CHANGES revision 356290
1 2 OpenSSL CHANGES 3 _______________ 4 5 This is a high-level summary of the most important changes. 6 For a full list of changes, see the git commit log; for example, 7 https://github.com/openssl/openssl/commits/ and pick the appropriate 8 release branch. 9 10 Changes between 1.0.2t and 1.0.2u [20 Dec 2019] 11 12 *) Fixed an an overflow bug in the x64_64 Montgomery squaring procedure 13 used in exponentiation with 512-bit moduli. No EC algorithms are 14 affected. Analysis suggests that attacks against 2-prime RSA1024, 15 3-prime RSA1536, and DSA1024 as a result of this defect would be very 16 difficult to perform and are not believed likely. Attacks against DH512 17 are considered just feasible. However, for an attack the target would 18 have to re-use the DH512 private key, which is not recommended anyway. 19 Also applications directly using the low level API BN_mod_exp may be 20 affected if they use BN_FLG_CONSTTIME. 21 (CVE-2019-1551) 22 [Andy Polyakov] 23 24 Changes between 1.0.2s and 1.0.2t [10 Sep 2019] 25 26 *) For built-in EC curves, ensure an EC_GROUP built from the curve name is 27 used even when parsing explicit parameters, when loading a serialized key 28 or calling `EC_GROUP_new_from_ecpkparameters()`/ 29 `EC_GROUP_new_from_ecparameters()`. 30 This prevents bypass of security hardening and performance gains, 31 especially for curves with specialized EC_METHODs. 32 By default, if a key encoded with explicit parameters is loaded and later 33 serialized, the output is still encoded with explicit parameters, even if 34 internally a "named" EC_GROUP is used for computation. 35 [Nicola Tuveri] 36 37 *) Compute ECC cofactors if not provided during EC_GROUP construction. Before 38 this change, EC_GROUP_set_generator would accept order and/or cofactor as 39 NULL. After this change, only the cofactor parameter can be NULL. It also 40 does some minimal sanity checks on the passed order. 41 (CVE-2019-1547) 42 [Billy Bob Brumley] 43 44 *) Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey. 45 An attack is simple, if the first CMS_recipientInfo is valid but the 46 second CMS_recipientInfo is chosen ciphertext. If the second 47 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct 48 encryption key will be replaced by garbage, and the message cannot be 49 decoded, but if the RSA decryption fails, the correct encryption key is 50 used and the recipient will not notice the attack. 51 As a work around for this potential attack the length of the decrypted 52 key must be equal to the cipher default key length, in case the 53 certifiate is not given and all recipientInfo are tried out. 54 The old behaviour can be re-enabled in the CMS code by setting the 55 CMS_DEBUG_DECRYPT flag. 56 (CVE-2019-1563) 57 [Bernd Edlinger] 58 59 *) Document issue with installation paths in diverse Windows builds 60 61 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL 62 binaries and run-time config file. 63 (CVE-2019-1552) 64 [Richard Levitte] 65 66 Changes between 1.0.2r and 1.0.2s [28 May 2019] 67 68 *) Change the default RSA, DSA and DH size to 2048 bit instead of 1024. 69 This changes the size when using the genpkey app when no size is given. It 70 fixes an omission in earlier changes that changed all RSA, DSA and DH 71 generation apps to use 2048 bits by default. 72 [Kurt Roeckx] 73 74 *) Add FIPS support for Android Arm 64-bit 75 76 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object 77 Module in Version 2.0.10. For some reason, the corresponding target 78 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be 79 built with FIPS support on Android Arm 64-bit. This omission has been 80 fixed. 81 [Matthias St. Pierre] 82 83 Changes between 1.0.2q and 1.0.2r [26 Feb 2019] 84 85 *) 0-byte record padding oracle 86 87 If an application encounters a fatal protocol error and then calls 88 SSL_shutdown() twice (once to send a close_notify, and once to receive one) 89 then OpenSSL can respond differently to the calling application if a 0 byte 90 record is received with invalid padding compared to if a 0 byte record is 91 received with an invalid MAC. If the application then behaves differently 92 based on that in a way that is detectable to the remote peer, then this 93 amounts to a padding oracle that could be used to decrypt data. 94 95 In order for this to be exploitable "non-stitched" ciphersuites must be in 96 use. Stitched ciphersuites are optimised implementations of certain 97 commonly used ciphersuites. Also the application must call SSL_shutdown() 98 twice even if a protocol error has occurred (applications should not do 99 this but some do anyway). 100 101 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod 102 Aviram, with additional investigation by Steven Collison and Andrew 103 Hourselt. It was reported to OpenSSL on 10th December 2018. 104 (CVE-2019-1559) 105 [Matt Caswell] 106 107 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0(). 108 [Richard Levitte] 109 110 Changes between 1.0.2p and 1.0.2q [20 Nov 2018] 111 112 *) Microarchitecture timing vulnerability in ECC scalar multiplication 113 114 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been 115 shown to be vulnerable to a microarchitecture timing side channel attack. 116 An attacker with sufficient access to mount local timing attacks during 117 ECDSA signature generation could recover the private key. 118 119 This issue was reported to OpenSSL on 26th October 2018 by Alejandro 120 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and 121 Nicola Tuveri. 122 (CVE-2018-5407) 123 [Billy Brumley] 124 125 *) Timing vulnerability in DSA signature generation 126 127 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a 128 timing side channel attack. An attacker could use variations in the signing 129 algorithm to recover the private key. 130 131 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser. 132 (CVE-2018-0734) 133 [Paul Dale] 134 135 *) Resolve a compatibility issue in EC_GROUP handling with the FIPS Object 136 Module, accidentally introduced while backporting security fixes from the 137 development branch and hindering the use of ECC in FIPS mode. 138 [Nicola Tuveri] 139 140 Changes between 1.0.2o and 1.0.2p [14 Aug 2018] 141 142 *) Client DoS due to large DH parameter 143 144 During key agreement in a TLS handshake using a DH(E) based ciphersuite a 145 malicious server can send a very large prime value to the client. This will 146 cause the client to spend an unreasonably long period of time generating a 147 key for this prime resulting in a hang until the client has finished. This 148 could be exploited in a Denial Of Service attack. 149 150 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken 151 (CVE-2018-0732) 152 [Guido Vranken] 153 154 *) Cache timing vulnerability in RSA Key Generation 155 156 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to 157 a cache timing side channel attack. An attacker with sufficient access to 158 mount cache timing attacks during the RSA key generation process could 159 recover the private key. 160 161 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera 162 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia. 163 (CVE-2018-0737) 164 [Billy Brumley] 165 166 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str 167 parameter is no longer accepted, as it leads to a corrupt table. NULL 168 pem_str is reserved for alias entries only. 169 [Richard Levitte] 170 171 *) Revert blinding in ECDSA sign and instead make problematic addition 172 length-invariant. Switch even to fixed-length Montgomery multiplication. 173 [Andy Polyakov] 174 175 *) Change generating and checking of primes so that the error rate of not 176 being prime depends on the intended use based on the size of the input. 177 For larger primes this will result in more rounds of Miller-Rabin. 178 The maximal error rate for primes with more than 1080 bits is lowered 179 to 2^-128. 180 [Kurt Roeckx, Annie Yousar] 181 182 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64. 183 [Kurt Roeckx] 184 185 *) Add blinding to ECDSA and DSA signatures to protect against side channel 186 attacks discovered by Keegan Ryan (NCC Group). 187 [Matt Caswell] 188 189 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we 190 now allow empty (zero character) pass phrases. 191 [Richard Levitte] 192 193 *) Certificate time validation (X509_cmp_time) enforces stricter 194 compliance with RFC 5280. Fractional seconds and timezone offsets 195 are no longer allowed. 196 [Emilia K��sper] 197 198 Changes between 1.0.2n and 1.0.2o [27 Mar 2018] 199 200 *) Constructed ASN.1 types with a recursive definition could exceed the stack 201 202 Constructed ASN.1 types with a recursive definition (such as can be found 203 in PKCS7) could eventually exceed the stack given malicious input with 204 excessive recursion. This could result in a Denial Of Service attack. There 205 are no such structures used within SSL/TLS that come from untrusted sources 206 so this is considered safe. 207 208 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz 209 project. 