked => OPENSSL_malloc_locked 8400 Realloc => OPENSSL_realloc 8401 Free => OPENSSL_free 8402 [Richard Levitte] 8403 8404 *) New function BN_mod_exp_mont_word for small bases (roughly 15% 8405 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 8406 [Bodo Moeller] 8407 8408 *) CygWin32 support. 8409 [John Jarvie <jjarvie@newsguy.com>] 8410 8411 *) The type-safe stack code has been rejigged. It is now only compiled 8412 in when OpenSSL is configured with the DEBUG_SAFESTACK option and 8413 by default all type-specific stack functions are "#define"d back to 8414 standard stack functions. This results in more streamlined output 8415 but retains the type-safety checking possibilities of the original 8416 approach. 8417 [Geoff Thorpe] 8418 8419 *) The STACK code has been cleaned up, and certain type declarations 8420 that didn't make a lot of sense have been brought in line. This has 8421 also involved a cleanup of sorts in safestack.h to more correctly 8422 map type-safe stack functions onto their plain stack counterparts. 8423 This work has also resulted in a variety of "const"ifications of 8424 lots of the code, especially "_cmp" operations which should normally 8425 be prototyped with "const" parameters anyway. 8426 [Geoff Thorpe] 8427 8428 *) When generating bytes for the first time in md_rand.c, 'stir the pool' 8429 by seeding with STATE_SIZE dummy bytes (with zero entropy count). 8430 (The PRNG state consists of two parts, the large pool 'state' and 'md', 8431 where all of 'md' is used each time the PRNG is used, but 'state' 8432 is used only indexed by a cyclic counter. As entropy may not be 8433 well distributed from the beginning, 'md' is important as a 8434 chaining variable. However, the output function chains only half 8435 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 8436 all of 'md', and seeding with STATE_SIZE dummy bytes will result 8437 in all of 'state' being rewritten, with the new values depending 8438 on virtually all of 'md'. This overcomes the 80 bit limitation.) 8439 [Bodo Moeller] 8440 8441 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 8442 the handshake is continued after ssl_verify_cert_chain(); 8443 otherwise, if SSL_VERIFY_NONE is set, remaining error codes 8444 can lead to 'unexplainable' connection aborts later. 8445 [Bodo Moeller; problem tracked down by Lutz Jaenicke] 8446 8447 *) Major EVP API cipher revision. 8448 Add hooks for extra EVP features. This allows various cipher 8449 parameters to be set in the EVP interface. Support added for variable 8450 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 8451 setting of RC2 and RC5 parameters. 8452 8453 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 8454 ciphers. 8455 8456 Remove lots of duplicated code from the EVP library. For example *every* 8457 cipher init() function handles the 'iv' in the same way according to the 8458 cipher mode. They also all do nothing if the 'key' parameter is NULL and 8459 for CFB and OFB modes they zero ctx->num. 8460 8461 New functionality allows removal of S/MIME code RC2 hack. 8462 8463 Most of the routines have the same form and so can be declared in terms 8464 of macros. 8465 8466 By shifting this to the top level EVP_CipherInit() it can be removed from 8467 all individual ciphers. If the cipher wants to handle IVs or keys 8468 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 8469 flags. 8470 8471 Change lots of functions like EVP_EncryptUpdate() to now return a 8472 value: although software versions of the algorithms cannot fail 8473 any installed hardware versions can. 8474 [Steve Henson] 8475 8476 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 8477 this option is set, tolerate broken clients that send the negotiated 8478 protocol version number instead of the requested protocol version 8479 number. 8480 [Bodo Moeller] 8481 8482 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 8483 i.e. non-zero for export ciphersuites, zero otherwise. 8484 Previous versions had this flag inverted, inconsistent with 8485 rsa_tmp_cb (..._TMP_RSA_CB). 8486 [Bodo Moeller; problem reported by Amit Chopra] 8487 8488 *) Add missing DSA library text string. Work around for some IIS 8489 key files with invalid SEQUENCE encoding. 8490 [Steve Henson] 8491 8492 *) Add a document (doc/standards.txt) that list all kinds of standards 8493 and so on that are implemented in OpenSSL. 8494 [Richard Levitte] 8495 8496 *) Enhance c_rehash script. Old version would mishandle certificates 8497 with the same subject name hash and wouldn't handle CRLs at all. 8498 Added -fingerprint option to crl utility, to support new c_rehash 8499 features. 8500 [Steve Henson] 8501 8502 *) Eliminate non-ANSI declarations in crypto.h and stack.h. 8503 [Ulf M��ller] 8504 8505 *) Fix for SSL server purpose checking. Server checking was 8506 rejecting certificates which had extended key usage present 8507 but no ssl client purpose. 8508 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 8509 8510 *) Make PKCS#12 code work with no password. The PKCS#12 spec 8511 is a little unclear about how a blank password is handled. 8512 Since the password in encoded as a BMPString with terminating 8513 double NULL a zero length password would end up as just the 8514 double NULL. However no password at all is different and is 8515 handled differently in the PKCS#12 key generation code. NS 8516 treats a blank password as zero length. MSIE treats it as no 8517 password on export: but it will try both on import. We now do 8518 the same: PKCS12_parse() tries zero length and no password if 8519 the password is set to "" or NULL (NULL is now a valid password: 8520 it wasn't before) as does the pkcs12 application. 8521 [Steve Henson] 8522 8523 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 8524 perror when PEM_read_bio_X509_REQ fails, the error message must 8525 be obtained from the error queue. 8526 [Bodo Moeller] 8527 8528 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 8529 it in ERR_remove_state if appropriate, and change ERR_get_state 8530 accordingly to avoid race conditions (this is necessary because 8531 thread_hash is no longer constant once set). 8532 [Bodo Moeller] 8533 8534 *) Bugfix for linux-elf makefile.one. 8535 [Ulf M��ller] 8536 8537 *) RSA_get_default_method() will now cause a default 8538 RSA_METHOD to be chosen if one doesn't exist already. 8539 Previously this was only set during a call to RSA_new() 8540 or RSA_new_method(NULL) meaning it was possible for 8541 RSA_get_default_method() to return NULL. 8542 [Geoff Thorpe] 8543 8544 *) Added native name translation to the existing DSO code 8545 that will convert (if the flag to do so is set) filenames 8546 that are sufficiently small and have no path information 8547 into a canonical native form. Eg. "blah" converted to 8548 "libblah.so" or "blah.dll" etc. 8549 [Geoff Thorpe] 8550 8551 *) New function ERR_error_string_n(e, buf, len) which is like 8552 ERR_error_string(e, buf), but writes at most 'len' bytes 8553 including the 0 terminator. For ERR_error_string_n, 'buf' 8554 may not be NULL. 8555 [Damien Miller <djm@mindrot.org>, Bodo Moeller] 8556 8557 *) CONF library reworked to become more general. A new CONF 8558 configuration file reader "class" is implemented as well as a 8559 new functions (NCONF_*, for "New CONF") to handle it. The now 8560 old CONF_* functions are still there, but are reimplemented to 8561 work in terms of the new functions. Also, a set of functions 8562 to handle the internal storage of the configuration data is 8563 provided to make it easier to write new configuration file 8564 reader "classes" (I can definitely see something reading a 8565 configuration file in XML format, for example), called _CONF_*, 8566 or "the configuration storage API"... 8567 8568 The new configuration file reading functions are: 8569 8570 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 8571 NCONF_get_section, NCONF_get_string, NCONF_get_numbre 8572 8573 NCONF_default, NCONF_WIN32 8574 8575 NCONF_dump_fp, NCONF_dump_bio 8576 8577 NCONF_default and NCONF_WIN32 are method (or "class") choosers, 8578 NCONF_new creates a new CONF object. This works in the same way 8579 as other interfaces in OpenSSL, like the BIO interface. 8580 NCONF_dump_* dump the internal storage of the configuration file, 8581 which is useful for debugging. All other functions take the same 8582 arguments as the old CONF_* functions wth the exception of the 8583 first that must be a `CONF *' instead of a `LHASH *'. 8584 8585 To make it easer to use the new classes with the old CONF_* functions, 8586 the function CONF_set_default_method is provided. 8587 [Richard Levitte] 8588 8589 *) Add '-tls1' option to 'openssl ciphers', which was already 8590 mentioned in the documentation but had not been implemented. 8591 (This option is not yet really useful because even the additional 8592 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 8593 [Bodo Moeller] 8594 8595 *) Initial DSO code added into libcrypto for letting OpenSSL (and 8596 OpenSSL-based applications) load shared libraries and bind to 8597 them in a portable way. 8598 [Geoff Thorpe, with contributions from Richard Levitte] 8599 8600 Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 8601 8602 *) Make sure _lrotl and _lrotr are only used with MSVC. 8603 8604 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 8605 (the default implementation of RAND_status). 8606 8607 *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 8608 to '-clrext' (= clear extensions), as intended and documented. 8609 [Bodo Moeller; inconsistency pointed out by Michael Attili 8610 <attili@amaxo.com>] 8611 8612 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 8613 was larger than the MD block size. 8614 [Steve Henson, pointed out by Yost William <YostW@tce.com>] 8615 8616 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 8617 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 8618 using the passed key: if the passed key was a private key the result 8619 of X509_print(), for example, would be to print out all the private key 8620 components. 8621 [Steve Henson] 8622 8623 *) des_quad_cksum() byte order bug fix. 8624 [Ulf M��ller, using the problem description in krb4-0.9.7, where 8625 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 8626 8627 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 8628 discouraged. 8629 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 8630 8631 *) For easily testing in shell scripts whether some command 8632 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 8633 returns with exit code 0 iff no command of the given name is available. 8634 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 8635 the output goes to stdout and nothing is printed to stderr. 8636 Additional arguments are always ignored. 8637 8638 Since for each cipher there is a command of the same name, 8639 the 'no-cipher' compilation switches can be tested this way. 8640 8641 ('openssl no-XXX' is not able to detect pseudo-commands such 8642 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 8643 [Bodo Moeller] 8644 8645 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 8646 [Bodo Moeller] 8647 8648 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 8649 is set; it will be thrown away anyway because each handshake creates 8650 its own key. 8651 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 8652 to parameters -- in previous versions (since OpenSSL 0.9.3) the 8653 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 8654 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 8655 [Bodo Moeller] 8656 8657 *) New s_client option -ign_eof: EOF at stdin is ignored, and 8658 'Q' and 'R' lose their special meanings (quit/renegotiate). 8659 This is part of what -quiet does; unlike -quiet, -ign_eof 8660 does not suppress any output. 8661 [Richard Levitte] 8662 8663 *) Add compatibility options to the purpose and trust code. The 8664 purpose X509_PURPOSE_ANY is "any purpose" which automatically 8665 accepts a certificate or CA, this was the previous behaviour, 8666 with all the associated security issues. 8667 8668 X509_TRUST_COMPAT is the old trust behaviour: only and 8669 automatically trust self signed roots in certificate store. A 8670 new trust setting X509_TRUST_DEFAULT is used to specify that 8671 a purpose has no associated trust setting and it should instead 8672 use the value in the default purpose. 8673 [Steve Henson] 8674 8675 *) Fix the PKCS#8 DSA private key code so it decodes keys again 8676 and fix a memory leak. 8677 [Steve Henson] 8678 8679 *) In util/mkerr.pl (which implements 'make errors'), preserve 8680 reason strings from the previous version of the .c file, as 8681 the default to have only downcase letters (and digits) in 8682 automatically generated reasons codes is not always appropriate. 8683 [Bodo Moeller] 8684 8685 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 8686 using strerror. Previously, ERR_reason_error_string() returned 8687 library names as reason strings for SYSerr; but SYSerr is a special 8688 case where small numbers are errno values, not library numbers. 8689 [Bodo Moeller] 8690 8691 *) Add '-dsaparam' option to 'openssl dhparam' application. This 8692 converts DSA parameters into DH parameters. (When creating parameters, 8693 DSA_generate_parameters is used.) 8694 [Bodo Moeller] 8695 8696 *) Include 'length' (recommended exponent length) in C code generated 8697 by 'openssl dhparam -C'. 8698 [Bodo Moeller] 8699 8700 *) The second argument to set_label in perlasm was already being used 8701 so couldn't be used as a "file scope" flag. Moved to third argument 8702 which was free. 8703 [Steve Henson] 8704 8705 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 8706 instead of RAND_bytes for encryption IVs and salts. 8707 [Bodo Moeller] 8708 8709 *) Include RAND_status() into RAND_METHOD instead of implementing 8710 it only for md_rand.c Otherwise replacing the PRNG by calling 8711 RAND_set_rand_method would be impossible. 8712 [Bodo Moeller] 8713 8714 *) Don't let DSA_generate_key() enter an infinite loop if the random 8715 number generation fails. 8716 [Bodo Moeller] 8717 8718 *) New 'rand' application for creating pseudo-random output. 8719 [Bodo Moeller] 8720 8721 *) Added configuration support for Linux/IA64 8722 [Rolf Haberrecker <rolf@suse.de>] 8723 8724 *) Assembler module support for Mingw32. 8725 [Ulf M��ller] 8726 8727 *) Shared library support for HPUX (in shlib/). 8728 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 8729 8730 *) Shared library support for Solaris gcc. 8731 [Lutz Behnke <behnke@trustcenter.de>] 8732 8733 Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 8734 8735 *) PKCS7_encrypt() was adding text MIME headers twice because they 8736 were added manually and by SMIME_crlf_copy(). 8737 [Steve Henson] 8738 8739 *) In bntest.c don't call BN_rand with zero bits argument. 8740 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 8741 8742 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 8743 case was implemented. This caused BN_div_recp() to fail occasionally. 8744 [Ulf M��ller] 8745 8746 *) Add an optional second argument to the set_label() in the perl 8747 assembly language builder. If this argument exists and is set 8748 to 1 it signals that the assembler should use a symbol whose 8749 scope is the entire file, not just the current function. This 8750 is needed with MASM which uses the format label:: for this scope. 8751 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 8752 8753 *) Change the ASN1 types so they are typedefs by default. Before 8754 almost all types were #define'd to ASN1_STRING which was causing 8755 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 8756 for example. 8757 [Steve Henson] 8758 8759 *) Change names of new functions to the new get1/get0 naming 8760 convention: After 'get1', the caller owns a reference count 8761 and has to call ..._free; 'get0' returns a pointer to some 8762 data structure without incrementing reference counters. 8763 (Some of the existing 'get' functions increment a reference 8764 counter, some don't.) 8765 Similarly, 'set1' and 'add1' functions increase reference 8766 counters or duplicate objects. 8767 [Steve Henson] 8768 8769 *) Allow for the possibility of temp RSA key generation failure: 8770 the code used to assume it always worked and crashed on failure. 8771 [Steve Henson] 8772 8773 *) Fix potential buffer overrun problem in BIO_printf(). 