210 (CVE-2018-0739) 211 [Matt Caswell] 212 213 Changes between 1.0.2m and 1.0.2n [7 Dec 2017] 214 215 *) Read/write after SSL object in error state 216 217 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state" 218 mechanism. The intent was that if a fatal error occurred during a handshake 219 then OpenSSL would move into the error state and would immediately fail if 220 you attempted to continue the handshake. This works as designed for the 221 explicit handshake functions (SSL_do_handshake(), SSL_accept() and 222 SSL_connect()), however due to a bug it does not work correctly if 223 SSL_read() or SSL_write() is called directly. In that scenario, if the 224 handshake fails then a fatal error will be returned in the initial function 225 call. If SSL_read()/SSL_write() is subsequently called by the application 226 for the same SSL object then it will succeed and the data is passed without 227 being decrypted/encrypted directly from the SSL/TLS record layer. 228 229 In order to exploit this issue an application bug would have to be present 230 that resulted in a call to SSL_read()/SSL_write() being issued after having 231 already received a fatal error. 232 233 This issue was reported to OpenSSL by David Benjamin (Google). 234 (CVE-2017-3737) 235 [Matt Caswell] 236 237 *) rsaz_1024_mul_avx2 overflow bug on x86_64 238 239 There is an overflow bug in the AVX2 Montgomery multiplication procedure 240 used in exponentiation with 1024-bit moduli. No EC algorithms are affected. 241 Analysis suggests that attacks against RSA and DSA as a result of this 242 defect would be very difficult to perform and are not believed likely. 243 Attacks against DH1024 are considered just feasible, because most of the 244 work necessary to deduce information about a private key may be performed 245 offline. The amount of resources required for such an attack would be 246 significant. However, for an attack on TLS to be meaningful, the server 247 would have to share the DH1024 private key among multiple clients, which is 248 no longer an option since CVE-2016-0701. 249 250 This only affects processors that support the AVX2 but not ADX extensions 251 like Intel Haswell (4th generation). 252 253 This issue was reported to OpenSSL by David Benjamin (Google). The issue 254 was originally found via the OSS-Fuzz project. 255 (CVE-2017-3738) 256 [Andy Polyakov] 257 258 Changes between 1.0.2l and 1.0.2m [2 Nov 2017] 259 260 *) bn_sqrx8x_internal carry bug on x86_64 261 262 There is a carry propagating bug in the x86_64 Montgomery squaring 263 procedure. No EC algorithms are affected. Analysis suggests that attacks 264 against RSA and DSA as a result of this defect would be very difficult to 265 perform and are not believed likely. Attacks against DH are considered just 266 feasible (although very difficult) because most of the work necessary to 267 deduce information about a private key may be performed offline. The amount 268 of resources required for such an attack would be very significant and 269 likely only accessible to a limited number of attackers. An attacker would 270 additionally need online access to an unpatched system using the target 271 private key in a scenario with persistent DH parameters and a private 272 key that is shared between multiple clients. 273 274 This only affects processors that support the BMI1, BMI2 and ADX extensions 275 like Intel Broadwell (5th generation) and later or AMD Ryzen. 276 277 This issue was reported to OpenSSL by the OSS-Fuzz project. 278 (CVE-2017-3736) 279 [Andy Polyakov] 280 281 *) Malformed X.509 IPAddressFamily could cause OOB read 282 283 If an X.509 certificate has a malformed IPAddressFamily extension, 284 OpenSSL could do a one-byte buffer overread. The most likely result 285 would be an erroneous display of the certificate in text format. 286 287 This issue was reported to OpenSSL by the OSS-Fuzz project. 288 (CVE-2017-3735) 289 [Rich Salz] 290 291 Changes between 1.0.2k and 1.0.2l [25 May 2017] 292 293 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target 294 platform rather than 'mingw'. 295 [Richard Levitte] 296 297 Changes between 1.0.2j and 1.0.2k [26 Jan 2017] 298 299 *) Truncated packet could crash via OOB read 300 301 If one side of an SSL/TLS path is running on a 32-bit host and a specific 302 cipher is being used, then a truncated packet can cause that host to 303 perform an out-of-bounds read, usually resulting in a crash. 304 305 This issue was reported to OpenSSL by Robert ��wi��cki of Google. 306 (CVE-2017-3731) 307 [Andy Polyakov] 308 309 *) BN_mod_exp may produce incorrect results on x86_64 310 311 There is a carry propagating bug in the x86_64 Montgomery squaring 312 procedure. No EC algorithms are affected. Analysis suggests that attacks 313 against RSA and DSA as a result of this defect would be very difficult to 314 perform and are not believed likely. Attacks against DH are considered just 315 feasible (although very difficult) because most of the work necessary to 316 deduce information about a private key may be performed offline. The amount 317 of resources required for such an attack would be very significant and 318 likely only accessible to a limited number of attackers. An attacker would 319 additionally need online access to an unpatched system using the target 320 private key in a scenario with persistent DH parameters and a private 321 key that is shared between multiple clients. For example this can occur by 322 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very 323 similar to CVE-2015-3193 but must be treated as a separate problem. 324 325 This issue was reported to OpenSSL by the OSS-Fuzz project. 326 (CVE-2017-3732) 327 [Andy Polyakov] 328 329 *) Montgomery multiplication may produce incorrect results 330 331 There is a carry propagating bug in the Broadwell-specific Montgomery 332 multiplication procedure that handles input lengths divisible by, but 333 longer than 256 bits. Analysis suggests that attacks against RSA, DSA 334 and DH private keys are impossible. This is because the subroutine in 335 question is not used in operations with the private key itself and an input 336 of the attacker's direct choice. Otherwise the bug can manifest itself as 337 transient authentication and key negotiation failures or reproducible 338 erroneous outcome of public-key operations with specially crafted input. 339 Among EC algorithms only Brainpool P-512 curves are affected and one 340 presumably can attack ECDH key negotiation. Impact was not analyzed in 341 detail, because pre-requisites for attack are considered unlikely. Namely 342 multiple clients have to choose the curve in question and the server has to 343 share the private key among them, neither of which is default behaviour. 344 Even then only clients that chose the curve will be affected. 345 346 This issue was publicly reported as transient failures and was not 347 initially recognized as a security issue. Thanks to Richard Morgan for 348 providing reproducible case. 349 (CVE-2016-7055) 350 [Andy Polyakov] 351 352 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0 353 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to 354 prevent issues where no progress is being made and the peer continually 355 sends unrecognised record types, using up resources processing them. 356 [Matt Caswell] 357 358 Changes between 1.0.2i and 1.0.2j [26 Sep 2016] 359 360 *) Missing CRL sanity check 361 362 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0 363 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use 364 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception. 365 366 This issue only affects the OpenSSL 1.0.2i 367 (CVE-2016-7052) 368 [Matt Caswell] 369 370 Changes between 1.0.2h and 1.0.2i [22 Sep 2016] 371 372 *) OCSP Status Request extension unbounded memory growth 373 374 A malicious client can send an excessively large OCSP Status Request 375 extension. If that client continually requests renegotiation, sending a 376 large OCSP Status Request extension each time, then there will be unbounded 377 memory growth on the server. This will eventually lead to a Denial Of 378 Service attack through memory exhaustion. Servers with a default 379 configuration are vulnerable even if they do not support OCSP. Builds using 380 the "no-ocsp" build time option are not affected. 381 382 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 383 (CVE-2016-6304) 384 [Matt Caswell] 385 386 *) In order to mitigate the SWEET32 attack, the DES ciphers were moved from 387 HIGH to MEDIUM. 388 389 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan 390 Leurent (INRIA) 391 (CVE-2016-2183) 392 [Rich Salz] 393 394 *) OOB write in MDC2_Update() 395 396 An overflow can occur in MDC2_Update() either if called directly or 397 through the EVP_DigestUpdate() function using MDC2. If an attacker 398 is able to supply very large amounts of input data after a previous 399 call to EVP_EncryptUpdate() with a partial block then a length check 400 can overflow resulting in a heap corruption. 