8774 [Ulf M��ller, using public domain code by Patrick Powell; problem 8775 pointed out by David Sacerdote <das33@cornell.edu>] 8776 8777 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 8778 RAND_egd() and RAND_status(). In the command line application, 8779 the EGD socket can be specified like a seed file using RANDFILE 8780 or -rand. 8781 [Ulf M��ller] 8782 8783 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 8784 Some CAs (e.g. Verisign) distribute certificates in this form. 8785 [Steve Henson] 8786 8787 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 8788 list to exclude them. This means that no special compilation option 8789 is needed to use anonymous DH: it just needs to be included in the 8790 cipher list. 8791 [Steve Henson] 8792 8793 *) Change the EVP_MD_CTX_type macro so its meaning consistent with 8794 EVP_MD_type. The old functionality is available in a new macro called 8795 EVP_MD_md(). Change code that uses it and update docs. 8796 [Steve Henson] 8797 8798 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 8799 where the 'void *' argument is replaced by a function pointer argument. 8800 Previously 'void *' was abused to point to functions, which works on 8801 many platforms, but is not correct. As these functions are usually 8802 called by macros defined in OpenSSL header files, most source code 8803 should work without changes. 8804 [Richard Levitte] 8805 8806 *) <openssl/opensslconf.h> (which is created by Configure) now contains 8807 sections with information on -D... compiler switches used for 8808 compiling the library so that applications can see them. To enable 8809 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 8810 must be defined. E.g., 8811 #define OPENSSL_ALGORITHM_DEFINES 8812 #include <openssl/opensslconf.h> 8813 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 8814 [Richard Levitte, Ulf and Bodo M��ller] 8815 8816 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 8817 record layer. 8818 [Bodo Moeller] 8819 8820 *) Change the 'other' type in certificate aux info to a STACK_OF 8821 X509_ALGOR. Although not an AlgorithmIdentifier as such it has 8822 the required ASN1 format: arbitrary types determined by an OID. 8823 [Steve Henson] 8824 8825 *) Add some PEM_write_X509_REQ_NEW() functions and a command line 8826 argument to 'req'. This is not because the function is newer or 8827 better than others it just uses the work 'NEW' in the certificate 8828 request header lines. Some software needs this. 8829 [Steve Henson] 8830 8831 *) Reorganise password command line arguments: now passwords can be 8832 obtained from various sources. Delete the PEM_cb function and make 8833 it the default behaviour: i.e. if the callback is NULL and the 8834 usrdata argument is not NULL interpret it as a null terminated pass 8835 phrase. If usrdata and the callback are NULL then the pass phrase 8836 is prompted for as usual. 8837 [Steve Henson] 8838 8839 *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 8840 the support is automatically enabled. The resulting binaries will 8841 autodetect the card and use it if present. 8842 [Ben Laurie and Compaq Inc.] 8843 8844 *) Work around for Netscape hang bug. This sends certificate request 8845 and server done in one record. Since this is perfectly legal in the 8846 SSL/TLS protocol it isn't a "bug" option and is on by default. See 8847 the bugs/SSLv3 entry for more info. 8848 [Steve Henson] 8849 8850 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 8851 [Andy Polyakov] 8852 8853 *) Add -rand argument to smime and pkcs12 applications and read/write 8854 of seed file. 8855 [Steve Henson] 8856 8857 *) New 'passwd' tool for crypt(3) and apr1 password hashes. 8858 [Bodo Moeller] 8859 8860 *) Add command line password options to the remaining applications. 8861 [Steve Henson] 8862 8863 *) Bug fix for BN_div_recp() for numerators with an even number of 8864 bits. 8865 [Ulf M��ller] 8866 8867 *) More tests in bntest.c, and changed test_bn output. 8868 [Ulf M��ller] 8869 8870 *) ./config recognizes MacOS X now. 8871 [Andy Polyakov] 8872 8873 *) Bug fix for BN_div() when the first words of num and divsor are 8874 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 8875 [Ulf M��ller] 8876 8877 *) Add support for various broken PKCS#8 formats, and command line 8878 options to produce them. 8879 [Steve Henson] 8880 8881 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 8882 get temporary BIGNUMs from a BN_CTX. 8883 [Ulf M��ller] 8884 8885 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 8886 for p == 0. 8887 [Ulf M��ller] 8888 8889 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 8890 include a #define from the old name to the new. The original intent 8891 was that statically linked binaries could for example just call 8892 SSLeay_add_all_ciphers() to just add ciphers to the table and not 8893 link with digests. This never worked becayse SSLeay_add_all_digests() 8894 and SSLeay_add_all_ciphers() were in the same source file so calling 8895 one would link with the other. They are now in separate source files. 8896 [Steve Henson] 8897 8898 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 8899 [Steve Henson] 8900 8901 *) Use a less unusual form of the Miller-Rabin primality test (it used 8902 a binary algorithm for exponentiation integrated into the Miller-Rabin 8903 loop, our standard modexp algorithms are faster). 8904 [Bodo Moeller] 8905 8906 *) Support for the EBCDIC character set completed. 8907 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 8908 8909 *) Source code cleanups: use const where appropriate, eliminate casts, 8910 use void * instead of char * in lhash. 8911 [Ulf M��ller] 8912 8913 *) Bugfix: ssl3_send_server_key_exchange was not restartable 8914 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 8915 this the server could overwrite ephemeral keys that the client 8916 has already seen). 8917 [Bodo Moeller] 8918 8919 *) Turn DSA_is_prime into a macro that calls BN_is_prime, 8920 using 50 iterations of the Rabin-Miller test. 8921 8922 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 8923 iterations of the Rabin-Miller test as required by the appendix 8924 to FIPS PUB 186[-1]) instead of DSA_is_prime. 8925 As BN_is_prime_fasttest includes trial division, DSA parameter 8926 generation becomes much faster. 8927 8928 This implies a change for the callback functions in DSA_is_prime 8929 and DSA_generate_parameters: The callback function is called once 8930 for each positive witness in the Rabin-Miller test, not just 8931 occasionally in the inner loop; and the parameters to the 8932 callback function now provide an iteration count for the outer 8933 loop rather than for the current invocation of the inner loop. 8934 DSA_generate_parameters additionally can call the callback 8935 function with an 'iteration count' of -1, meaning that a 8936 candidate has passed the trial division test (when q is generated 8937 from an application-provided seed, trial division is skipped). 8938 [Bodo Moeller] 8939 8940 *) New function BN_is_prime_fasttest that optionally does trial 8941 division before starting the Rabin-Miller test and has 8942 an additional BN_CTX * argument (whereas BN_is_prime always 8943 has to allocate at least one BN_CTX). 8944 'callback(1, -1, cb_arg)' is called when a number has passed the 8945 trial division stage. 8946 [Bodo Moeller] 8947 8948 *) Fix for bug in CRL encoding. The validity dates weren't being handled 8949 as ASN1_TIME. 8950 [Steve Henson] 8951 8952 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 8953 [Steve Henson] 8954 8955 *) New function BN_pseudo_rand(). 8956 [Ulf M��ller] 8957 8958 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 8959 bignum version of BN_from_montgomery() with the working code from 8960 SSLeay 0.9.0 (the word based version is faster anyway), and clean up 8961 the comments. 8962 [Ulf M��ller] 8963 8964 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 8965 made it impossible to use the same SSL_SESSION data structure in 8966 SSL2 clients in multiple threads. 8967 [Bodo Moeller] 8968 8969 *) The return value of RAND_load_file() no longer counts bytes obtained 8970 by stat(). RAND_load_file(..., -1) is new and uses the complete file 8971 to seed the PRNG (previously an explicit byte count was required). 8972 [Ulf M��ller, Bodo M��ller] 8973 8974 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 8975 used (char *) instead of (void *) and had casts all over the place. 8976 [Steve Henson] 8977 8978 *) Make BN_generate_prime() return NULL on error if ret!=NULL. 8979 [Ulf M��ller] 8980 8981 *) Retain source code compatibility for BN_prime_checks macro: 8982 BN_is_prime(..., BN_prime_checks, ...) now uses 8983 BN_prime_checks_for_size to determine the appropriate number of 8984 Rabin-Miller iterations. 8985 [Ulf M��ller] 8986 8987 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 8988 DH_CHECK_P_NOT_SAFE_PRIME. 8989 (Check if this is true? OpenPGP calls them "strong".) 8990 [Ulf M��ller] 8991 8992 *) Merge the functionality of "dh" and "gendh" programs into a new program 8993 "dhparam". The old programs are retained for now but will handle DH keys 8994 (instead of parameters) in future. 8995 [Steve Henson] 8996 8997 *) Make the ciphers, s_server and s_client programs check the return values 8998 when a new cipher list is set. 8999 [Steve Henson] 9000 9001 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 9002 ciphers. Before when the 56bit ciphers were enabled the sorting was 9003 wrong. 9004 9005 The syntax for the cipher sorting has been extended to support sorting by 9006 cipher-strength (using the strength_bits hard coded in the tables). 9007 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 9008 9009 Fix a bug in the cipher-command parser: when supplying a cipher command 9010 string with an "undefined" symbol (neither command nor alphanumeric 9011 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 9012 an error is flagged. 9013 9014 Due to the strength-sorting extension, the code of the 9015 ssl_create_cipher_list() function was completely rearranged. I hope that 9016 the readability was also increased :-) 9017 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 9018 9019 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 9020 for the first serial number and places 2 in the serial number file. This 9021 avoids problems when the root CA is created with serial number zero and 9022 the first user certificate has the same issuer name and serial number 9023 as the root CA. 9024 [Steve Henson] 9025 9026 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 9027 the new code. Add documentation for this stuff. 9028 [Steve Henson] 9029 9030 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 9031 X509_*() to X509at_*() on the grounds that they don't handle X509 9032 structures and behave in an analagous way to the X509v3 functions: 9033 they shouldn't be called directly but wrapper functions should be used 9034 instead. 9035 9036 So we also now have some wrapper functions that call the X509at functions 9037 when passed certificate requests. (TO DO: similar things can be done with 9038 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 9039 things. Some of these need some d2i or i2d and print functionality 9040 because they handle more complex structures.) 9041 [Steve Henson] 9042 9043 *) Add missing #ifndefs that caused missing symbols when building libssl 9044 as a shared library without RSA. Use #ifndef NO_SSL2 instead of 9045 NO_RSA in ssl/s2*.c. 9046 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M��ller] 9047 9048 *) Precautions against using the PRNG uninitialized: RAND_bytes() now 9049 has a return value which indicates the quality of the random data 9050 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 9051 error queue. New function RAND_pseudo_bytes() generates output that is 9052 guaranteed to be unique but not unpredictable. RAND_add is like 9053 RAND_seed, but takes an extra argument for an entropy estimate 9054 (RAND_seed always assumes full entropy). 9055 [Ulf M��ller] 9056 9057 *) Do more iterations of Rabin-Miller probable prime test (specifically, 9058 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 9059 instead of only 2 for all lengths; see BN_prime_checks_for_size definition 9060 in crypto/bn/bn_prime.c for the complete table). This guarantees a 9061 false-positive rate of at most 2^-80 for random input. 9062 [Bodo Moeller] 9063 9064 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 9065 [Bodo Moeller] 9066 9067 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 9068 in the 0.9.5 release), this returns the chain 9069 from an X509_CTX structure with a dup of the stack and all 9070 the X509 reference counts upped: so the stack will exist 9071 after X509_CTX_cleanup() has been called. Modify pkcs12.c 9072 to use this. 9073 9074 Also make SSL_SESSION_print() print out the verify return 9075 code. 9076 [Steve Henson] 9077 9078 *) Add manpage for the pkcs12 command. Also change the default 9079 behaviour so MAC iteration counts are used unless the new 9080 -nomaciter option is used. This improves file security and 9081 only older versions of MSIE (4.0 for example) need it. 9082 [Steve Henson] 9083 9084 *) Honor the no-xxx Configure options when creating .DEF files. 9085 [Ulf M��ller] 9086 9087 *) Add PKCS#10 attributes to field table: challengePassword, 9088 unstructuredName and unstructuredAddress. These are taken from 9089 draft PKCS#9 v2.0 but are compatible with v1.2 provided no 9090 international characters are used. 9091 9092 More changes to X509_ATTRIBUTE code: allow the setting of types 9093 based on strings. Remove the 'loc' parameter when adding 9094 attributes because these will be a SET OF encoding which is sorted 9095 in ASN1 order. 9096 [Steve Henson] 9097 9098 *) Initial changes to the 'req' utility to allow request generation 9099 automation. This will allow an application to just generate a template 9100 file containing all the field values and have req construct the 9101 request. 9102 9103 Initial support for X509_ATTRIBUTE handling. Stacks of these are 9104 used all over the place including certificate requests and PKCS#7 9105 structures. They are currently handled manually where necessary with 9106 some primitive wrappers for PKCS#7. The new functions behave in a 9107 manner analogous to the X509 extension functions: they allow 9108 attributes to be looked up by NID and added. 9109 9110 Later something similar to the X509V3 code would be desirable to 9111 automatically handle the encoding, decoding and printing of the 9112 more complex types. The string types like challengePassword can 9113 be handled by the string table functions. 9114 9115 Also modified the multi byte string table handling. Now there is 9116 a 'global mask' which masks out certain types. The table itself 9117 can use the flag STABLE_NO_MASK to ignore the mask setting: this 9118 is useful when for example there is only one permissible type 9119 (as in countryName) and using the mask might result in no valid 9120 types at all. 9121 [Steve Henson] 9122 9123 *) Clean up 'Finished' handling, and add functions SSL_get_finished and 9124 SSL_get_peer_finished to allow applications to obtain the latest 9125 Finished messages sent to the peer or expected from the peer, 9126 respectively. (SSL_get_peer_finished is usually the Finished message 9127 actually received from the peer, otherwise the protocol will be aborted.) 9128 9129 As the Finished message are message digests of the complete handshake 9130 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 9131 be used for external authentication procedures when the authentication 9132 provided by SSL/TLS is not desired or is not enough. 9133 [Bodo Moeller] 9134 9135 *) Enhanced support for Alpha Linux is added. Now ./config checks if 9136 the host supports BWX extension and if Compaq C is present on the 9137 $PATH. Just exploiting of the BWX extension results in 20-30% 9138 performance kick for some algorithms, e.g. DES and RC4 to mention 9139 a couple. Compaq C in turn generates ~20% faster code for MD5 and 9140 SHA1. 9141 [Andy Polyakov] 9142 9143 *) Add support for MS "fast SGC". This is arguably a violation of the 9144 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 9145 weak crypto and after checking the certificate is SGC a second one 9146 with strong crypto. MS SGC stops the first handshake after receiving 9147 the server certificate message and sends a second client hello. Since 9148 a server will typically do all the time consuming operations before 9149 expecting any further messages from the client (server key exchange 9150 is the most expensive) there is little difference between the two. 9151 9152 To get OpenSSL to support MS SGC we have to permit a second client 9153 hello message after we have sent server done. In addition we have to 9154 reset the MAC if we do get this second client hello. 9155 [Steve Henson] 9156 9157 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 9158 if a DER encoded private key is RSA or DSA traditional format. Changed 9159 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 9160 format DER encoded private key. Newer code should use PKCS#8 format which 9161 has the key type encoded in the ASN1 structure. Added DER private key 9162 support to pkcs8 application. 9163 [Steve Henson] 9164 9165 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 9166 ciphersuites has been selected (as required by the SSL 3/TLS 1 9167 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 9168 is set, we interpret this as a request to violate the specification 9169 (the worst that can happen is a handshake failure, and 'correct' 9170 behaviour would result in a handshake failure anyway). 9171 [Bodo Moeller] 9172 9173 *) In SSL_CTX_add_session, take into account that there might be multiple 9174 SSL_SESSION structures with the same session ID (e.g. when two threads 9175 concurrently obtain them from an external cache). 9176 The internal cache can handle only one SSL_SESSION with a given ID, 9177 so if there's a conflict, we now throw out the old one to achieve 9178 consistency. 9179 [Bodo Moeller] 9180 9181 *) Add OIDs for idea and blowfish in CBC mode. This will allow both 9182 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 9183 some routines that use cipher OIDs: some ciphers do not have OIDs 9184 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 9185 example. 9186 [Steve Henson] 9187 9188 *) Simplify the trust setting structure and code. Now we just have 9189 two sequences of OIDs for trusted and rejected settings. These will 9190 typically have values the same as the extended key usage extension 9191 and any application specific purposes. 9192 9193 The trust checking code now has a default behaviour: it will just 9194 check for an object with the same NID as the passed id. Functions can 9195 be provided to override either the default behaviour or the behaviour 9196 for a given id. SSL client, server and email already have functions 9197 in place for compatibility: they check the NID and also return "trusted" 9198 if the certificate is self signed. 9199 [Steve Henson] 9200 9201 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 9202 traditional format into an EVP_PKEY structure. 9203 [Steve Henson] 9204 9205 *) Add a password callback function PEM_cb() which either prompts for 9206 a password if usr_data is NULL or otherwise assumes it is a null 9207 terminated password. Allow passwords to be passed on command line 9208 environment or config files in a few more utilities. 9209 [Steve Henson] 9210 9211 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 9212 keys. Add some short names for PKCS#8 PBE algorithms and allow them 9213 to be specified on the command line for the pkcs8 and pkcs12 utilities. 9214 Update documentation. 9215 [Steve Henson] 9216 9217 *) Support for ASN1 "NULL" type. This could be handled before by using 9218 ASN1_TYPE but there wasn't any function that would try to read a NULL 9219 and produce an error if it couldn't. For compatibility we also have 9220 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 9221 don't allocate anything because they don't need to. 9222 [Steve Henson] 9223 9224 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 9225 for details. 9226 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 9227 9228 *) Rebuild of the memory allocation routines used by OpenSSL code and 9229 possibly others as well. The purpose is to make an interface that 9230 provide hooks so anyone can build a separate set of allocation and 9231 deallocation routines to be used by OpenSSL, for example memory 9232 pool implementations, or something else, which was previously hard 9233 since Malloc(), Realloc() and Free() were defined as macros having 9234 the values malloc, realloc and free, respectively (except for Win32 9235 compilations). The same is provided for memory debugging code. 9236 OpenSSL already comes with functionality to find memory leaks, but 9237 this gives people a chance to debug other memory problems. 9238 9239 With these changes, a new set of functions and macros have appeared: 9240 9241 CRYPTO_set_mem_debug_functions() [F] 9242 CRYPTO_get_mem_debug_functions() [F] 9243 CRYPTO_dbg_set_options() [F] 9244 CRYPTO_dbg_get_options() [F] 9245 CRYPTO_malloc_debug_init() [M] 9246 9247 The memory debug functions are NULL by default, unless the library 9248 is compiled with CRYPTO_MDEBUG or friends is defined. If someone 9249 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 9250 gives the standard debugging functions that come with OpenSSL) or 9251 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 9252 provided by the library user) must be used. When the standard 9253 debugging functions are used, CRYPTO_dbg_set_options can be used to 9254 request additional information: 9255 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 9256 the CRYPTO_MDEBUG_xxx macro when compiling the library. 9257 9258 Also, things like CRYPTO_set_mem_functions will always give the 9259 expected result (the new set of functions is used for allocation 9260 and deallocation) at all times, regardless of platform and compiler 9261 options. 9262 9263 To finish it up, some functions that were never use in any other 9264 way than through macros have a new API and new semantic: 9265 9266 CRYPTO_dbg_malloc() 9267 CRYPTO_dbg_realloc() 9268 CRYPTO_dbg_free() 9269 9270 All macros of value have retained their old syntax. 9271 [Richard Levitte and Bodo Moeller] 9272 9273 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 9274 ordering of SMIMECapabilities wasn't in "strength order" and there 9275 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 9276 algorithm. 9277 [Steve Henson] 9278 9279 *) Some ASN1 types with illegal zero length encoding (INTEGER, 9280 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 9281 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 9282 9283 *) Merge in my S/MIME library for OpenSSL. This provides a simple 9284 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 9285 functionality to handle multipart/signed properly) and a utility 9286 called 'smime' to call all this stuff. This is based on code I 9287 originally wrote for Celo who have kindly allowed it to be 9288 included in OpenSSL. 9289 [Steve Henson] 9290 9291 *) Add variants des_set_key_checked and des_set_key_unchecked of 9292 des_set_key (aka des_key_sched). Global variable des_check_key 9293 decides which of these is called by des_set_key; this way 9294 des_check_key behaves as it always did, but applications and 9295 the library itself, which was buggy for des_check_key == 1, 9296 have a cleaner way to pick the version they need. 9297 [Bodo Moeller] 9298 9299 *) New function PKCS12_newpass() which changes the password of a 9300 PKCS12 structure. 9301 [Steve Henson] 9302 9303 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 9304 dynamic mix. In both cases the ids can be used as an index into the 9305 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 9306 functions so they accept a list of the field values and the 9307 application doesn't need to directly manipulate the X509_TRUST 9308 structure. 9309 [Steve Henson] 9310 9311 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 9312 need initialising. 9313 [Steve Henson] 9314 9315 *) Modify the way the V3 extension code looks up extensions. This now 9316 works in a similar way to the object code: we have some "standard" 9317 extensions in a static table which is searched with OBJ_bsearch() 9318 and the application can add dynamic ones if needed. The file 9319 crypto/x509v3/ext_dat.h now has the info: this file needs to be 9320 updated whenever a new extension is added to the core code and kept 9321 in ext_nid order. There is a simple program 'tabtest.c' which checks 9322 this. New extensions are not added too often so this file can readily 9323 be maintained manually. 9324 9325 There are two big advantages in doing things this way. The extensions 9326 can be looked up immediately and no longer need to be "added" using 9327 X509V3_add_standard_extensions(): this function now does nothing. 9328 [Side note: I get *lots* of email saying the extension code doesn't 9329 work because people forget to call this function] 9330 Also no dynamic allocation is done unless new extensions are added: 9331 so if we don't add custom extensions there is no need to call 9332 X509V3_EXT_cleanup(). 9333 [Steve Henson] 9334 9335 *) Modify enc utility's salting as follows: make salting the default. Add a 9336 magic header, so unsalted files fail gracefully instead of just decrypting 9337 to garbage. This is because not salting is a big security hole, so people 9338 should be discouraged from doing it. 9339 [Ben Laurie] 9340 9341 *) Fixes and enhancements to the 'x509' utility. It allowed a message 9342 digest to be passed on the command line but it only used this 9343 parameter when signing a certificate. Modified so all relevant 9344 operations are affected by the digest parameter including the 9345 -fingerprint and -x509toreq options. Also -x509toreq choked if a 9346 DSA key was used because it didn't fix the digest. 9347 [Steve Henson] 9348 9349 *) Initial certificate chain verify code. Currently tests the untrusted 9350 certificates for consistency with the verify purpose (which is set 9351 when the X509_STORE_CTX structure is set up) and checks the pathlength. 9352 9353 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 9354 this is because it will reject chains with invalid extensions whereas 9355 every previous version of OpenSSL and SSLeay made no checks at all. 9356 9357 Trust code: checks the root CA for the relevant trust settings. Trust 9358 settings have an initial value consistent with the verify purpose: e.g. 9359 if the verify purpose is for SSL client use it expects the CA to be 9360 trusted for SSL client use. However the default value can be changed to 9361 permit custom trust settings: one example of this would be to only trust 9362 certificates from a specific "secure" set of CAs. 9363 9364 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 9365 which should be used for version portability: especially since the 9366 verify structure is likely to change more often now. 9367 9368 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 9369 to set them. If not set then assume SSL clients will verify SSL servers 9370 and vice versa. 9371 9372 Two new options to the verify program: -untrusted allows a set of 9373 untrusted certificates to be passed in and -purpose which sets the 9374 intended purpose of the certificate. If a purpose is set then the 9375 new chain verify code is used to check extension consistency. 9376 [Steve Henson] 9377 9378 *) Support for the authority information access extension. 9379 [Steve Henson] 9380 9381 *) Modify RSA and DSA PEM read routines to transparently handle 9382 PKCS#8 format private keys. New *_PUBKEY_* functions that handle 9383 public keys in a format compatible with certificate 9384 SubjectPublicKeyInfo structures. Unfortunately there were already 9385 functions called *_PublicKey_* which used various odd formats so 9386 these are retained for compatibility: however the DSA variants were 9387 never in a public release so they have been deleted. Changed dsa/rsa 9388 utilities to handle the new format: note no releases ever handled public 9389 keys so we should be OK. 9390 9391 The primary motivation for this change is to avoid the same fiasco 9392 that dogs private keys: there are several incompatible private key 9393 formats some of which are standard and some OpenSSL specific and 9394 require various evil hacks to allow partial transparent handling and 9395 even then it doesn't work with DER formats. Given the option anything 9396 other than PKCS#8 should be dumped: but the other formats have to 9397 stay in the name of compatibility. 9398 9399 With public keys and the benefit of hindsight one standard format 9400 is used which works with EVP_PKEY, RSA or DSA structures: though 9401 it clearly returns an error if you try to read the wrong kind of key. 9402 9403 Added a -pubkey option to the 'x509' utility to output the public key. 9404 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 9405 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 9406 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 9407 that do the same as the EVP_PKEY_assign_*() except they up the 9408 reference count of the added key (they don't "swallow" the 9409 supplied key). 9410 [Steve Henson] 9411 9412 *) Fixes to crypto/x509/by_file.c the code to read in certificates and 9413 CRLs would fail if the file contained no certificates or no CRLs: 9414 added a new function to read in both types and return the number 9415 read: this means that if none are read it will be an error. The 9416 DER versions of the certificate and CRL reader would always fail 9417 because it isn't possible to mix certificates and CRLs in DER format 9418 without choking one or the other routine. Changed this to just read 9419 a certificate: this is the best we can do. Also modified the code 9420 in apps/verify.c to take notice of return codes: it was previously 9421 attempting to read in certificates from NULL pointers and ignoring 9422 any errors: this is one reason why the cert and CRL reader seemed 9423 to work. It doesn't check return codes from the default certificate 9424 routines: these may well fail if the certificates aren't installed. 9425 [Steve Henson] 9426 9427 *) Code to support otherName option in GeneralName. 9428 [Steve Henson] 9429 9430 *) First update to verify code. Change the verify utility 9431 so it warns if it is passed a self signed certificate: 9432 for consistency with the normal behaviour. X509_verify 9433 has been modified to it will now verify a self signed 9434 certificate if *exactly* the same certificate appears 9435 in the store: it was previously impossible to trust a 9436 single self signed certificate. This means that: 9437 openssl verify ss.pem 9438 now gives a warning about a self signed certificate but 9439 openssl verify -CAfile ss.pem ss.pem 9440 is OK. 9441 [Steve Henson] 9442 9443 *) For servers, store verify_result in SSL_SESSION data structure 9444 (and add it to external session representation). 9445 This is needed when client certificate verifications fails, 9446 but an application-provided verification callback (set by 9447 SSL_CTX_set_cert_verify_callback) allows accepting the session 9448 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 9449 but returns 1): When the session is reused, we have to set 9450 ssl->verify_result to the appropriate error code to avoid 9451 security holes. 9452 [Bodo Moeller, problem pointed out by Lutz Jaenicke] 9453 9454 *) Fix a bug in the new PKCS#7 code: it didn't consider the 9455 case in PKCS7_dataInit() where the signed PKCS7 structure 9456 didn't contain any existing data because it was being created. 9457 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 9458 9459 *) Add a salt to the key derivation routines in enc.c. This 9460 forms the first 8 bytes of the encrypted file. Also add a 9461 -S option to allow a salt to be input on the command line. 9462 [Steve Henson] 9463 9464 *) New function X509_cmp(). Oddly enough there wasn't a function 9465 to compare two certificates. We do this by working out the SHA1 9466 hash and comparing that. X509_cmp() will be needed by the trust 9467 code. 9468 [Steve Henson] 9469 9470 *) SSL_get1_session() is like SSL_get_session(), but increments 9471 the reference count in the SSL_SESSION returned. 9472 [Geoff Thorpe <geoff@eu.c2.net>] 9473 9474 *) Fix for 'req': it was adding a null to request attributes. 9475 Also change the X509_LOOKUP and X509_INFO code to handle 9476 certificate auxiliary information. 9477 [Steve Henson] 9478 9479 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 9480 the 'enc' command. 9481 [Steve Henson] 9482 9483 *) Add the possibility to add extra information to the memory leak 9484 detecting output, to form tracebacks, showing from where each 9485 allocation was originated: CRYPTO_push_info("constant string") adds 9486 the string plus current file name and line number to a per-thread 9487 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 9488 is like calling CYRPTO_pop_info() until the stack is empty. 