401 402 The amount of data needed is comparable to SIZE_MAX which is impractical 403 on most platforms. 404 405 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 406 (CVE-2016-6303) 407 [Stephen Henson] 408 409 *) Malformed SHA512 ticket DoS 410 411 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a 412 DoS attack where a malformed ticket will result in an OOB read which will 413 ultimately crash. 414 415 The use of SHA512 in TLS session tickets is comparatively rare as it requires 416 a custom server callback and ticket lookup mechanism. 417 418 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 419 (CVE-2016-6302) 420 [Stephen Henson] 421 422 *) OOB write in BN_bn2dec() 423 424 The function BN_bn2dec() does not check the return value of BN_div_word(). 425 This can cause an OOB write if an application uses this function with an 426 overly large BIGNUM. This could be a problem if an overly large certificate 427 or CRL is printed out from an untrusted source. TLS is not affected because 428 record limits will reject an oversized certificate before it is parsed. 429 430 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 431 (CVE-2016-2182) 432 [Stephen Henson] 433 434 *) OOB read in TS_OBJ_print_bio() 435 436 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is 437 the total length the OID text representation would use and not the amount 438 of data written. This will result in OOB reads when large OIDs are 439 presented. 440 441 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 442 (CVE-2016-2180) 443 [Stephen Henson] 444 445 *) Pointer arithmetic undefined behaviour 446 447 Avoid some undefined pointer arithmetic 448 449 A common idiom in the codebase is to check limits in the following manner: 450 "p + len > limit" 451 452 Where "p" points to some malloc'd data of SIZE bytes and 453 limit == p + SIZE 454 455 "len" here could be from some externally supplied data (e.g. from a TLS 456 message). 457 458 The rules of C pointer arithmetic are such that "p + len" is only well 459 defined where len <= SIZE. Therefore the above idiom is actually 460 undefined behaviour. 461 462 For example this could cause problems if some malloc implementation 463 provides an address for "p" such that "p + len" actually overflows for 464 values of len that are too big and therefore p + len < limit. 465 466 This issue was reported to OpenSSL by Guido Vranken 467 (CVE-2016-2177) 468 [Matt Caswell] 469 470 *) Constant time flag not preserved in DSA signing 471 472 Operations in the DSA signing algorithm should run in constant time in 473 order to avoid side channel attacks. A flaw in the OpenSSL DSA 474 implementation means that a non-constant time codepath is followed for 475 certain operations. This has been demonstrated through a cache-timing 476 attack to be sufficient for an attacker to recover the private DSA key. 477 478 This issue was reported by C��sar Pereida (Aalto University), Billy Brumley 479 (Tampere University of Technology), and Yuval Yarom (The University of 480 Adelaide and NICTA). 481 (CVE-2016-2178) 482 [C��sar Pereida] 483 484 *) DTLS buffered message DoS 485 486 In a DTLS connection where handshake messages are delivered out-of-order 487 those messages that OpenSSL is not yet ready to process will be buffered 488 for later use. Under certain circumstances, a flaw in the logic means that 489 those messages do not get removed from the buffer even though the handshake 490 has been completed. An attacker could force up to approx. 15 messages to 491 remain in the buffer when they are no longer required. These messages will 492 be cleared when the DTLS connection is closed. The default maximum size for 493 a message is 100k. Therefore the attacker could force an additional 1500k 494 to be consumed per connection. By opening many simulataneous connections an 495 attacker could cause a DoS attack through memory exhaustion. 496 497 This issue was reported to OpenSSL by Quan Luo. 498 (CVE-2016-2179) 499 [Matt Caswell] 500 501 *) DTLS replay protection DoS 502 503 A flaw in the DTLS replay attack protection mechanism means that records 504 that arrive for future epochs update the replay protection "window" before 505 the MAC for the record has been validated. This could be exploited by an 506 attacker by sending a record for the next epoch (which does not have to 507 decrypt or have a valid MAC), with a very large sequence number. This means 508 that all subsequent legitimate packets are dropped causing a denial of 509 service for a specific DTLS connection. 510 511 This issue was reported to OpenSSL by the OCAP audit team. 512 (CVE-2016-2181) 513 [Matt Caswell] 514 515 *) Certificate message OOB reads 516 517 In OpenSSL 1.0.2 and earlier some missing message length checks can result 518 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a 519 theoretical DoS risk but this has not been observed in practice on common 520 platforms. 521 522 The messages affected are client certificate, client certificate request 523 and server certificate. As a result the attack can only be performed 524 against a client or a server which enables client authentication. 525 526 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 527 (CVE-2016-6306) 528 [Stephen Henson] 529 530 Changes between 1.0.2g and 1.0.2h [3 May 2016] 531 532 *) Prevent padding oracle in AES-NI CBC MAC check 533 534 A MITM attacker can use a padding oracle attack to decrypt traffic 535 when the connection uses an AES CBC cipher and the server support 536 AES-NI. 537 538 This issue was introduced as part of the fix for Lucky 13 padding 539 attack (CVE-2013-0169). The padding check was rewritten to be in 540 constant time by making sure that always the same bytes are read and 541 compared against either the MAC or padding bytes. But it no longer 542 checked that there was enough data to have both the MAC and padding 543 bytes. 544 545 This issue was reported by Juraj Somorovsky using TLS-Attacker. 546 (CVE-2016-2107) 547 [Kurt Roeckx] 548 549 *) Fix EVP_EncodeUpdate overflow 550 551 An overflow can occur in the EVP_EncodeUpdate() function which is used for 552 Base64 encoding of binary data. If an attacker is able to supply very large 553 amounts of input data then a length check can overflow resulting in a heap 554 corruption. 555 556 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by 557 the PEM_write_bio* family of functions. These are mainly used within the 558 OpenSSL command line applications, so any application which processes data 559 from an untrusted source and outputs it as a PEM file should be considered 560 vulnerable to this issue. User applications that call these APIs directly 561 with large amounts of untrusted data may also be vulnerable. 562 563 This issue was reported by Guido Vranken. 564 (CVE-2016-2105) 565 [Matt Caswell] 566 567 *) Fix EVP_EncryptUpdate overflow 568 569 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker 570 is able to supply very large amounts of input data after a previous call to 571 EVP_EncryptUpdate() with a partial block then a length check can overflow 572 resulting in a heap corruption. Following an analysis of all OpenSSL 573 internal usage of the EVP_EncryptUpdate() function all usage is one of two 574 forms. The first form is where the EVP_EncryptUpdate() call is known to be 575 the first called function after an EVP_EncryptInit(), and therefore that 576 specific call must be safe. The second form is where the length passed to 577 EVP_EncryptUpdate() can be seen from the code to be some small value and 578 therefore there is no possibility of an overflow. Since all instances are 579 one of these two forms, it is believed that there can be no overflows in 580 internal code due to this problem. It should be noted that 581 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths. 582 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances 583 of these calls have also been analysed too and it is believed there are no 584 instances in internal usage where an overflow could occur. 585 586 This issue was reported by Guido Vranken. 587 (CVE-2016-2106) 588 [Matt Caswell] 589 590 *) Prevent ASN.1 BIO excessive memory allocation 591 592 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio() 593 a short invalid encoding can casuse allocation of large amounts of memory 594 potentially consuming excessive resources or exhausting memory. 595 596 Any application parsing untrusted data through d2i BIO functions is 597 affected. The memory based functions such as d2i_X509() are *not* affected. 598 Since the memory based functions are used by the TLS library, TLS 599 applications are not affected. 600 601 This issue was reported by Brian Carpenter. 602 (CVE-2016-2109) 603 [Stephen Henson] 604 605 *) EBCDIC overread 606 607 ASN1 Strings that are over 1024 bytes can cause an overread in applications 608 using the X509_NAME_oneline() function on EBCDIC systems. This could result 609 in arbitrary stack data being returned in the buffer. 