9489 Also updated memory leak detection code to be multi-thread-safe. 9490 [Richard Levitte] 9491 9492 *) Add options -text and -noout to pkcs7 utility and delete the 9493 encryption options which never did anything. Update docs. 9494 [Steve Henson] 9495 9496 *) Add options to some of the utilities to allow the pass phrase 9497 to be included on either the command line (not recommended on 9498 OSes like Unix) or read from the environment. Update the 9499 manpages and fix a few bugs. 9500 [Steve Henson] 9501 9502 *) Add a few manpages for some of the openssl commands. 9503 [Steve Henson] 9504 9505 *) Fix the -revoke option in ca. It was freeing up memory twice, 9506 leaking and not finding already revoked certificates. 9507 [Steve Henson] 9508 9509 *) Extensive changes to support certificate auxiliary information. 9510 This involves the use of X509_CERT_AUX structure and X509_AUX 9511 functions. An X509_AUX function such as PEM_read_X509_AUX() 9512 can still read in a certificate file in the usual way but it 9513 will also read in any additional "auxiliary information". By 9514 doing things this way a fair degree of compatibility can be 9515 retained: existing certificates can have this information added 9516 using the new 'x509' options. 9517 9518 Current auxiliary information includes an "alias" and some trust 9519 settings. The trust settings will ultimately be used in enhanced 9520 certificate chain verification routines: currently a certificate 9521 can only be trusted if it is self signed and then it is trusted 9522 for all purposes. 9523 [Steve Henson] 9524 9525 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 9526 The problem was that one of the replacement routines had not been working 9527 since SSLeay releases. For now the offending routine has been replaced 9528 with non-optimised assembler. Even so, this now gives around 95% 9529 performance improvement for 1024 bit RSA signs. 9530 [Mark Cox] 9531 9532 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 9533 handling. Most clients have the effective key size in bits equal to 9534 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 9535 A few however don't do this and instead use the size of the decrypted key 9536 to determine the RC2 key length and the AlgorithmIdentifier to determine 9537 the effective key length. In this case the effective key length can still 9538 be 40 bits but the key length can be 168 bits for example. This is fixed 9539 by manually forcing an RC2 key into the EVP_PKEY structure because the 9540 EVP code can't currently handle unusual RC2 key sizes: it always assumes 9541 the key length and effective key length are equal. 9542 [Steve Henson] 9543 9544 *) Add a bunch of functions that should simplify the creation of 9545 X509_NAME structures. Now you should be able to do: 9546 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 9547 and have it automatically work out the correct field type and fill in 9548 the structures. The more adventurous can try: 9549 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 9550 and it will (hopefully) work out the correct multibyte encoding. 9551 [Steve Henson] 9552 9553 *) Change the 'req' utility to use the new field handling and multibyte 9554 copy routines. Before the DN field creation was handled in an ad hoc 9555 way in req, ca, and x509 which was rather broken and didn't support 9556 BMPStrings or UTF8Strings. Since some software doesn't implement 9557 BMPStrings or UTF8Strings yet, they can be enabled using the config file 9558 using the dirstring_type option. See the new comment in the default 9559 openssl.cnf for more info. 9560 [Steve Henson] 9561 9562 *) Make crypto/rand/md_rand.c more robust: 9563 - Assure unique random numbers after fork(). 9564 - Make sure that concurrent threads access the global counter and 9565 md serializably so that we never lose entropy in them 9566 or use exactly the same state in multiple threads. 9567 Access to the large state is not always serializable because 9568 the additional locking could be a performance killer, and 9569 md should be large enough anyway. 9570 [Bodo Moeller] 9571 9572 *) New file apps/app_rand.c with commonly needed functionality 9573 for handling the random seed file. 9574 9575 Use the random seed file in some applications that previously did not: 9576 ca, 9577 dsaparam -genkey (which also ignored its '-rand' option), 9578 s_client, 9579 s_server, 9580 x509 (when signing). 9581 Except on systems with /dev/urandom, it is crucial to have a random 9582 seed file at least for key creation, DSA signing, and for DH exchanges; 9583 for RSA signatures we could do without one. 9584 9585 gendh and gendsa (unlike genrsa) used to read only the first byte 9586 of each file listed in the '-rand' option. The function as previously 9587 found in genrsa is now in app_rand.c and is used by all programs 9588 that support '-rand'. 9589 [Bodo Moeller] 9590 9591 *) In RAND_write_file, use mode 0600 for creating files; 9592 don't just chmod when it may be too late. 9593 [Bodo Moeller] 9594 9595 *) Report an error from X509_STORE_load_locations 9596 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 9597 [Bill Perry] 9598 9599 *) New function ASN1_mbstring_copy() this copies a string in either 9600 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 9601 into an ASN1_STRING type. A mask of permissible types is passed 9602 and it chooses the "minimal" type to use or an error if not type 9603 is suitable. 9604 [Steve Henson] 9605 9606 *) Add function equivalents to the various macros in asn1.h. The old 9607 macros are retained with an M_ prefix. Code inside the library can 9608 use the M_ macros. External code (including the openssl utility) 9609 should *NOT* in order to be "shared library friendly". 9610 [Steve Henson] 9611 9612 *) Add various functions that can check a certificate's extensions 9613 to see if it usable for various purposes such as SSL client, 9614 server or S/MIME and CAs of these types. This is currently 9615 VERY EXPERIMENTAL but will ultimately be used for certificate chain 9616 verification. Also added a -purpose flag to x509 utility to 9617 print out all the purposes. 9618 [Steve Henson] 9619 9620 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 9621 functions. 9622 [Steve Henson] 9623 9624 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 9625 for, obtain and decode and extension and obtain its critical flag. 9626 This allows all the necessary extension code to be handled in a 9627 single function call. 9628 [Steve Henson] 9629 9630 *) RC4 tune-up featuring 30-40% performance improvement on most RISC 9631 platforms. See crypto/rc4/rc4_enc.c for further details. 9632 [Andy Polyakov] 9633 9634 *) New -noout option to asn1parse. This causes no output to be produced 9635 its main use is when combined with -strparse and -out to extract data 9636 from a file (which may not be in ASN.1 format). 9637 [Steve Henson] 9638 9639 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 9640 when producing the local key id. 9641 [Richard Levitte <levitte@stacken.kth.se>] 9642 9643 *) New option -dhparam in s_server. This allows a DH parameter file to be 9644 stated explicitly. If it is not stated then it tries the first server 9645 certificate file. The previous behaviour hard coded the filename 9646 "server.pem". 9647 [Steve Henson] 9648 9649 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 9650 a public key to be input or output. For example: 9651 openssl rsa -in key.pem -pubout -out pubkey.pem 9652 Also added necessary DSA public key functions to handle this. 9653 [Steve Henson] 9654 9655 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 9656 in the message. This was handled by allowing 9657 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 9658 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 9659 9660 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 9661 to the end of the strings whereas this didn't. This would cause problems 9662 if strings read with d2i_ASN1_bytes() were later modified. 9663 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 9664 9665 *) Fix for base64 decode bug. When a base64 bio reads only one line of 9666 data and it contains EOF it will end up returning an error. This is 9667 caused by input 46 bytes long. The cause is due to the way base64 9668 BIOs find the start of base64 encoded data. They do this by trying a 9669 trial decode on each line until they find one that works. When they 9670 do a flag is set and it starts again knowing it can pass all the 9671 data directly through the decoder. Unfortunately it doesn't reset 9672 the context it uses. This means that if EOF is reached an attempt 9673 is made to pass two EOFs through the context and this causes the 9674 resulting error. This can also cause other problems as well. As is 9675 usual with these problems it takes *ages* to find and the fix is 9676 trivial: move one line. 9677 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 9678 9679 *) Ugly workaround to get s_client and s_server working under Windows. The 9680 old code wouldn't work because it needed to select() on sockets and the 9681 tty (for keypresses and to see if data could be written). Win32 only 9682 supports select() on sockets so we select() with a 1s timeout on the 9683 sockets and then see if any characters are waiting to be read, if none 9684 are present then we retry, we also assume we can always write data to 9685 the tty. This isn't nice because the code then blocks until we've 9686 received a complete line of data and it is effectively polling the 9687 keyboard at 1s intervals: however it's quite a bit better than not 9688 working at all :-) A dedicated Windows application might handle this 9689 with an event loop for example. 9690 [Steve Henson] 9691 9692 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 9693 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 9694 will be called when RSA_sign() and RSA_verify() are used. This is useful 9695 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 9696 For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 9697 should *not* be used: RSA_sign() and RSA_verify() must be used instead. 9698 This necessitated the support of an extra signature type NID_md5_sha1 9699 for SSL signatures and modifications to the SSL library to use it instead 9700 of calling RSA_public_decrypt() and RSA_private_encrypt(). 9701 [Steve Henson] 9702 9703 *) Add new -verify -CAfile and -CApath options to the crl program, these 9704 will lookup a CRL issuers certificate and verify the signature in a 9705 similar way to the verify program. Tidy up the crl program so it 9706 no longer accesses structures directly. Make the ASN1 CRL parsing a bit 9707 less strict. It will now permit CRL extensions even if it is not 9708 a V2 CRL: this will allow it to tolerate some broken CRLs. 9709 [Steve Henson] 9710 9711 *) Initialize all non-automatic variables each time one of the openssl 9712 sub-programs is started (this is necessary as they may be started 9713 multiple times from the "OpenSSL>" prompt). 9714 [Lennart Bang, Bodo Moeller] 9715 9716 *) Preliminary compilation option RSA_NULL which disables RSA crypto without 9717 removing all other RSA functionality (this is what NO_RSA does). This 9718 is so (for example) those in the US can disable those operations covered 9719 by the RSA patent while allowing storage and parsing of RSA keys and RSA 9720 key generation. 9721 [Steve Henson] 9722 9723 *) Non-copying interface to BIO pairs. 9724 (still largely untested) 9725 [Bodo Moeller] 9726 9727 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 9728 ASCII string. This was handled independently in various places before. 9729 [Steve Henson] 9730 9731 *) New functions UTF8_getc() and UTF8_putc() that parse and generate 9732 UTF8 strings a character at a time. 9733 [Steve Henson] 9734 9735 *) Use client_version from client hello to select the protocol 9736 (s23_srvr.c) and for RSA client key exchange verification 9737 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 9738 [Bodo Moeller] 9739 9740 *) Add various utility functions to handle SPKACs, these were previously 9741 handled by poking round in the structure internals. Added new function 9742 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 9743 print, verify and generate SPKACs. Based on an original idea from 9744 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 9745 [Steve Henson] 9746 9747 *) RIPEMD160 is operational on all platforms and is back in 'make test'. 9748 [Andy Polyakov] 9749 9750 *) Allow the config file extension section to be overwritten on the 9751 command line. Based on an original idea from Massimiliano Pala 9752 <madwolf@comune.modena.it>. The new option is called -extensions 9753 and can be applied to ca, req and x509. Also -reqexts to override 9754 the request extensions in req and -crlexts to override the crl extensions 9755 in ca. 9756 [Steve Henson] 9757 9758 *) Add new feature to the SPKAC handling in ca. Now you can include 9759 the same field multiple times by preceding it by "XXXX." for example: 9760 1.OU="Unit name 1" 9761 2.OU="Unit name 2" 9762 this is the same syntax as used in the req config file. 9763 [Steve Henson] 9764 9765 *) Allow certificate extensions to be added to certificate requests. These 9766 are specified in a 'req_extensions' option of the req section of the 9767 config file. They can be printed out with the -text option to req but 9768 are otherwise ignored at present. 9769 [Steve Henson] 9770 9771 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 9772 data read consists of only the final block it would not decrypted because 9773 EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 9774 A misplaced 'break' also meant the decrypted final block might not be 9775 copied until the next read. 9776 [Steve Henson] 9777 9778 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 9779 a few extra parameters to the DH structure: these will be useful if 9780 for example we want the value of 'q' or implement X9.42 DH. 9781 [Steve Henson] 9782 9783 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 9784 provides hooks that allow the default DSA functions or functions on a 9785 "per key" basis to be replaced. This allows hardware acceleration and 9786 hardware key storage to be handled without major modification to the 9787 library. Also added low level modexp hooks and CRYPTO_EX structure and 9788 associated functions. 9789 [Steve Henson] 9790 9791 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 9792 as "read only": it can't be written to and the buffer it points to will 9793 not be freed. Reading from a read only BIO is much more efficient than 9794 a normal memory BIO. This was added because there are several times when 9795 an area of memory needs to be read from a BIO. The previous method was 9796 to create a memory BIO and write the data to it, this results in two 9797 copies of the data and an O(n^2) reading algorithm. There is a new 9798 function BIO_new_mem_buf() which creates a read only memory BIO from 9799 an area of memory. Also modified the PKCS#7 routines to use read only 9800 memory BIOs. 9801 [Steve Henson] 9802 9803 *) Bugfix: ssl23_get_client_hello did not work properly when called in 9804 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 9805 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 9806 but a retry condition occured while trying to read the rest. 9807 [Bodo Moeller] 9808 9809 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 9810 NID_pkcs7_encrypted by default: this was wrong since this should almost 9811 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 9812 the encrypted data type: this is a more sensible place to put it and it 9813 allows the PKCS#12 code to be tidied up that duplicated this 9814 functionality. 9815 [Steve Henson] 9816 9817 *) Changed obj_dat.pl script so it takes its input and output files on 9818 the command line. This should avoid shell escape redirection problems 9819 under Win32. 9820 [Steve Henson] 9821 9822 *) Initial support for certificate extension requests, these are included 9823 in things like Xenroll certificate requests. Included functions to allow 9824 extensions to be obtained and added. 9825 [Steve Henson] 9826 9827 *) -crlf option to s_client and s_server for sending newlines as 9828 CRLF (as required by many protocols). 9829 [Bodo Moeller] 9830 9831 Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 9832 9833 *) Install libRSAglue.a when OpenSSL is built with RSAref. 9834 [Ralf S. Engelschall] 9835 9836 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 9837 [Andrija Antonijevic <TheAntony2@bigfoot.com>] 9838 9839 *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 9840 program. 9841 [Steve Henson] 9842 9843 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 9844 DH parameters/keys (q is lost during that conversion, but the resulting 9845 DH parameters contain its length). 9846 9847 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 9848 much faster than DH_generate_parameters (which creates parameters 9849 where p = 2*q + 1), and also the smaller q makes DH computations 9850 much more efficient (160-bit exponentiation instead of 1024-bit 9851 exponentiation); so this provides a convenient way to support DHE 9852 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 9853 utter importance to use 9854 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 9855 or 9856 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 9857 when such DH parameters are used, because otherwise small subgroup 9858 attacks may become possible! 9859 [Bodo Moeller] 9860 9861 *) Avoid memory leak in i2d_DHparams. 9862 [Bodo Moeller] 9863 9864 *) Allow the -k option to be used more than once in the enc program: 9865 this allows the same encrypted message to be read by multiple recipients. 9866 [Steve Henson] 9867 9868 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 9869 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 9870 it will always use the numerical form of the OID, even if it has a short 9871 or long name. 9872 [Steve Henson] 9873 9874 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 9875 method only got called if p,q,dmp1,dmq1,iqmp components were present, 9876 otherwise bn_mod_exp was called. In the case of hardware keys for example 9877 no private key components need be present and it might store extra data 9878 in the RSA structure, which cannot be accessed from bn_mod_exp. 9879 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 9880 private key operations. 9881 [Steve Henson] 9882 9883 *) Added support for SPARC Linux. 9884 [Andy Polyakov] 9885 9886 *) pem_password_cb function type incompatibly changed from 9887 typedef int pem_password_cb(char *buf, int size, int rwflag); 9888 to 9889 ....(char *buf, int size, int rwflag, void *userdata); 9890 so that applications can pass data to their callbacks: 9891 The PEM[_ASN1]_{read,write}... functions and macros now take an 9892 additional void * argument, which is just handed through whenever 9893 the password callback is called. 9894 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 9895 9896 New function SSL_CTX_set_default_passwd_cb_userdata. 9897 9898 Compatibility note: As many C implementations push function arguments 9899 onto the stack in reverse order, the new library version is likely to 9900 interoperate with programs that have been compiled with the old 9901 pem_password_cb definition (PEM_whatever takes some data that 9902 happens to be on the stack as its last argument, and the callback 9903 just ignores this garbage); but there is no guarantee whatsoever that 9904 this will work. 9905 9906 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 9907 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 9908 problems not only on Windows, but also on some Unix platforms. 9909 To avoid problematic command lines, these definitions are now in an 9910 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 9911 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 9912 [Bodo Moeller] 9913 9914 *) MIPS III/IV assembler module is reimplemented. 9915 [Andy Polyakov] 9916 9917 *) More DES library cleanups: remove references to srand/rand and 9918 delete an unused file. 9919 [Ulf M��ller] 9920 9921 *) Add support for the the free Netwide assembler (NASM) under Win32, 9922 since not many people have MASM (ml) and it can be hard to obtain. 9923 This is currently experimental but it seems to work OK and pass all 9924 the tests. Check out INSTALL.W32 for info. 9925 [Steve Henson] 9926 9927 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 9928 without temporary keys kept an extra copy of the server key, 9929 and connections with temporary keys did not free everything in case 9930 of an error. 9931 [Bodo Moeller] 9932 9933 *) New function RSA_check_key and new openssl rsa option -check 9934 for verifying the consistency of RSA keys. 9935 [Ulf Moeller, Bodo Moeller] 9936 9937 *) Various changes to make Win32 compile work: 9938 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 9939 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 9940 comparison" warnings. 9941 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 9942 [Steve Henson] 9943 9944 *) Add a debugging option to PKCS#5 v2 key generation function: when 9945 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 9946 derived keys are printed to stderr. 9947 [Steve Henson] 9948 9949 *) Copy the flags in ASN1_STRING_dup(). 9950 [Roman E. Pavlov <pre@mo.msk.ru>] 9951 9952 *) The x509 application mishandled signing requests containing DSA 9953 keys when the signing key was also DSA and the parameters didn't match. 9954 9955 It was supposed to omit the parameters when they matched the signing key: 9956 the verifying software was then supposed to automatically use the CA's 9957 parameters if they were absent from the end user certificate. 9958 9959 Omitting parameters is no longer recommended. The test was also 9960 the wrong way round! This was probably due to unusual behaviour in 9961 EVP_cmp_parameters() which returns 1 if the parameters match. 9962 This meant that parameters were omitted when they *didn't* match and 9963 the certificate was useless. Certificates signed with 'ca' didn't have 9964 this bug. 9965 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 9966 9967 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 9968 The interface is as follows: 9969 Applications can use 9970 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 9971 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 9972 "off" is now the default. 9973 The library internally uses 9974 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 9975 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 9976 to disable memory-checking temporarily. 9977 9978 Some inconsistent states that previously were possible (and were 9979 even the default) are now avoided. 9980 9981 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 9982 with each memory chunk allocated; this is occasionally more helpful 9983 than just having a counter. 9984 9985 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 9986 9987 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 9988 extensions. 9989 [Bodo Moeller] 9990 9991 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 9992 which largely parallels "options", but is for changing API behaviour, 9993 whereas "options" are about protocol behaviour. 9994 Initial "mode" flags are: 9995 9996 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 9997 a single record has been written. 9998 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 9999 retries use the same buffer location. 10000 (But all of the contents must be 10001 copied!) 10002 [Bodo Moeller] 10003 10004 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 10005 worked. 10006 10007 *) Fix problems with no-hmac etc. 10008 [Ulf M��ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 10009 10010 *) New functions RSA_get_default_method(), RSA_set_method() and 10011 RSA_get_method(). These allows replacement of RSA_METHODs without having 10012 to mess around with the internals of an RSA structure. 10013 [Steve Henson] 10014 10015 *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 10016 Also really enable memory leak checks in openssl.c and in some 10017 test programs. 10018 [Chad C. Mulligan, Bodo Moeller] 10019 10020 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 10021 up the length of negative integers. This has now been simplified to just 10022 store the length when it is first determined and use it later, rather 10023 than trying to keep track of where data is copied and updating it to 10024 point to the end. 10025 [Steve Henson, reported by Brien Wheeler 10026 <bwheeler@authentica-security.com>] 10027 10028 *) Add a new function PKCS7_signatureVerify. This allows the verification 10029 of a PKCS#7 signature but with the signing certificate passed to the 10030 function itself. This contrasts with PKCS7_dataVerify which assumes the 10031 certificate is present in the PKCS#7 structure. This isn't always the 10032 case: certificates can be omitted from a PKCS#7 structure and be 10033 distributed by "out of band" means (such as a certificate database). 10034 [Steve Henson] 10035 10036 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 10037 function prototypes in pem.h, also change util/mkdef.pl to add the 10038 necessary function names. 10039 [Steve Henson] 10040 10041 *) mk1mf.pl (used by Windows builds) did not properly read the 10042 options set by Configure in the top level Makefile, and Configure 10043 was not even able to write more than one option correctly. 10044 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 10045 [Bodo Moeller] 10046 10047 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 10048 file to be loaded from a BIO or FILE pointer. The BIO version will 10049 for example allow memory BIOs to contain config info. 10050 [Steve Henson] 10051 10052 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 10053 Whoever hopes to achieve shared-library compatibility across versions 10054 must use this, not the compile-time macro. 10055 (Exercise 0.9.4: Which is the minimum library version required by 10056 such programs?) 10057 Note: All this applies only to multi-threaded programs, others don't 10058 need locks. 10059 [Bodo Moeller] 10060 10061 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 10062 through a BIO pair triggered the default case, i.e. 10063 SSLerr(...,SSL_R_UNKNOWN_STATE). 10064 [Bodo Moeller] 10065 10066 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 10067 can use the SSL library even if none of the specific BIOs is 10068 appropriate. 10069 [Bodo Moeller] 10070 10071 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 10072 for the encoded length. 10073 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 10074 10075 *) Add initial documentation of the X509V3 functions. 10076 [Steve Henson] 10077 10078 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 10079 PEM_write_bio_PKCS8PrivateKey() that are equivalent to 10080 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 10081 secure PKCS#8 private key format with a high iteration count. 10082 [Steve Henson] 10083 10084 *) Fix determination of Perl interpreter: A perl or perl5 10085 _directory_ in $PATH was also accepted as the interpreter. 10086 [Ralf S. Engelschall] 10087 10088 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 10089 wrong with it but it was very old and did things like calling 10090 PEM_ASN1_read() directly and used MD5 for the hash not to mention some 10091 unusual formatting. 10092 [Steve Henson] 10093 10094 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 10095 to use the new extension code. 10096 [Steve Henson] 10097 10098 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 10099 with macros. This should make it easier to change their form, add extra 10100 arguments etc. Fix a few PEM prototypes which didn't have cipher as a 10101 constant. 10102 [Steve Henson] 10103 10104 *) Add to configuration table a new entry that can specify an alternative 10105 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 10106 according to Mark Crispin <MRC@Panda.COM>. 10107 [Bodo Moeller] 10108 10109#if 0 10110 *) DES CBC did not update the IV. Weird. 10111 [Ben Laurie] 10112#else 10113 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 10114 Changing the behaviour of the former might break existing programs -- 10115 where IV updating is needed, des_ncbc_encrypt can be used. 10116#endif 10117 10118 *) When bntest is run from "make test" it drives bc to check its 10119 calculations, as well as internally checking them. If an internal check 10120 fails, it needs to cause bc to give a non-zero result or make test carries 10121 on without noticing the failure. Fixed. 10122 [Ben Laurie] 10123 10124 *) DES library cleanups. 10125 [Ulf M��ller] 10126 10127 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 10128 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 10129 ciphers. NOTE: although the key derivation function has been verified 10130 against some published test vectors it has not been extensively tested 10131 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 10132 of v2.0. 10133 [Steve Henson] 10134 10135 *) Instead of "mkdir -p", which is not fully portable, use new 10136 Perl script "util/mkdir-p.pl". 10137 [Bodo Moeller] 10138 10139 *) Rewrite the way password based encryption (PBE) is handled. It used to 10140 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 10141 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 10142 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 10143 the 'parameter' field of the AlgorithmIdentifier is passed to the 10144 underlying key generation function so it must do its own ASN1 parsing. 10145 This has also changed the EVP_PBE_CipherInit() function which now has a 10146 'parameter' argument instead of literal salt and iteration count values 10147 and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 10148 [Steve Henson] 10149 10150 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 10151 and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 10152 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 10153 KEY" because this clashed with PKCS#8 unencrypted string. Since this 10154 value was just used as a "magic string" and not used directly its 10155 value doesn't matter. 10156 [Steve Henson] 10157 10158 *) Introduce some semblance of const correctness to BN. Shame C doesn't 10159 support mutable. 10160 [Ben Laurie] 10161 10162 *) "linux-sparc64" configuration (ultrapenguin). 10163 [Ray Miller <ray.miller@oucs.ox.ac.uk>] 10164 "linux-sparc" configuration. 10165 [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 10166 10167 *) config now generates no-xxx options for missing ciphers. 10168 [Ulf M��ller] 10169 10170 *) Support the EBCDIC character set (work in progress). 10171 File ebcdic.c not yet included because it has a different license. 10172 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 10173 10174 *) Support BS2000/OSD-POSIX. 10175 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 10176 10177 *) Make callbacks for key generation use void * instead of char *. 10178 [Ben Laurie] 10179 10180 *) Make S/MIME samples compile (not yet tested). 10181 [Ben Laurie] 10182 10183 *) Additional typesafe stacks. 10184 [Ben Laurie] 10185 10186 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 10187 [Bodo Moeller] 10188 10189 10190 Changes between 0.9.3 and 0.9.3a [29 May 1999] 10191 10192 *) New configuration variant "sco5-gcc". 10193 10194 *) Updated some demos. 10195 [Sean O Riordain, Wade Scholine] 10196 10197 *) Add missing BIO_free at exit of pkcs12 application. 10198 [Wu Zhigang] 10199 10200 *) Fix memory leak in conf.c. 10201 [Steve Henson] 10202 10203 *) Updates for Win32 to assembler version of MD5. 10204 [Steve Henson] 10205 10206 *) Set #! path to perl in apps/der_chop to where we found it 10207 instead of using a fixed path. 10208 [Bodo Moeller] 10209 10210 *) SHA library changes for irix64-mips4-cc. 10211 [Andy Polyakov] 10212 10213 *) Improvements for VMS support. 10214 [Richard Levitte] 10215 10216 10217 Changes between 0.9.2b and 0.9.3 [24 May 1999] 10218 10219 *) Bignum library bug fix. IRIX 6 passes "make test" now! 10220 This also avoids the problems with SC4.2 and unpatched SC5. 10221 [Andy Polyakov <appro@fy.chalmers.se>] 10222 10223 *) New functions sk_num, sk_value and sk_set to replace the previous macros. 10224 These are required because of the typesafe stack would otherwise break 10225 existing code. If old code used a structure member which used to be STACK 10226 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 10227 sk_num or sk_value it would produce an error because the num, data members 10228 are not present in STACK_OF. Now it just produces a warning. sk_set 10229 replaces the old method of assigning a value to sk_value 10230 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 10231 that does this will no longer work (and should use sk_set instead) but 10232 this could be regarded as a "questionable" behaviour anyway. 