610 611 This issue was reported by Guido Vranken. 612 (CVE-2016-2176) 613 [Matt Caswell] 614 615 *) Modify behavior of ALPN to invoke callback after SNI/servername 616 callback, such that updates to the SSL_CTX affect ALPN. 617 [Todd Short] 618 619 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the 620 default. 621 [Kurt Roeckx] 622 623 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the 624 methods are enabled and ssl2 is disabled the methods return NULL. 625 [Kurt Roeckx] 626 627 Changes between 1.0.2f and 1.0.2g [1 Mar 2016] 628 629 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL. 630 Builds that are not configured with "enable-weak-ssl-ciphers" will not 631 provide any "EXPORT" or "LOW" strength ciphers. 632 [Viktor Dukhovni] 633 634 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2 635 is by default disabled at build-time. Builds that are not configured with 636 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used, 637 users who want to negotiate SSLv2 via the version-flexible SSLv23_method() 638 will need to explicitly call either of: 639 640 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2); 641 or 642 SSL_clear_options(ssl, SSL_OP_NO_SSLv2); 643 644 as appropriate. Even if either of those is used, or the application 645 explicitly uses the version-specific SSLv2_method() or its client and 646 server variants, SSLv2 ciphers vulnerable to exhaustive search key 647 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT 648 ciphers, and SSLv2 56-bit DES are no longer available. 649 (CVE-2016-0800) 650 [Viktor Dukhovni] 651 652 *) Fix a double-free in DSA code 653 654 A double free bug was discovered when OpenSSL parses malformed DSA private 655 keys and could lead to a DoS attack or memory corruption for applications 656 that receive DSA private keys from untrusted sources. This scenario is 657 considered rare. 658 659 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using 660 libFuzzer. 661 (CVE-2016-0705) 662 [Stephen Henson] 663 664 *) Disable SRP fake user seed to address a server memory leak. 665 666 Add a new method SRP_VBASE_get1_by_user that handles the seed properly. 667 668 SRP_VBASE_get_by_user had inconsistent memory management behaviour. 669 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user 670 was changed to ignore the "fake user" SRP seed, even if the seed 671 is configured. 672 673 Users should use SRP_VBASE_get1_by_user instead. Note that in 674 SRP_VBASE_get1_by_user, caller must free the returned value. Note 675 also that even though configuring the SRP seed attempts to hide 676 invalid usernames by continuing the handshake with fake 677 credentials, this behaviour is not constant time and no strong 678 guarantees are made that the handshake is indistinguishable from 679 that of a valid user. 680 (CVE-2016-0798) 681 [Emilia K��sper] 682 683 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption 684 685 In the BN_hex2bn function the number of hex digits is calculated using an 686 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For 687 large values of |i| this can result in |bn_expand| not allocating any 688 memory because |i * 4| is negative. This can leave the internal BIGNUM data 689 field as NULL leading to a subsequent NULL ptr deref. For very large values 690 of |i|, the calculation |i * 4| could be a positive value smaller than |i|. 691 In this case memory is allocated to the internal BIGNUM data field, but it 692 is insufficiently sized leading to heap corruption. A similar issue exists 693 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn 694 is ever called by user applications with very large untrusted hex/dec data. 695 This is anticipated to be a rare occurrence. 696 697 All OpenSSL internal usage of these functions use data that is not expected 698 to be untrusted, e.g. config file data or application command line 699 arguments. If user developed applications generate config file data based 700 on untrusted data then it is possible that this could also lead to security 701 consequences. This is also anticipated to be rare. 702 703 This issue was reported to OpenSSL by Guido Vranken. 704 (CVE-2016-0797) 705 [Matt Caswell] 706 707 *) Fix memory issues in BIO_*printf functions 708 709 The internal |fmtstr| function used in processing a "%s" format string in 710 the BIO_*printf functions could overflow while calculating the length of a 711 string and cause an OOB read when printing very long strings. 712 713 Additionally the internal |doapr_outch| function can attempt to write to an 714 OOB memory location (at an offset from the NULL pointer) in the event of a 715 memory allocation failure. In 1.0.2 and below this could be caused where 716 the size of a buffer to be allocated is greater than INT_MAX. E.g. this 717 could be in processing a very long "%s" format string. Memory leaks can 718 also occur. 719 720 The first issue may mask the second issue dependent on compiler behaviour. 721 These problems could enable attacks where large amounts of untrusted data 722 is passed to the BIO_*printf functions. If applications use these functions 723 in this way then they could be vulnerable. OpenSSL itself uses these 724 functions when printing out human-readable dumps of ASN.1 data. Therefore 725 applications that print this data could be vulnerable if the data is from 726 untrusted sources. OpenSSL command line applications could also be 727 vulnerable where they print out ASN.1 data, or if untrusted data is passed 728 as command line arguments. 729 730 Libssl is not considered directly vulnerable. Additionally certificates etc 731 received via remote connections via libssl are also unlikely to be able to 732 trigger these issues because of message size limits enforced within libssl. 733 734 This issue was reported to OpenSSL Guido Vranken. 735 (CVE-2016-0799) 736 [Matt Caswell] 737 738 *) Side channel attack on modular exponentiation 739 740 A side-channel attack was found which makes use of cache-bank conflicts on 741 the Intel Sandy-Bridge microarchitecture which could lead to the recovery 742 of RSA keys. The ability to exploit this issue is limited as it relies on 743 an attacker who has control of code in a thread running on the same 744 hyper-threaded core as the victim thread which is performing decryptions. 745 746 This issue was reported to OpenSSL by Yuval Yarom, The University of 747 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and 748 Nadia Heninger, University of Pennsylvania with more information at 749 http://cachebleed.info. 750 (CVE-2016-0702) 751 [Andy Polyakov] 752 753 *) Change the req app to generate a 2048-bit RSA/DSA key by default, 754 if no keysize is specified with default_bits. This fixes an 755 omission in an earlier change that changed all RSA/DSA key generation 756 apps to use 2048 bits by default. 757 [Emilia K��sper] 758 759 Changes between 1.0.2e and 1.0.2f [28 Jan 2016] 760 761 *) DH small subgroups 762 763 Historically OpenSSL only ever generated DH parameters based on "safe" 764 primes. More recently (in version 1.0.2) support was provided for 765 generating X9.42 style parameter files such as those required for RFC 5114 766 support. The primes used in such files may not be "safe". Where an 767 application is using DH configured with parameters based on primes that are 768 not "safe" then an attacker could use this fact to find a peer's private 769 DH exponent. This attack requires that the attacker complete multiple 770 handshakes in which the peer uses the same private DH exponent. For example 771 this could be used to discover a TLS server's private DH exponent if it's 772 reusing the private DH exponent or it's using a static DH ciphersuite. 773 774 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in 775 TLS. It is not on by default. If the option is not set then the server 776 reuses the same private DH exponent for the life of the server process and 777 would be vulnerable to this attack. It is believed that many popular 778 applications do set this option and would therefore not be at risk. 779 780 The fix for this issue adds an additional check where a "q" parameter is 781 available (as is the case in X9.42 based parameters). This detects the 782 only known attack, and is the only possible defense for static DH 783 ciphersuites. This could have some performance impact. 784 785 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by 786 default and cannot be disabled. This could have some performance impact. 787 788 This issue was reported to OpenSSL by Antonio Sanso (Adobe). 789 (CVE-2016-0701) 790 [Matt Caswell] 791 792 *) SSLv2 doesn't block disabled ciphers 793 794 A malicious client can negotiate SSLv2 ciphers that have been disabled on 795 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have 796 been disabled, provided that the SSLv2 protocol was not also disabled via 797 SSL_OP_NO_SSLv2. 798 799 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram 800 and Sebastian Schinzel. 801 (CVE-2015-3197) 802 [Viktor Dukhovni] 803 804 *) Reject DH handshakes with parameters shorter than 1024 bits. 805 [Kurt Roeckx] 806 807 Changes between 1.0.2d and 1.0.2e [3 Dec 2015] 808 809 *) BN_mod_exp may produce incorrect results on x86_64 810 811 There is a carry propagating bug in the x86_64 Montgomery squaring 812 procedure. No EC algorithms are affected. Analysis suggests that attacks 813 against RSA and DSA as a result of this defect would be very difficult to 814 perform and are not believed likely. Attacks against DH are considered just 815 feasible (although very difficult) because most of the work necessary to 816 deduce information about a private key may be performed offline. The amount 817 of resources required for such an attack would be very significant and 818 likely only accessible to a limited number of attackers. An attacker would 819 additionally need online access to an unpatched system using the target 820 private key in a scenario with persistent DH parameters and a private 821 key that is shared between multiple clients. For example this can occur by 822 default in OpenSSL DHE based SSL/TLS ciphersuites. 823 824 This issue was reported to OpenSSL by Hanno B��ck. 825 (CVE-2015-3193) 826 [Andy Polyakov] 827 828 *) Certificate verify crash with missing PSS parameter 829 830 The signature verification routines will crash with a NULL pointer 831 dereference if presented with an ASN.1 signature using the RSA PSS 832 algorithm and absent mask generation function parameter. Since these 833 routines are used to verify certificate signature algorithms this can be 834 used to crash any certificate verification operation and exploited in a 835 DoS attack. Any application which performs certificate verification is 836 vulnerable including OpenSSL clients and servers which enable client 837 authentication. 838 839 This issue was reported to OpenSSL by Lo��c Jonas Etienne (Qnective AG). 840 (CVE-2015-3194) 841 [Stephen Henson] 842 843 *) X509_ATTRIBUTE memory leak 844 845 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak 846 memory. This structure is used by the PKCS#7 and CMS routines so any 847 application which reads PKCS#7 or CMS data from untrusted sources is 848 affected. SSL/TLS is not affected. 849 850 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using 851 libFuzzer. 852 (CVE-2015-3195) 853 [Stephen Henson] 854 855 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs. 856 This changes the decoding behaviour for some invalid messages, 857 though the change is mostly in the more lenient direction, and 858 legacy behaviour is preserved as much as possible. 859 [Emilia K��sper] 860 861 *) In DSA_generate_parameters_ex, if the provided seed is too short, 862 use a random seed, as already documented. 863 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>] 864 865 Changes between 1.0.2c and 1.0.2d [9 Jul 2015] 866 867 *) Alternate chains certificate forgery 868 869 During certificate verfification, OpenSSL will attempt to find an 870 alternative certificate chain if the first attempt to build such a chain 871 fails. An error in the implementation of this logic can mean that an 872 attacker could cause certain checks on untrusted certificates to be 873 bypassed, such as the CA flag, enabling them to use a valid leaf 874 certificate to act as a CA and "issue" an invalid certificate. 875 876 This issue was reported to OpenSSL by Adam Langley/David Benjamin 877 (Google/BoringSSL). 878 (CVE-2015-1793) 879 [Matt Caswell] 880 881 *) Race condition handling PSK identify hint 882 883 If PSK identity hints are received by a multi-threaded client then 884 the values are wrongly updated in the parent SSL_CTX structure. This can 885 result in a race condition potentially leading to a double free of the 886 identify hint data. 887 (CVE-2015-3196) 888 [Stephen Henson] 889 890 Changes between 1.0.2b and 1.0.2c [12 Jun 2015] 891 892 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI 893 incompatibility in the handling of HMAC. The previous ABI has now been 894 restored. 895 896 Changes between 1.0.2a and 1.0.2b [11 Jun 2015] 897 898 *) Malformed ECParameters causes infinite loop 899 900 When processing an ECParameters structure OpenSSL enters an infinite loop 901 if the curve specified is over a specially malformed binary polynomial 902 field. 903 904 This can be used to perform denial of service against any 905 system which processes public keys, certificate requests or 906 certificates. This includes TLS clients and TLS servers with 907 client authentication enabled. 908 909 This issue was reported to OpenSSL by Joseph Barr-Pixton. 910 (CVE-2015-1788) 911 [Andy Polyakov] 912 913 *) Exploitable out-of-bounds read in X509_cmp_time 914 915 X509_cmp_time does not properly check the length of the ASN1_TIME 916 string and can read a few bytes out of bounds. In addition, 917 X509_cmp_time accepts an arbitrary number of fractional seconds in the 918 time string. 919 920 An attacker can use this to craft malformed certificates and CRLs of 921 various sizes and potentially cause a segmentation fault, resulting in 922 a DoS on applications that verify certificates or CRLs. TLS clients 923 that verify CRLs are affected. TLS clients and servers with client 924 authentication enabled may be affected if they use custom verification 925 callbacks. 926 927 This issue was reported to OpenSSL by Robert Swiecki (Google), and 928 independently by Hanno B��ck. 929 (CVE-2015-1789) 930 [Emilia K��sper] 931 932 *) PKCS7 crash with missing EnvelopedContent 933 934 The PKCS#7 parsing code does not handle missing inner EncryptedContent 935 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs 936 with missing content and trigger a NULL pointer dereference on parsing. 937 938 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7 939 structures from untrusted sources are affected. OpenSSL clients and 940 servers are not affected. 941 942 This issue was reported to OpenSSL by Michal Zalewski (Google). 943 (CVE-2015-1790) 944 [Emilia K��sper] 945 946 *) CMS verify infinite loop with unknown hash function 947 948 When verifying a signedData message the CMS code can enter an infinite loop 949 if presented with an unknown hash function OID. This can be used to perform 950 denial of service against any system which verifies signedData messages using 951 the CMS code. 952 This issue was reported to OpenSSL by Johannes Bauer. 953 (CVE-2015-1792) 954 [Stephen Henson] 955 956 *) Race condition handling NewSessionTicket 957 958 If a NewSessionTicket is received by a multi-threaded client when attempting to 959 reuse a previous ticket then a race condition can occur potentially leading to 960 a double free of the ticket data. 961 (CVE-2015-1791) 962 [Matt Caswell] 963 964 *) Removed support for the two export grade static DH ciphersuites 965 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites 966 were newly added (along with a number of other static DH ciphersuites) to 967 1.0.2. However the two export ones have *never* worked since they were 968 introduced. It seems strange in any case to be adding new export 969 ciphersuites, and given "logjam" it also does not seem correct to fix them. 970 [Matt Caswell] 971 972 *) Only support 256-bit or stronger elliptic curves with the 973 'ecdh_auto' setting (server) or by default (client). Of supported 974 curves, prefer P-256 (both). 975 [Emilia Kasper] 976 977 *) Reject DH handshakes with parameters shorter than 768 bits. 978 [Kurt Roeckx and Emilia Kasper] 979 980 Changes between 1.0.2 and 1.0.2a [19 Mar 2015] 981 982 *) ClientHello sigalgs DoS fix 983 984 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an 985 invalid signature algorithms extension a NULL pointer dereference will 986 occur. This can be exploited in a DoS attack against the server. 987 988 This issue was was reported to OpenSSL by David Ramos of Stanford 989 University. 990 (CVE-2015-0291) 991 [Stephen Henson and Matt Caswell] 992 993 *) Multiblock corrupted pointer fix 994 995 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This 996 feature only applies on 64 bit x86 architecture platforms that support AES 997 NI instructions. A defect in the implementation of "multiblock" can cause 998 OpenSSL's internal write buffer to become incorrectly set to NULL when 999 using non-blocking IO. Typically, when the user application is using a 1000 socket BIO for writing, this will only result in a failed connection. 1001 However if some other BIO is used then it is likely that a segmentation 1002 fault will be triggered, thus enabling a potential DoS attack. 