10233 [Steve Henson] 10234 10235 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 10236 correctly handle encrypted S/MIME data. 10237 [Steve Henson] 10238 10239 *) Change type of various DES function arguments from des_cblock 10240 (which means, in function argument declarations, pointer to char) 10241 to des_cblock * (meaning pointer to array with 8 char elements), 10242 which allows the compiler to do more typechecking; it was like 10243 that back in SSLeay, but with lots of ugly casts. 10244 10245 Introduce new type const_des_cblock. 10246 [Bodo Moeller] 10247 10248 *) Reorganise the PKCS#7 library and get rid of some of the more obvious 10249 problems: find RecipientInfo structure that matches recipient certificate 10250 and initialise the ASN1 structures properly based on passed cipher. 10251 [Steve Henson] 10252 10253 *) Belatedly make the BN tests actually check the results. 10254 [Ben Laurie] 10255 10256 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 10257 to and from BNs: it was completely broken. New compilation option 10258 NEG_PUBKEY_BUG to allow for some broken certificates that encode public 10259 key elements as negative integers. 10260 [Steve Henson] 10261 10262 *) Reorganize and speed up MD5. 10263 [Andy Polyakov <appro@fy.chalmers.se>] 10264 10265 *) VMS support. 10266 [Richard Levitte <richard@levitte.org>] 10267 10268 *) New option -out to asn1parse to allow the parsed structure to be 10269 output to a file. This is most useful when combined with the -strparse 10270 option to examine the output of things like OCTET STRINGS. 10271 [Steve Henson] 10272 10273 *) Make SSL library a little more fool-proof by not requiring any longer 10274 that SSL_set_{accept,connect}_state be called before 10275 SSL_{accept,connect} may be used (SSL_set_..._state is omitted 10276 in many applications because usually everything *appeared* to work as 10277 intended anyway -- now it really works as intended). 10278 [Bodo Moeller] 10279 10280 *) Move openssl.cnf out of lib/. 10281 [Ulf M��ller] 10282 10283 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 10284 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 10285 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 10286 [Ralf S. Engelschall] 10287 10288 *) Various fixes to the EVP and PKCS#7 code. It may now be able to 10289 handle PKCS#7 enveloped data properly. 10290 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 10291 10292 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 10293 copying pointers. The cert_st handling is changed by this in 10294 various ways (and thus what used to be known as ctx->default_cert 10295 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 10296 any longer when s->cert does not give us what we need). 10297 ssl_cert_instantiate becomes obsolete by this change. 10298 As soon as we've got the new code right (possibly it already is?), 10299 we have solved a couple of bugs of the earlier code where s->cert 10300 was used as if it could not have been shared with other SSL structures. 10301 10302 Note that using the SSL API in certain dirty ways now will result 10303 in different behaviour than observed with earlier library versions: 10304 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 10305 does not influence s as it used to. 10306 10307 In order to clean up things more thoroughly, inside SSL_SESSION 10308 we don't use CERT any longer, but a new structure SESS_CERT 10309 that holds per-session data (if available); currently, this is 10310 the peer's certificate chain and, for clients, the server's certificate 10311 and temporary key. CERT holds only those values that can have 10312 meaningful defaults in an SSL_CTX. 10313 [Bodo Moeller] 10314 10315 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 10316 from the internal representation. Various PKCS#7 fixes: remove some 10317 evil casts and set the enc_dig_alg field properly based on the signing 10318 key type. 10319 [Steve Henson] 10320 10321 *) Allow PKCS#12 password to be set from the command line or the 10322 environment. Let 'ca' get its config file name from the environment 10323 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 10324 and 'x509'). 10325 [Steve Henson] 10326 10327 *) Allow certificate policies extension to use an IA5STRING for the 10328 organization field. This is contrary to the PKIX definition but 10329 VeriSign uses it and IE5 only recognises this form. Document 'x509' 10330 extension option. 10331 [Steve Henson] 10332 10333 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 10334 without disallowing inline assembler and the like for non-pedantic builds. 10335 [Ben Laurie] 10336 10337 *) Support Borland C++ builder. 10338 [Janez Jere <jj@void.si>, modified by Ulf M��ller] 10339 10340 *) Support Mingw32. 10341 [Ulf M��ller] 10342 10343 *) SHA-1 cleanups and performance enhancements. 10344 [Andy Polyakov <appro@fy.chalmers.se>] 10345 10346 *) Sparc v8plus assembler for the bignum library. 10347 [Andy Polyakov <appro@fy.chalmers.se>] 10348 10349 *) Accept any -xxx and +xxx compiler options in Configure. 10350 [Ulf M��ller] 10351 10352 *) Update HPUX configuration. 10353 [Anonymous] 10354 10355 *) Add missing sk_<type>_unshift() function to safestack.h 10356 [Ralf S. Engelschall] 10357 10358 *) New function SSL_CTX_use_certificate_chain_file that sets the 10359 "extra_cert"s in addition to the certificate. (This makes sense 10360 only for "PEM" format files, as chains as a whole are not 10361 DER-encoded.) 10362 [Bodo Moeller] 10363 10364 *) Support verify_depth from the SSL API. 10365 x509_vfy.c had what can be considered an off-by-one-error: 10366 Its depth (which was not part of the external interface) 10367 was actually counting the number of certificates in a chain; 10368 now it really counts the depth. 10369 [Bodo Moeller] 10370 10371 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 10372 instead of X509err, which often resulted in confusing error 10373 messages since the error codes are not globally unique 10374 (e.g. an alleged error in ssl3_accept when a certificate 10375 didn't match the private key). 10376 10377 *) New function SSL_CTX_set_session_id_context that allows to set a default 10378 value (so that you don't need SSL_set_session_id_context for each 10379 connection using the SSL_CTX). 10380 [Bodo Moeller] 10381 10382 *) OAEP decoding bug fix. 10383 [Ulf M��ller] 10384 10385 *) Support INSTALL_PREFIX for package builders, as proposed by 10386 David Harris. 10387 [Bodo Moeller] 10388 10389 *) New Configure options "threads" and "no-threads". For systems 10390 where the proper compiler options are known (currently Solaris 10391 and Linux), "threads" is the default. 10392 [Bodo Moeller] 10393 10394 *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 10395 [Bodo Moeller] 10396 10397 *) Install various scripts to $(OPENSSLDIR)/misc, not to 10398 $(INSTALLTOP)/bin -- they shouldn't clutter directories 10399 such as /usr/local/bin. 10400 [Bodo Moeller] 10401 10402 *) "make linux-shared" to build shared libraries. 10403 [Niels Poppe <niels@netbox.org>] 10404 10405 *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 10406 [Ulf M��ller] 10407 10408 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 10409 extension adding in x509 utility. 10410 [Steve Henson] 10411 10412 *) Remove NOPROTO sections and error code comments. 10413 [Ulf M��ller] 10414 10415 *) Partial rewrite of the DEF file generator to now parse the ANSI 10416 prototypes. 10417 [Steve Henson] 10418 10419 *) New Configure options --prefix=DIR and --openssldir=DIR. 10420 [Ulf M��ller] 10421 10422 *) Complete rewrite of the error code script(s). It is all now handled 10423 by one script at the top level which handles error code gathering, 10424 header rewriting and C source file generation. It should be much better 10425 than the old method: it now uses a modified version of Ulf's parser to 10426 read the ANSI prototypes in all header files (thus the old K&R definitions 10427 aren't needed for error creation any more) and do a better job of 10428 translating function codes into names. The old 'ASN1 error code imbedded 10429 in a comment' is no longer necessary and it doesn't use .err files which 10430 have now been deleted. Also the error code call doesn't have to appear all 10431 on one line (which resulted in some large lines...). 10432 [Steve Henson] 10433 10434 *) Change #include filenames from <foo.h> to <openssl/foo.h>. 10435 [Bodo Moeller] 10436 10437 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 10438 0 (which usually indicates a closed connection), but continue reading. 10439 [Bodo Moeller] 10440 10441 *) Fix some race conditions. 10442 [Bodo Moeller] 10443 10444 *) Add support for CRL distribution points extension. Add Certificate 10445 Policies and CRL distribution points documentation. 10446 [Steve Henson] 10447 10448 *) Move the autogenerated header file parts to crypto/opensslconf.h. 10449 [Ulf M��ller] 10450 10451 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 10452 8 of keying material. Merlin has also confirmed interop with this fix 10453 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 10454 [Merlin Hughes <merlin@baltimore.ie>] 10455 10456 *) Fix lots of warnings. 10457 [Richard Levitte <levitte@stacken.kth.se>] 10458 10459 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 10460 the directory spec didn't end with a LIST_SEPARATOR_CHAR. 10461 [Richard Levitte <levitte@stacken.kth.se>] 10462 10463 *) Fix problems with sizeof(long) == 8. 10464 [Andy Polyakov <appro@fy.chalmers.se>] 10465 10466 *) Change functions to ANSI C. 10467 [Ulf M��ller] 10468 10469 *) Fix typos in error codes. 10470 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M��ller] 10471 10472 *) Remove defunct assembler files from Configure. 10473 [Ulf M��ller] 10474 10475 *) SPARC v8 assembler BIGNUM implementation. 10476 [Andy Polyakov <appro@fy.chalmers.se>] 10477 10478 *) Support for Certificate Policies extension: both print and set. 10479 Various additions to support the r2i method this uses. 10480 [Steve Henson] 10481 10482 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 10483 return a const string when you are expecting an allocated buffer. 10484 [Ben Laurie] 10485 10486 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 10487 types DirectoryString and DisplayText. 10488 [Steve Henson] 10489 10490 *) Add code to allow r2i extensions to access the configuration database, 10491 add an LHASH database driver and add several ctx helper functions. 10492 [Steve Henson] 10493 10494 *) Fix an evil bug in bn_expand2() which caused various BN functions to 10495 fail when they extended the size of a BIGNUM. 10496 [Steve Henson] 10497 10498 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 10499 support typesafe stack. 10500 [Steve Henson] 10501 10502 *) Fix typo in SSL_[gs]et_options(). 10503 [Nils Frostberg <nils@medcom.se>] 10504 10505 *) Delete various functions and files that belonged to the (now obsolete) 10506 old X509V3 handling code. 10507 [Steve Henson] 10508 10509 *) New Configure option "rsaref". 10510 [Ulf M��ller] 10511 10512 *) Don't auto-generate pem.h. 10513 [Bodo Moeller] 10514 10515 *) Introduce type-safe ASN.1 SETs. 10516 [Ben Laurie] 10517 10518 *) Convert various additional casted stacks to type-safe STACK_OF() variants. 10519 [Ben Laurie, Ralf S. Engelschall, Steve Henson] 10520 10521 *) Introduce type-safe STACKs. This will almost certainly break lots of code 10522 that links with OpenSSL (well at least cause lots of warnings), but fear 10523 not: the conversion is trivial, and it eliminates loads of evil casts. A 10524 few STACKed things have been converted already. Feel free to convert more. 10525 In the fullness of time, I'll do away with the STACK type altogether. 10526 [Ben Laurie] 10527 10528 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 10529 specified in <certfile> by updating the entry in the index.txt file. 10530 This way one no longer has to edit the index.txt file manually for 10531 revoking a certificate. The -revoke option does the gory details now. 10532 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 10533 10534 *) Fix `openssl crl -noout -text' combination where `-noout' killed the 10535 `-text' option at all and this way the `-noout -text' combination was 10536 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 10537 [Ralf S. Engelschall] 10538 10539 *) Make sure a corresponding plain text error message exists for the 10540 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 10541 verify callback function determined that a certificate was revoked. 10542 [Ralf S. Engelschall] 10543 10544 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 10545 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 10546 all available cipers including rc5, which was forgotten until now. 10547 In order to let the testing shell script know which algorithms 10548 are available, a new (up to now undocumented) command 10549 "openssl list-cipher-commands" is used. 10550 [Bodo Moeller] 10551 10552 *) Bugfix: s_client occasionally would sleep in select() when 10553 it should have checked SSL_pending() first. 10554 [Bodo Moeller] 10555 10556 *) New functions DSA_do_sign and DSA_do_verify to provide access to 10557 the raw DSA values prior to ASN.1 encoding. 10558 [Ulf M��ller] 10559 10560 *) Tweaks to Configure 10561 [Niels Poppe <niels@netbox.org>] 10562 10563 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 10564 yet... 10565 [Steve Henson] 10566 10567 *) New variables $(RANLIB) and $(PERL) in the Makefiles. 10568 [Ulf M��ller] 10569 10570 *) New config option to avoid instructions that are illegal on the 80386. 10571 The default code is faster, but requires at least a 486. 10572 [Ulf M��ller] 10573 10574 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 10575 SSL2_SERVER_VERSION (not used at all) macros, which are now the 10576 same as SSL2_VERSION anyway. 10577 [Bodo Moeller] 10578 10579 *) New "-showcerts" option for s_client. 10580 [Bodo Moeller] 10581 10582 *) Still more PKCS#12 integration. Add pkcs12 application to openssl 10583 application. Various cleanups and fixes. 10584 [Steve Henson] 10585 10586 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 10587 modify error routines to work internally. Add error codes and PBE init 10588 to library startup routines. 10589 [Steve Henson] 10590 10591 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 10592 packing functions to asn1 and evp. Changed function names and error 10593 codes along the way. 10594 [Steve Henson] 10595 10596 *) PKCS12 integration: and so it begins... First of several patches to 10597 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 10598 objects to objects.h 10599 [Steve Henson] 10600 10601 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 10602 and display support for Thawte strong extranet extension. 10603 [Steve Henson] 10604 10605 *) Add LinuxPPC support. 10606 [Jeff Dubrule <igor@pobox.org>] 10607 10608 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 10609 bn_div_words in alpha.s. 10610 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 10611 10612 *) Make sure the RSA OAEP test is skipped under -DRSAref because 10613 OAEP isn't supported when OpenSSL is built with RSAref. 10614 [Ulf Moeller <ulf@fitug.de>] 10615 10616 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 10617 so they no longer are missing under -DNOPROTO. 10618 [Soren S. Jorvang <soren@t.dk>] 10619 10620 10621 Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 10622 10623 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 10624 doesn't work when the session is reused. Coming soon! 10625 [Ben Laurie] 10626 10627 *) Fix a security hole, that allows sessions to be reused in the wrong 10628 context thus bypassing client cert protection! All software that uses 10629 client certs and session caches in multiple contexts NEEDS PATCHING to 10630 allow session reuse! A fuller solution is in the works. 10631 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 10632 10633 *) Some more source tree cleanups (removed obsolete files 10634 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 10635 permission on "config" script to be executable) and a fix for the INSTALL 10636 document. 10637 [Ulf Moeller <ulf@fitug.de>] 10638 10639 *) Remove some legacy and erroneous uses of malloc, free instead of 10640 Malloc, Free. 10641 [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 10642 10643 *) Make rsa_oaep_test return non-zero on error. 10644 [Ulf Moeller <ulf@fitug.de>] 10645 10646 *) Add support for native Solaris shared libraries. Configure 10647 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 10648 if someone would make that last step automatic. 10649 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 10650 10651 *) ctx_size was not built with the right compiler during "make links". Fixed. 