1003 1004 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller. 1005 (CVE-2015-0290) 1006 [Matt Caswell] 1007 1008 *) Segmentation fault in DTLSv1_listen fix 1009 1010 The DTLSv1_listen function is intended to be stateless and processes the 1011 initial ClientHello from many peers. It is common for user code to loop 1012 over the call to DTLSv1_listen until a valid ClientHello is received with 1013 an associated cookie. A defect in the implementation of DTLSv1_listen means 1014 that state is preserved in the SSL object from one invocation to the next 1015 that can lead to a segmentation fault. Errors processing the initial 1016 ClientHello can trigger this scenario. An example of such an error could be 1017 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only 1018 server. 1019 1020 This issue was reported to OpenSSL by Per Allansson. 1021 (CVE-2015-0207) 1022 [Matt Caswell] 1023 1024 *) Segmentation fault in ASN1_TYPE_cmp fix 1025 1026 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is 1027 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check 1028 certificate signature algorithm consistency this can be used to crash any 1029 certificate verification operation and exploited in a DoS attack. Any 1030 application which performs certificate verification is vulnerable including 1031 OpenSSL clients and servers which enable client authentication. 1032 (CVE-2015-0286) 1033 [Stephen Henson] 1034 1035 *) Segmentation fault for invalid PSS parameters fix 1036 1037 The signature verification routines will crash with a NULL pointer 1038 dereference if presented with an ASN.1 signature using the RSA PSS 1039 algorithm and invalid parameters. Since these routines are used to verify 1040 certificate signature algorithms this can be used to crash any 1041 certificate verification operation and exploited in a DoS attack. Any 1042 application which performs certificate verification is vulnerable including 1043 OpenSSL clients and servers which enable client authentication. 1044 1045 This issue was was reported to OpenSSL by Brian Carpenter. 1046 (CVE-2015-0208) 1047 [Stephen Henson] 1048 1049 *) ASN.1 structure reuse memory corruption fix 1050 1051 Reusing a structure in ASN.1 parsing may allow an attacker to cause 1052 memory corruption via an invalid write. Such reuse is and has been 1053 strongly discouraged and is believed to be rare. 1054 1055 Applications that parse structures containing CHOICE or ANY DEFINED BY 1056 components may be affected. Certificate parsing (d2i_X509 and related 1057 functions) are however not affected. OpenSSL clients and servers are 1058 not affected. 1059 (CVE-2015-0287) 1060 [Stephen Henson] 1061 1062 *) PKCS7 NULL pointer dereferences fix 1063 1064 The PKCS#7 parsing code does not handle missing outer ContentInfo 1065 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with 1066 missing content and trigger a NULL pointer dereference on parsing. 1067 1068 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or 1069 otherwise parse PKCS#7 structures from untrusted sources are 1070 affected. OpenSSL clients and servers are not affected. 1071 1072 This issue was reported to OpenSSL by Michal Zalewski (Google). 1073 (CVE-2015-0289) 1074 [Emilia K��sper] 1075 1076 *) DoS via reachable assert in SSLv2 servers fix 1077 1078 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in 1079 servers that both support SSLv2 and enable export cipher suites by sending 1080 a specially crafted SSLv2 CLIENT-MASTER-KEY message. 1081 1082 This issue was discovered by Sean Burford (Google) and Emilia K��sper 1083 (OpenSSL development team). 1084 (CVE-2015-0293) 1085 [Emilia K��sper] 1086 1087 *) Empty CKE with client auth and DHE fix 1088 1089 If client auth is used then a server can seg fault in the event of a DHE 1090 ciphersuite being selected and a zero length ClientKeyExchange message 1091 being sent by the client. This could be exploited in a DoS attack. 1092 (CVE-2015-1787) 1093 [Matt Caswell] 1094 1095 *) Handshake with unseeded PRNG fix 1096 1097 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake 1098 with an unseeded PRNG. The conditions are: 1099 - The client is on a platform where the PRNG has not been seeded 1100 automatically, and the user has not seeded manually 1101 - A protocol specific client method version has been used (i.e. not 1102 SSL_client_methodv23) 1103 - A ciphersuite is used that does not require additional random data from 1104 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA). 1105 1106 If the handshake succeeds then the client random that has been used will 1107 have been generated from a PRNG with insufficient entropy and therefore the 1108 output may be predictable. 1109 1110 For example using the following command with an unseeded openssl will 1111 succeed on an unpatched platform: 1112 1113 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA 1114 (CVE-2015-0285) 1115 [Matt Caswell] 1116 1117 *) Use After Free following d2i_ECPrivatekey error fix 1118 1119 A malformed EC private key file consumed via the d2i_ECPrivateKey function 1120 could cause a use after free condition. This, in turn, could cause a double 1121 free in several private key parsing functions (such as d2i_PrivateKey 1122 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption 1123 for applications that receive EC private keys from untrusted 1124 sources. This scenario is considered rare. 1125 1126 This issue was discovered by the BoringSSL project and fixed in their 1127 commit 517073cd4b. 1128 (CVE-2015-0209) 1129 [Matt Caswell] 1130 1131 *) X509_to_X509_REQ NULL pointer deref fix 1132 1133 The function X509_to_X509_REQ will crash with a NULL pointer dereference if 1134 the certificate key is invalid. This function is rarely used in practice. 1135 1136 This issue was discovered by Brian Carpenter. 1137 (CVE-2015-0288) 1138 [Stephen Henson] 1139 1140 *) Removed the export ciphers from the DEFAULT ciphers 1141 [Kurt Roeckx] 1142 1143 Changes between 1.0.1l and 1.0.2 [22 Jan 2015] 1144 1145 *) Change RSA and DH/DSA key generation apps to generate 2048-bit 1146 keys by default. 1147 [Kurt Roeckx] 1148 1149 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g. 1150 ARMv5 through ARMv8, as opposite to "locking" it to single one. 1151 So far those who have to target multiple plaforms would compromise 1152 and argue that binary targeting say ARMv5 would still execute on 1153 ARMv8. "Universal" build resolves this compromise by providing 1154 near-optimal performance even on newer platforms. 1155 [Andy Polyakov] 1156 1157 *) Accelerated NIST P-256 elliptic curve implementation for x86_64 1158 (other platforms pending). 1159 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov] 1160 1161 *) Add support for the SignedCertificateTimestampList certificate and 1162 OCSP response extensions from RFC6962. 1163 [Rob Stradling] 1164 1165 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) 1166 for corner cases. (Certain input points at infinity could lead to 1167 bogus results, with non-infinity inputs mapped to infinity too.) 1168 [Bodo Moeller] 1169 1170 *) Initial support for PowerISA 2.0.7, first implemented in POWER8. 1171 This covers AES, SHA256/512 and GHASH. "Initial" means that most 1172 common cases are optimized and there still is room for further 1173 improvements. Vector Permutation AES for Altivec is also added. 1174 [Andy Polyakov] 1175 1176 *) Add support for little-endian ppc64 Linux target. 1177 [Marcelo Cerri (IBM)] 1178 1179 *) Initial support for AMRv8 ISA crypto extensions. This covers AES, 1180 SHA1, SHA256 and GHASH. "Initial" means that most common cases 1181 are optimized and there still is room for further improvements. 1182 Both 32- and 64-bit modes are supported. 1183 [Andy Polyakov, Ard Biesheuvel (Linaro)] 1184 1185 *) Improved ARMv7 NEON support. 1186 [Andy Polyakov] 1187 1188 *) Support for SPARC Architecture 2011 crypto extensions, first 1189 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1, 1190 SHA256/512, MD5, GHASH and modular exponentiation. 1191 [Andy Polyakov, David Miller] 1192 1193 *) Accelerated modular exponentiation for Intel processors, a.k.a. 1194 RSAZ. 1195 [Shay Gueron & Vlad Krasnov (Intel Corp)] 1196 1197 *) Support for new and upcoming Intel processors, including AVX2, 1198 BMI and SHA ISA extensions. This includes additional "stitched" 1199 implementations, AESNI-SHA256 and GCM, and multi-buffer support 1200 for TLS encrypt. 1201 1202 This work was sponsored by Intel Corp. 1203 [Andy Polyakov] 1204 1205 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method() 1206 supports both DTLS 1.2 and 1.0 and should use whatever version the peer 1207 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only. 1208 [Steve Henson] 1209 1210 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file(): 1211 this fixes a limiation in previous versions of OpenSSL. 1212 [Steve Henson] 1213 1214 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest, 1215 MGF1 digest and OAEP label. 1216 [Steve Henson] 1217 1218 *) Add EVP support for key wrapping algorithms, to avoid problems with 1219 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in 1220 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap 1221 algorithms and include tests cases. 