10652 [Ben Laurie] 10653 10654 *) Change the meaning of 'ALL' in the cipher list. It now means "everything 10655 except NULL ciphers". This means the default cipher list will no longer 10656 enable NULL ciphers. They need to be specifically enabled e.g. with 10657 the string "DEFAULT:eNULL". 10658 [Steve Henson] 10659 10660 *) Fix to RSA private encryption routines: if p < q then it would 10661 occasionally produce an invalid result. This will only happen with 10662 externally generated keys because OpenSSL (and SSLeay) ensure p > q. 10663 [Steve Henson] 10664 10665 *) Be less restrictive and allow also `perl util/perlpath.pl 10666 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 10667 because this way one can also use an interpreter named `perl5' (which is 10668 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 10669 installed as `perl'). 10670 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 10671 10672 *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 10673 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 10674 10675 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 10676 advapi32.lib to Win32 build and change the pem test comparision 10677 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 10678 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 10679 and crypto/des/ede_cbcm_enc.c. 10680 [Steve Henson] 10681 10682 *) DES quad checksum was broken on big-endian architectures. Fixed. 10683 [Ben Laurie] 10684 10685 *) Comment out two functions in bio.h that aren't implemented. Fix up the 10686 Win32 test batch file so it (might) work again. The Win32 test batch file 10687 is horrible: I feel ill.... 10688 [Steve Henson] 10689 10690 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 10691 in e_os.h. Audit of header files to check ANSI and non ANSI 10692 sections: 10 functions were absent from non ANSI section and not exported 10693 from Windows DLLs. Fixed up libeay.num for new functions. 10694 [Steve Henson] 10695 10696 *) Make `openssl version' output lines consistent. 10697 [Ralf S. Engelschall] 10698 10699 *) Fix Win32 symbol export lists for BIO functions: Added 10700 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 10701 to ms/libeay{16,32}.def. 10702 [Ralf S. Engelschall] 10703 10704 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 10705 fine under Unix and passes some trivial tests I've now added. But the 10706 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 10707 added to make sure no one expects that this stuff really works in the 10708 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 10709 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 10710 openssl_bio.xs. 10711 [Ralf S. Engelschall] 10712 10713 *) Fix the generation of two part addresses in perl. 10714 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 10715 10716 *) Add config entry for Linux on MIPS. 10717 [John Tobey <jtobey@channel1.com>] 10718 10719 *) Make links whenever Configure is run, unless we are on Windoze. 10720 [Ben Laurie] 10721 10722 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 10723 Currently only issuerAltName and AuthorityKeyIdentifier make any sense 10724 in CRLs. 10725 [Steve Henson] 10726 10727 *) Add a useful kludge to allow package maintainers to specify compiler and 10728 other platforms details on the command line without having to patch the 10729 Configure script everytime: One now can use ``perl Configure 10730 <id>:<details>'', i.e. platform ids are allowed to have details appended 10731 to them (seperated by colons). This is treated as there would be a static 10732 pre-configured entry in Configure's %table under key <id> with value 10733 <details> and ``perl Configure <id>'' is called. So, when you want to 10734 perform a quick test-compile under FreeBSD 3.1 with pgcc and without 10735 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 10736 now, which overrides the FreeBSD-elf entry on-the-fly. 10737 [Ralf S. Engelschall] 10738 10739 *) Disable new TLS1 ciphersuites by default: they aren't official yet. 10740 [Ben Laurie] 10741 10742 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 10743 on the `perl Configure ...' command line. This way one can compile 10744 OpenSSL libraries with Position Independent Code (PIC) which is needed 10745 for linking it into DSOs. 10746 [Ralf S. Engelschall] 10747 10748 *) Remarkably, export ciphers were totally broken and no-one had noticed! 10749 Fixed. 10750 [Ben Laurie] 10751 10752 *) Cleaned up the LICENSE document: The official contact for any license 10753 questions now is the OpenSSL core team under openssl-core@openssl.org. 10754 And add a paragraph about the dual-license situation to make sure people 10755 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 10756 to the OpenSSL toolkit. 10757 [Ralf S. Engelschall] 10758 10759 *) General source tree makefile cleanups: Made `making xxx in yyy...' 10760 display consistent in the source tree and replaced `/bin/rm' by `rm'. 10761 Additonally cleaned up the `make links' target: Remove unnecessary 10762 semicolons, subsequent redundant removes, inline point.sh into mklink.sh 10763 to speed processing and no longer clutter the display with confusing 10764 stuff. Instead only the actually done links are displayed. 10765 [Ralf S. Engelschall] 10766 10767 *) Permit null encryption ciphersuites, used for authentication only. It used 10768 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 10769 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 10770 encryption. 10771 [Ben Laurie] 10772 10773 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 10774 signed attributes when verifying signatures (this would break them), 10775 the detached data encoding was wrong and public keys obtained using 10776 X509_get_pubkey() weren't freed. 10777 [Steve Henson] 10778 10779 *) Add text documentation for the BUFFER functions. Also added a work around 10780 to a Win95 console bug. This was triggered by the password read stuff: the 10781 last character typed gets carried over to the next fread(). If you were 10782 generating a new cert request using 'req' for example then the last 10783 character of the passphrase would be CR which would then enter the first 10784 field as blank. 10785 [Steve Henson] 10786 10787 *) Added the new `Includes OpenSSL Cryptography Software' button as 10788 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 10789 button and can be used by applications based on OpenSSL to show the 10790 relationship to the OpenSSL project. 10791 [Ralf S. Engelschall] 10792 10793 *) Remove confusing variables in function signatures in files 10794 ssl/ssl_lib.c and ssl/ssl.h. 10795 [Lennart Bong <lob@kulthea.stacken.kth.se>] 10796 10797 *) Don't install bss_file.c under PREFIX/include/ 10798 [Lennart Bong <lob@kulthea.stacken.kth.se>] 10799 10800 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 10801 functions that return function pointers and has support for NT specific 10802 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 10803 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 10804 unsigned to signed types: this was killing the Win32 compile. 10805 [Steve Henson] 10806 10807 *) Add new certificate file to stack functions, 10808 SSL_add_dir_cert_subjects_to_stack() and 10809 SSL_add_file_cert_subjects_to_stack(). These largely supplant 10810 SSL_load_client_CA_file(), and can be used to add multiple certs easily 10811 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 10812 This means that Apache-SSL and similar packages don't have to mess around 10813 to add as many CAs as they want to the preferred list. 10814 [Ben Laurie] 10815 10816 *) Experiment with doxygen documentation. Currently only partially applied to 10817 ssl/ssl_lib.c. 10818 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 10819 openssl.doxy as the configuration file. 10820 [Ben Laurie] 10821 10822 *) Get rid of remaining C++-style comments which strict C compilers hate. 10823 [Ralf S. Engelschall, pointed out by Carlos Amengual] 10824 10825 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 10826 compiled in by default: it has problems with large keys. 10827 [Steve Henson] 10828 10829 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 10830 DH private keys and/or callback functions which directly correspond to 10831 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 10832 is needed for applications which have to configure certificates on a 10833 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 10834 (e.g. s_server). 10835 For the RSA certificate situation is makes no difference, but 10836 for the DSA certificate situation this fixes the "no shared cipher" 10837 problem where the OpenSSL cipher selection procedure failed because the 10838 temporary keys were not overtaken from the context and the API provided 10839 no way to reconfigure them. 10840 The new functions now let applications reconfigure the stuff and they 10841 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 10842 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 10843 non-public-API function ssl_cert_instantiate() is used as a helper 10844 function and also to reduce code redundancy inside ssl_rsa.c. 10845 [Ralf S. Engelschall] 10846 10847 *) Move s_server -dcert and -dkey options out of the undocumented feature 10848 area because they are useful for the DSA situation and should be 10849 recognized by the users. 10850 [Ralf S. Engelschall] 10851 10852 *) Fix the cipher decision scheme for export ciphers: the export bits are 10853 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 10854 SSL_EXP_MASK. So, the original variable has to be used instead of the 10855 already masked variable. 10856 [Richard Levitte <levitte@stacken.kth.se>] 10857 10858 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 10859 [Richard Levitte <levitte@stacken.kth.se>] 10860 10861 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 10862 from `int' to `unsigned int' because it's a length and initialized by 10863 EVP_DigestFinal() which expects an `unsigned int *'. 10864 [Richard Levitte <levitte@stacken.kth.se>] 10865 10866 *) Don't hard-code path to Perl interpreter on shebang line of Configure 10867 script. Instead use the usual Shell->Perl transition trick. 10868 [Ralf S. Engelschall] 10869 10870 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 10871 (in addition to RSA certificates) to match the behaviour of `openssl dsa 10872 -noout -modulus' as it's already the case for `openssl rsa -noout 10873 -modulus'. For RSA the -modulus is the real "modulus" while for DSA 10874 currently the public key is printed (a decision which was already done by 10875 `openssl dsa -modulus' in the past) which serves a similar purpose. 10876 Additionally the NO_RSA no longer completely removes the whole -modulus 10877 option; it now only avoids using the RSA stuff. Same applies to NO_DSA 10878 now, too. 10879 [Ralf S. Engelschall] 10880 10881 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 10882 BIO. See the source (crypto/evp/bio_ok.c) for more info. 10883 [Arne Ansper <arne@ats.cyber.ee>] 10884 10885 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 10886 to be added. Now both 'req' and 'ca' can use new objects defined in the 10887 config file. 10888 [Steve Henson] 10889 10890 *) Add cool BIO that does syslog (or event log on NT). 10891 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 10892 10893 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 10894 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 10895 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 10896 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 10897 [Ben Laurie] 10898 10899 *) Add preliminary config info for new extension code. 10900 [Steve Henson] 10901 10902 *) Make RSA_NO_PADDING really use no padding. 10903 [Ulf Moeller <ulf@fitug.de>] 10904 10905 *) Generate errors when private/public key check is done. 10906 [Ben Laurie] 10907 10908 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 10909 for some CRL extensions and new objects added. 10910 [Steve Henson] 10911 10912 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 10913 key usage extension and fuller support for authority key id. 10914 [Steve Henson] 10915 10916 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 10917 padding method for RSA, which is recommended for new applications in PKCS 10918 #1 v2.0 (RFC 2437, October 1998). 10919 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 10920 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 10921 against Bleichbacher's attack on RSA. 10922 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 10923 Ben Laurie] 10924 10925 *) Updates to the new SSL compression code 10926 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 10927 10928 *) Fix so that the version number in the master secret, when passed 10929 via RSA, checks that if TLS was proposed, but we roll back to SSLv3 10930 (because the server will not accept higher), that the version number 10931 is 0x03,0x01, not 0x03,0x00 10932 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 10933 10934 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 10935 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 10936 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 10937 [Steve Henson] 10938 10939 *) Support for RAW extensions where an arbitrary extension can be 10940 created by including its DER encoding. See apps/openssl.cnf for 10941 an example. 10942 [Steve Henson] 10943 10944 *) Make sure latest Perl versions don't interpret some generated C array 10945 code as Perl array code in the crypto/err/err_genc.pl script. 10946 [Lars Weber <3weber@informatik.uni-hamburg.de>] 10947 10948 *) Modify ms/do_ms.bat to not generate assembly language makefiles since 10949 not many people have the assembler. Various Win32 compilation fixes and 10950 update to the INSTALL.W32 file with (hopefully) more accurate Win32 10951 build instructions. 10952 [Steve Henson] 10953 10954 *) Modify configure script 'Configure' to automatically create crypto/date.h 10955 file under Win32 and also build pem.h from pem.org. New script 10956 util/mkfiles.pl to create the MINFO file on environments that can't do a 10957 'make files': perl util/mkfiles.pl >MINFO should work. 10958 [Steve Henson] 10959 10960 *) Major rework of DES function declarations, in the pursuit of correctness 10961 and purity. As a result, many evil casts evaporated, and some weirdness, 10962 too. You may find this causes warnings in your code. Zapping your evil 10963 casts will probably fix them. Mostly. 10964 [Ben Laurie] 10965 10966 *) Fix for a typo in asn1.h. Bug fix to object creation script 10967 obj_dat.pl. It considered a zero in an object definition to mean 10968 "end of object": none of the objects in objects.h have any zeros 10969 so it wasn't spotted. 10970 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 10971 10972 *) Add support for Triple DES Cipher Block Chaining with Output Feedback 10973 Masking (CBCM). In the absence of test vectors, the best I have been able 10974 to do is check that the decrypt undoes the encrypt, so far. Send me test 10975 vectors if you have them. 10976 [Ben Laurie] 10977 10978 *) Correct calculation of key length for export ciphers (too much space was 10979 allocated for null ciphers). This has not been tested! 10980 [Ben Laurie] 10981 10982 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 10983 message is now correct (it understands "crypto" and "ssl" on its 10984 command line). There is also now an "update" option. This will update 10985 the util/ssleay.num and util/libeay.num files with any new functions. 10986 If you do a: 10987 perl util/mkdef.pl crypto ssl update 10988 it will update them. 10989 [Steve Henson] 10990 10991 *) Overhauled the Perl interface (perl/*): 10992 - ported BN stuff to OpenSSL's different BN library 10993 - made the perl/ source tree CVS-aware 10994 - renamed the package from SSLeay to OpenSSL (the files still contain 10995 their history because I've copied them in the repository) 10996 - removed obsolete files (the test scripts will be replaced 10997 by better Test::Harness variants in the future) 10998 [Ralf S. Engelschall] 10999 11000 *) First cut for a very conservative source tree cleanup: 11001 1. merge various obsolete readme texts into doc/ssleay.txt 11002 where we collect the old documents and readme texts. 11003 2. remove the first part of files where I'm already sure that we no 11004 longer need them because of three reasons: either they are just temporary 11005 files which were left by Eric or they are preserved original files where 11006 I've verified that the diff is also available in the CVS via "cvs diff 11007 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 11008 the crypto/md/ stuff). 11009 [Ralf S. Engelschall] 11010 11011 *) More extension code. Incomplete support for subject and issuer alt 11012 name, issuer and authority key id. Change the i2v function parameters 11013 and add an extra 'crl' parameter in the X509V3_CTX structure: guess 11014 what that's for :-) Fix to ASN1 macro which messed up 11015 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 11016 [Steve Henson] 11017 11018 *) Preliminary support for ENUMERATED type. This is largely copied from the 11019 INTEGER code. 11020 [Steve Henson] 11021 11022 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 11023 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 11024 11025 *) Make sure `make rehash' target really finds the `openssl' program. 11026 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 11027 11028 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 11029 like to hear about it if this slows down other processors. 11030 [Ben Laurie] 11031 11032 *) Add CygWin32 platform information to Configure script. 11033 [Alan Batie <batie@aahz.jf.intel.com>] 11034 11035 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 11036 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 11037 11038 *) New program nseq to manipulate netscape certificate sequences 11039 [Steve Henson] 11040 11041 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 11042 few typos. 11043 [Steve Henson] 11044 11045 *) Fixes to BN code. Previously the default was to define BN_RECURSION 11046 but the BN code had some problems that would cause failures when 11047 doing certificate verification and some other functions. 11048 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 11049 11050 *) Add ASN1 and PEM code to support netscape certificate sequences. 11051 [Steve Henson] 11052 11053 *) Add ASN1 and PEM code to support netscape certificate sequences. 11054 [Steve Henson] 11055 11056 *) Add several PKIX and private extended key usage OIDs. 11057 [Steve Henson] 11058 11059 *) Modify the 'ca' program to handle the new extension code. Modify 11060 openssl.cnf for new extension format, add comments. 11061 [Steve Henson] 11062 11063 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 11064 and add a sample to openssl.cnf so req -x509 now adds appropriate 11065 CA extensions. 11066 [Steve Henson] 11067 11068 *) Continued X509 V3 changes. Add to other makefiles, integrate with the 11069 error code, add initial support to X509_print() and x509 application. 11070 [Steve Henson] 11071 11072 *) Takes a deep breath and start addding X509 V3 extension support code. Add 11073 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 11074 stuff is currently isolated and isn't even compiled yet. 11075 [Steve Henson] 11076 11077 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 11078 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 11079 Removed the versions check from X509 routines when loading extensions: 11080 this allows certain broken certificates that don't set the version 11081 properly to be processed. 11082 [Steve Henson] 11083 11084 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 11085 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 11086 can still be regenerated with "make depend". 11087 [Ben Laurie] 11088 11089 *) Spelling mistake in C version of CAST-128. 11090 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 11091 11092 *) Changes to the error generation code. The perl script err-code.pl 11093 now reads in the old error codes and retains the old numbers, only 11094 adding new ones if necessary. It also only changes the .err files if new 11095 codes are added. The makefiles have been modified to only insert errors 11096 when needed (to avoid needlessly modifying header files). This is done 11097 by only inserting errors if the .err file is newer than the auto generated 11098 C file. To rebuild all the error codes from scratch (the old behaviour) 11099 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 11100 or delete all the .err files. 11101 [Steve Henson] 11102 11103 *) CAST-128 was incorrectly implemented for short keys. The C version has 11104 been fixed, but is untested. The assembler versions are also fixed, but 11105 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 11106 to regenerate it if needed. 11107 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 11108 Hagino <itojun@kame.net>] 11109 11110 *) File was opened incorrectly in randfile.c. 11111 [Ulf M��ller <ulf@fitug.de>] 11112 11113 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 11114 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 11115 GeneralizedTime. ASN1_TIME is the proper type used in certificates et 11116 al: it's just almost always a UTCTime. Note this patch adds new error 11117 codes so do a "make errors" if there are problems. 11118 [Steve Henson] 11119 11120 *) Correct Linux 1 recognition in config. 11121 [Ulf M��ller <ulf@fitug.de>] 11122 11123 *) Remove pointless MD5 hash when using DSA keys in ca. 11124 [Anonymous <nobody@replay.com>] 11125 11126 *) Generate an error if given an empty string as a cert directory. Also 11127 generate an error if handed NULL (previously returned 0 to indicate an 11128 error, but didn't set one). 11129 [Ben Laurie, reported by Anonymous <nobody@replay.com>] 11130 11131 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 11132 [Ben Laurie] 11133 11134 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 11135 parameters. This was causing a warning which killed off the Win32 compile. 11136 [Steve Henson] 11137 11138 *) Remove C++ style comments from crypto/bn/bn_local.h. 11139 [Neil Costigan <neil.costigan@celocom.com>] 11140 11141 *) The function OBJ_txt2nid was broken. It was supposed to return a nid 11142 based on a text string, looking up short and long names and finally 11143 "dot" format. The "dot" format stuff didn't work. Added new function 11144 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 11145 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 11146 OID is not part of the table. 11147 [Steve Henson] 11148 11149 *) Add prototypes to X509 lookup/verify methods, fixing a bug in 11150 X509_LOOKUP_by_alias(). 11151 [Ben Laurie] 11152 11153 *) Sort openssl functions by name. 11154 [Ben Laurie] 11155 11156 *) Get the gendsa program working (hopefully) and add it to app list. Remove 11157 encryption from sample DSA keys (in case anyone is interested the password 11158 was "1234"). 11159 [Steve Henson] 11160 11161 *) Make _all_ *_free functions accept a NULL pointer. 11162 [Frans Heymans <fheymans@isaserver.be>] 11163 11164 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 11165 NULL pointers. 11166 [Anonymous <nobody@replay.com>] 11167 11168 *) s_server should send the CAfile as acceptable CAs, not its own cert. 11169 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 11170 11171 *) Don't blow it for numeric -newkey arguments to apps/req. 11172 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 11173 11174 *) Temp key "for export" tests were wrong in s3_srvr.c. 11175 [Anonymous <nobody@replay.com>] 11176 11177 *) Add prototype for temp key callback functions 11178 SSL_CTX_set_tmp_{rsa,dh}_callback(). 11179 [Ben Laurie] 11180 11181 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 11182 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 11183 [Steve Henson] 11184 11185 *) X509_name_add_entry() freed the wrong thing after an error. 11186 [Arne Ansper <arne@ats.cyber.ee>] 11187 11188 *) rsa_eay.c would attempt to free a NULL context. 11189 [Arne Ansper <arne@ats.cyber.ee>] 11190 11191 *) BIO_s_socket() had a broken should_retry() on Windoze. 11192 [Arne Ansper <arne@ats.cyber.ee>] 11193 11194 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 11195 [Arne Ansper <arne@ats.cyber.ee>] 11196 11197 *) Make sure the already existing X509_STORE->depth variable is initialized 11198 in X509_STORE_new(), but document the fact that this variable is still 11199 unused in the certificate verification process. 11200 [Ralf S. Engelschall] 11201 11202 *) Fix the various library and apps files to free up pkeys obtained from 11203 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 11204 [Steve Henson] 11205 11206 *) Fix reference counting in X509_PUBKEY_get(). This makes 11207 demos/maurice/example2.c work, amongst others, probably. 11208 [Steve Henson and Ben Laurie] 11209 11210 *) First cut of a cleanup for apps/. First the `ssleay' program is now named 11211 `openssl' and second, the shortcut symlinks for the `openssl <command>' 11212 are no longer created. This way we have a single and consistent command 11213 line interface `openssl <command>', similar to `cvs <command>'. 11214 [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 11215 11216 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 11217 BIT STRING wrapper always have zero unused bits. 11218 [Steve Henson] 11219 11220 *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 11221 [Steve Henson] 11222 11223 *) Make the top-level INSTALL documentation easier to understand. 11224 [Paul Sutton] 11225 11226 *) Makefiles updated to exit if an error occurs in a sub-directory 11227 make (including if user presses ^C) [Paul Sutton] 11228 11229 *) Make Montgomery context stuff explicit in RSA data structure. 11230 [Ben Laurie] 11231 11232 *) Fix build order of pem and err to allow for generated pem.h. 11233 [Ben Laurie] 11234 11235 *) Fix renumbering bug in X509_NAME_delete_entry(). 11236 [Ben Laurie] 11237 11238 *) Enhanced the err-ins.pl script so it makes the error library number 11239 global and can add a library name. This is needed for external ASN1 and 11240 other error libraries. 11241 [Steve Henson] 11242 11243 *) Fixed sk_insert which never worked properly. 11244 [Steve Henson] 11245 11246 *) Fix ASN1 macros so they can handle indefinite length construted 11247 EXPLICIT tags. Some non standard certificates use these: they can now 11248 be read in. 11249 [Steve Henson] 11250 11251 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 11252 into a single doc/ssleay.txt bundle. This way the information is still 11253 preserved but no longer messes up this directory. Now it's new room for 11254 the new set of documenation files. 11255 [Ralf S. Engelschall] 11256 11257 *) SETs were incorrectly DER encoded. This was a major pain, because they 11258 shared code with SEQUENCEs, which aren't coded the same. This means that 11259 almost everything to do with SETs or SEQUENCEs has either changed name or 11260 number of arguments. 11261 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 11262 11263 *) Fix test data to work with the above. 11264 [Ben Laurie] 11265 11266 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 11267 was already fixed by Eric for 0.9.1 it seems. 11268 [Ben Laurie - pointed out by Ulf M��ller <ulf@fitug.de>] 11269 11270 *) Autodetect FreeBSD3. 11271 [Ben Laurie] 11272 11273 *) Fix various bugs in Configure. This affects the following platforms: 11274 nextstep 11275 ncr-scde 11276 unixware-2.0 11277 unixware-2.0-pentium 11278 sco5-cc. 11279 [Ben Laurie] 11280 11281 *) Eliminate generated files from CVS. Reorder tests to regenerate files 11282 before they are needed. 11283 [Ben Laurie] 11284 11285 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 11286 [Ben Laurie] 11287 11288 11289 Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 11290 11291 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 11292 changed SSLeay to OpenSSL in version strings. 11293 [Ralf S. Engelschall] 11294 11295 *) Some fixups to the top-level documents. 11296 [Paul Sutton] 11297 11298 *) Fixed the nasty bug where rsaref.h was not found under compile-time 11299 because the symlink to include/ was missing. 11300 [Ralf S. Engelschall] 11301 11302 *) Incorporated the popular no-RSA/DSA-only patches 11303 which allow to compile a RSA-free SSLeay. 11304 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 11305 11306 *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 11307 when "ssleay" is still not found. 11308 [Ralf S. Engelschall] 11309 11310 *) Added more platforms to Configure: Cray T3E, HPUX 11, 11311 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 11312 11313 *) Updated the README file. 11314 [Ralf S. Engelschall] 11315 11316 *) Added various .cvsignore files in the CVS repository subdirs 11317 to make a "cvs update" really silent. 11318 [Ralf S. Engelschall] 11319 11320 *) Recompiled the error-definition header files and added 11321 missing symbols to the Win32 linker tables. 11322 [Ralf S. Engelschall] 11323 11324 *) Cleaned up the top-level documents; 11325 o new files: CHANGES and LICENSE 11326 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 11327 o merged COPYRIGHT into LICENSE 11328 o removed obsolete TODO file 11329 o renamed MICROSOFT to INSTALL.W32 11330 [Ralf S. Engelschall] 11331 11332 *) Removed dummy files from the 0.9.1b source tree: 11333 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 11334 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 11335 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 11336 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 11337 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 11338 [Ralf S. Engelschall] 11339 11340 *) Added various platform portability fixes. 11341 [Mark J. Cox] 11342 11343 *) The Genesis of the OpenSSL rpject: 11344 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 11345 Young and Tim J. Hudson created while they were working for C2Net until 11346 summer 1998. 11347 [The OpenSSL Project] 11348 11349 11350 Changes between 0.9.0b and 0.9.1b [not released] 11351 11352 *) Updated a few CA certificates under certs/ 11353 [Eric A. Young] 11354 11355 *) Changed some BIGNUM api stuff. 11356 [Eric A. Young] 11357 11358 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 11359 DGUX x86, Linux Alpha, etc. 11360 [Eric A. Young] 11361 11362 *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 11363 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 11364 available). 11365 [Eric A. Young] 11366 11367 *) Add -strparse option to asn1pars program which parses nested 11368 binary structures 11369 [Dr Stephen Henson <shenson@bigfoot.com>] 11370 11371 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 11372 [Eric A. Young] 11373 11374 *) DSA fix for "ca" program. 11375 [Eric A. Young] 11376 11377 *) Added "-genkey" option to "dsaparam" program. 11378 [Eric A. Young] 11379 11380 *) Added RIPE MD160 (rmd160) message digest. 11381 [Eric A. Young] 11382 11383 *) Added -a (all) option to "ssleay version" command. 11384 [Eric A. Young] 11385 11386 *) Added PLATFORM define which is the id given to Configure. 11387 [Eric A. Young] 11388 11389 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 11390 [Eric A. Young] 11391 11392 *) Extended the ASN.1 parser routines. 11393 [Eric A. Young] 11394 11395 *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 11396 [Eric A. Young] 11397 11398 *) Added a BN_CTX to the BN library. 11399 [Eric A. Young] 11400 11401 *) Fixed the weak key values in DES library 11402 [Eric A. Young] 11403 11404 *) Changed API in EVP library for cipher aliases. 11405 [Eric A. Young] 11406 11407 *) Added support for RC2/64bit cipher. 11408 [Eric A. Young] 11409 11410 *) Converted the lhash library to the crypto/mem.c functions. 11411 [Eric A. Young] 11412 11413 *) Added more recognized ASN.1 object ids. 11414 [Eric A. Young] 11415 11416 *) Added more RSA padding checks for SSL/TLS. 11417 [Eric A. Young] 11418 11419 *) Added BIO proxy/filter functionality. 11420 [Eric A. Young] 11421 11422 *) Added extra_certs to SSL_CTX which can be used 11423 send extra CA certificates to the client in the CA cert chain sending 11424 process. It can be configured with SSL_CTX_add_extra_chain_cert(). 11425 [Eric A. Young] 11426 11427 *) Now Fortezza is denied in the authentication phase because 11428 this is key exchange mechanism is not supported by SSLeay at all. 11429 [Eric A. Young] 11430 11431 *) Additional PKCS1 checks. 11432 [Eric A. Young] 11433 11434 *) Support the string "TLSv1" for all TLS v1 ciphers. 11435 [Eric A. Young] 11436 11437 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 11438 ex_data index of the SSL context in the X509_STORE_CTX ex_data. 11439 [Eric A. Young] 11440 11441 *) Fixed a few memory leaks. 11442 [Eric A. Young] 11443 11444 *) Fixed various code and comment typos. 11445 [Eric A. Young] 11446 11447 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 11448 bytes sent in the client random. 11449 [Edward Bishop <ebishop@spyglass.com>] 11450 11451