1222 [Steve Henson] 1223 1224 *) Add functions to allocate and set the fields of an ECDSA_METHOD 1225 structure. 1226 [Douglas E. Engert, Steve Henson] 1227 1228 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the 1229 difference in days and seconds between two tm or ASN1_TIME structures. 1230 [Steve Henson] 1231 1232 *) Add -rev test option to s_server to just reverse order of characters 1233 received by client and send back to server. Also prints an abbreviated 1234 summary of the connection parameters. 1235 [Steve Henson] 1236 1237 *) New option -brief for s_client and s_server to print out a brief summary 1238 of connection parameters. 1239 [Steve Henson] 1240 1241 *) Add callbacks for arbitrary TLS extensions. 1242 [Trevor Perrin <trevp@trevp.net> and Ben Laurie] 1243 1244 *) New option -crl_download in several openssl utilities to download CRLs 1245 from CRLDP extension in certificates. 1246 [Steve Henson] 1247 1248 *) New options -CRL and -CRLform for s_client and s_server for CRLs. 1249 [Steve Henson] 1250 1251 *) New function X509_CRL_diff to generate a delta CRL from the difference 1252 of two full CRLs. Add support to "crl" utility. 1253 [Steve Henson] 1254 1255 *) New functions to set lookup_crls function and to retrieve 1256 X509_STORE from X509_STORE_CTX. 1257 [Steve Henson] 1258 1259 *) Print out deprecated issuer and subject unique ID fields in 1260 certificates. 1261 [Steve Henson] 1262 1263 *) Extend OCSP I/O functions so they can be used for simple general purpose 1264 HTTP as well as OCSP. New wrapper function which can be used to download 1265 CRLs using the OCSP API. 1266 [Steve Henson] 1267 1268 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs. 1269 [Steve Henson] 1270 1271 *) SSL_CONF* functions. These provide a common framework for application 1272 configuration using configuration files or command lines. 1273 [Steve Henson] 1274 1275 *) SSL/TLS tracing code. This parses out SSL/TLS records using the 1276 message callback and prints the results. Needs compile time option 1277 "enable-ssl-trace". New options to s_client and s_server to enable 1278 tracing. 1279 [Steve Henson] 1280 1281 *) New ctrl and macro to retrieve supported points extensions. 1282 Print out extension in s_server and s_client. 1283 [Steve Henson] 1284 1285 *) New functions to retrieve certificate signature and signature 1286 OID NID. 1287 [Steve Henson] 1288 1289 *) Add functions to retrieve and manipulate the raw cipherlist sent by a 1290 client to OpenSSL. 1291 [Steve Henson] 1292 1293 *) New Suite B modes for TLS code. These use and enforce the requirements 1294 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and 1295 only use Suite B curves. The Suite B modes can be set by using the 1296 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring. 1297 [Steve Henson] 1298 1299 *) New chain verification flags for Suite B levels of security. Check 1300 algorithms are acceptable when flags are set in X509_verify_cert. 1301 [Steve Henson] 1302 1303 *) Make tls1_check_chain return a set of flags indicating checks passed 1304 by a certificate chain. Add additional tests to handle client 1305 certificates: checks for matching certificate type and issuer name 1306 comparison. 1307 [Steve Henson] 1308 1309 *) If an attempt is made to use a signature algorithm not in the peer 1310 preference list abort the handshake. If client has no suitable 1311 signature algorithms in response to a certificate request do not 1312 use the certificate. 1313 [Steve Henson] 1314 1315 *) If server EC tmp key is not in client preference list abort handshake. 1316 [Steve Henson] 1317 1318 *) Add support for certificate stores in CERT structure. This makes it 1319 possible to have different stores per SSL structure or one store in 1320 the parent SSL_CTX. Include distint stores for certificate chain 1321 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN 1322 to build and store a certificate chain in CERT structure: returing 1323 an error if the chain cannot be built: this will allow applications 1324 to test if a chain is correctly configured. 1325 1326 Note: if the CERT based stores are not set then the parent SSL_CTX 1327 store is used to retain compatibility with existing behaviour. 1328 1329 [Steve Henson] 1330 1331 *) New function ssl_set_client_disabled to set a ciphersuite disabled 1332 mask based on the current session, check mask when sending client 1333 hello and checking the requested ciphersuite. 1334 [Steve Henson] 1335 1336 *) New ctrls to retrieve and set certificate types in a certificate 1337 request message. Print out received values in s_client. If certificate 1338 types is not set with custom values set sensible values based on 1339 supported signature algorithms. 1340 [Steve Henson] 1341 1342 *) Support for distinct client and server supported signature algorithms. 1343 [Steve Henson] 1344 1345 *) Add certificate callback. If set this is called whenever a certificate 1346 is required by client or server. An application can decide which 1347 certificate chain to present based on arbitrary criteria: for example 1348 supported signature algorithms. Add very simple example to s_server. 1349 This fixes many of the problems and restrictions of the existing client 1350 certificate callback: for example you can now clear an existing 1351 certificate and specify the whole chain. 1352 [Steve Henson] 1353 1354 *) Add new "valid_flags" field to CERT_PKEY structure which determines what 1355 the certificate can be used for (if anything). Set valid_flags field 1356 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used 1357 to have similar checks in it. 1358 1359 Add new "cert_flags" field to CERT structure and include a "strict mode". 1360 This enforces some TLS certificate requirements (such as only permitting 1361 certificate signature algorithms contained in the supported algorithms 1362 extension) which some implementations ignore: this option should be used 1363 with caution as it could cause interoperability issues. 1364 [Steve Henson] 1365 1366 *) Update and tidy signature algorithm extension processing. Work out 1367 shared signature algorithms based on preferences and peer algorithms 1368 and print them out in s_client and s_server. Abort handshake if no 1369 shared signature algorithms. 1370 [Steve Henson] 1371 1372 *) Add new functions to allow customised supported signature algorithms 1373 for SSL and SSL_CTX structures. Add options to s_client and s_server 1374 to support them. 1375 [Steve Henson] 1376 1377 *) New function SSL_certs_clear() to delete all references to certificates 1378 from an SSL structure. Before this once a certificate had been added 1379 it couldn't be removed. 1380 [Steve Henson] 1381 1382 *) Integrate hostname, email address and IP address checking with certificate 1383 verification. New verify options supporting checking in opensl utility. 1384 [Steve Henson] 1385 1386 *) Fixes and wildcard matching support to hostname and email checking 1387 functions. Add manual page. 1388 [Florian Weimer (Red Hat Product Security Team)] 1389 1390 *) New functions to check a hostname email or IP address against a 1391 certificate. Add options x509 utility to print results of checks against 1392 a certificate. 1393 [Steve Henson] 1394 1395 *) Fix OCSP checking. 1396 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie] 1397 1398 *) Initial experimental support for explicitly trusted non-root CAs. 1399 OpenSSL still tries to build a complete chain to a root but if an 1400 intermediate CA has a trust setting included that is used. The first 1401 setting is used: whether to trust (e.g., -addtrust option to the x509 1402 utility) or reject. 1403 [Steve Henson] 1404 1405 *) Add -trusted_first option which attempts to find certificates in the 1406 trusted store even if an untrusted chain is also supplied. 1407 [Steve Henson] 1408 1409 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE, 1410 platform support for Linux and Android. 1411 [Andy Polyakov] 1412 1413 *) Support for linux-x32, ILP32 environment in x86_64 framework. 1414 [Andy Polyakov] 1415 1416 *) Experimental multi-implementation support for FIPS capable OpenSSL. 1417 When in FIPS mode the approved implementations are used as normal, 1418 when not in FIPS mode the internal unapproved versions are used instead. 1419 This means that the FIPS capable OpenSSL isn't forced to use the 1420 (often lower perfomance) FIPS implementations outside FIPS mode. 1421 [Steve Henson] 1422 1423 *) Transparently support X9.42 DH parameters when calling 1424 PEM_read_bio_DHparameters. This means existing applications can handle 1425 the new parameter format automatically. 1426 [Steve Henson] 1427 1428 *) Initial experimental support for X9.42 DH parameter format: mainly 1429 to support use of 'q' parameter for RFC5114 parameters. 1430 [Steve Henson] 1431 1432 *) Add DH parameters from RFC5114 including test data to dhtest. 1433 [Steve Henson] 1434 1435 *) Support for automatic EC temporary key parameter selection. If enabled 1436 the most preferred EC parameters are automatically used instead of 1437 hardcoded fixed parameters. Now a server just has to call: 1438 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically 1439 support ECDH and use the most appropriate parameters. 1440 [Steve Henson] 1441 1442 *) Enhance and tidy EC curve and point format TLS extension code. Use 1443 static structures instead of allocation if default values are used. 1444 New ctrls to set curves we wish to support and to retrieve shared curves. 1445 Print out shared curves in s_server. New options to s_server and s_client 1446 to set list of supported curves. 1447 [Steve Henson] 1448 1449 *) New ctrls to retrieve supported signature algorithms and 1450 supported curve values as an array of NIDs. Extend openssl utility 1451 to print out received values. 1452 [Steve Henson] 1453 1454 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert 1455 between NIDs and the more common NIST names such as "P-256". Enhance 1456 ecparam utility and ECC method to recognise the NIST names for curves. 1457 [Steve Henson] 1458 1459 *) Enhance SSL/TLS certificate chain handling to support different 1460 chains for each certificate instead of one chain in the parent SSL_CTX. 1461 [Steve Henson] 1462 1463 *) Support for fixed DH ciphersuite client authentication: where both 1464 server and client use DH certificates with common parameters. 1465 [Steve Henson] 1466 1467 *) Support for fixed DH ciphersuites: those requiring DH server 1468 certificates. 1469 [Steve Henson] 1470 1471 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of 1472 the certificate. 1473 Note: Related 1.0.2-beta specific macros X509_get_cert_info, 1474 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and 1475 X509_CINF_get_signature were reverted post internal team review. 1476 1477 Changes between 1.0.1k and 1.0.1l [15 Jan 2015] 1478 1479 *) Build fixes for the Windows and OpenVMS platforms 1480 [Matt Caswell and Richard Levitte] 1481 1482 Changes between 1.0.1j and 1.0.1k [8 Jan 2015] 1483 1484 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS 1485 message can cause a segmentation fault in OpenSSL due to a NULL pointer 1486 dereference. This could lead to a Denial Of Service attack. Thanks to 1487 Markus Stenberg of Cisco Systems, Inc. for reporting this issue. 1488 (CVE-2014-3571) 1489 [Steve Henson] 1490 1491 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the 1492 dtls1_buffer_record function under certain conditions. In particular this 1493 could occur if an attacker sent repeated DTLS records with the same 1494 sequence number but for the next epoch. The memory leak could be exploited 1495 by an attacker in a Denial of Service attack through memory exhaustion. 1496 Thanks to Chris Mueller for reporting this issue. 1497 (CVE-2015-0206) 1498 [Matt Caswell] 1499 1500 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is 1501 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl 1502 method would be set to NULL which could later result in a NULL pointer 1503 dereference. Thanks to Frank Schmirler for reporting this issue. 1504 (CVE-2014-3569) 1505 [Kurt Roeckx] 1506 1507 *) Abort handshake if server key exchange message is omitted for ephemeral 1508 ECDH ciphersuites. 1509 1510 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for 1511 reporting this issue. 1512 (CVE-2014-3572) 1513 [Steve Henson] 1514 1515 *) Remove non-export ephemeral RSA code on client and server. This code 1516 violated the TLS standard by allowing the use of temporary RSA keys in 1517 non-export ciphersuites and could be used by a server to effectively 1518 downgrade the RSA key length used to a value smaller than the server 1519 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at 1520 INRIA or reporting this issue. 1521 (CVE-2015-0204) 1522 [Steve Henson] 1523 1524 *) Fixed issue where DH client certificates are accepted without verification. 1525 An OpenSSL server will accept a DH certificate for client authentication 1526 without the certificate verify message. This effectively allows a client to 1527 authenticate without the use of a private key. This only affects servers 1528 which trust a client certificate authority which issues certificates 1529 containing DH keys: these are extremely rare and hardly ever encountered. 1530 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting 1531 this issue. 1532 (CVE-2015-0205) 1533 [Steve Henson] 1534 1535 *) Ensure that the session ID context of an SSL is updated when its 1536 SSL_CTX is updated via SSL_set_SSL_CTX. 1537 1538 The session ID context is typically set from the parent SSL_CTX, 1539 and can vary with the CTX. 1540 [Adam Langley] 1541 1542 *) Fix various certificate fingerprint issues. 1543 1544 By using non-DER or invalid encodings outside the signed portion of a 1545 certificate the fingerprint can be changed without breaking the signature. 1546 Although no details of the signed portion of the certificate can be changed 1547 this can cause problems with some applications: e.g. those using the 1548 certificate fingerprint for blacklists. 1549 1550 1. Reject signatures with non zero unused bits. 1551 1552 If the BIT STRING containing the signature has non zero unused bits reject 1553 the signature. All current signature algorithms require zero unused bits. 1554 1555 2. Check certificate algorithm consistency. 1556 1557 Check the AlgorithmIdentifier inside TBS matches the one in the 1558 certificate signature. NB: this will result in signature failure 1559 errors for some broken certificates. 1560 1561 Thanks to Konrad Kraszewski from Google for reporting this issue. 1562 1563 3. Check DSA/ECDSA signatures use DER. 1564 1565 Reencode DSA/ECDSA signatures and compare with the original received 1566 signature. Return an error if there is a mismatch. 1567 1568 This will reject various cases including garbage after signature 1569 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS 1570 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs 1571 (negative or with leading zeroes). 1572 1573 Further analysis was conducted and fixes were developed by Stephen Henson 1574 of the OpenSSL core team. 1575 1576 (CVE-2014-8275) 1577 [Steve Henson] 1578 1579 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect 1580 results on some platforms, including x86_64. This bug occurs at random 1581 with a very low probability, and is not known to be exploitable in any 1582 way, though its exact impact is difficult to determine. Thanks to Pieter 1583 Wuille (Blockstream) who reported this issue and also suggested an initial 1584 fix. Further analysis was conducted by the OpenSSL development team and 1585 Adam Langley of Google. The final fix was developed by Andy Polyakov of 1586 the OpenSSL core team. 1587 (CVE-2014-3570) 1588 [Andy Polyakov] 1589 1590 *) Do not resume sessions on the server if the negotiated protocol 1591 version does not match the session's version. Resuming with a different 1592 version, while not strictly forbidden by the RFC, is of questionable 1593 sanity and breaks all known clients. 1594 [David Benjamin, Emilia K��sper] 1595 1596 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject 1597 early CCS messages during renegotiation. (Note that because 1598 renegotiation is encrypted, this early CCS was not exploitable.) 1599 [Emilia K��sper] 1600 1601 *) Tighten client-side session ticket handling during renegotiation: 1602 ensure that the client only accepts a session ticket if the server sends 1603 the extension anew in the ServerHello. Previously, a TLS client would 1604 reuse the old extension state and thus accept a session ticket if one was 1605 announced in the initial ServerHello. 1606 1607 Similarly, ensure that the client requires a session ticket if one 1608 was advertised in the ServerHello. Previously, a TLS client would 1609 ignore a missing NewSessionTicket message. 1610 [Emilia K��sper] 1611 1612 Changes between 1.0.1i and 1.0.1j [15 Oct 2014] 1613 1614 *) SRTP Memory Leak. 1615 1616 A flaw in the DTLS SRTP extension parsing code allows an attacker, who 1617 sends a carefully crafted handshake message, to cause OpenSSL to fail 1618 to free up to 64k of memory causing a memory leak. This could be 1619 exploited in a Denial Of Service attack. This issue affects OpenSSL 1620 1.0.1 server implementations for both SSL/TLS and DTLS regardless of 1621 whether SRTP is used or configured. Implementations of OpenSSL that 1622 have been compiled with OPENSSL_NO_SRTP defined are not affected. 1623 1624 The fix was developed by the OpenSSL team. 1625 (CVE-2014-3513) 1626 [OpenSSL team] 1627 1628 *) Session Ticket Memory Leak. 1629 1630 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the 1631 integrity of that ticket is first verified. In the event of a session 1632 ticket integrity check failing, OpenSSL will fail to free memory 1633 causing a memory leak. By sending a large number of invalid session 1634 tickets an attacker could exploit this issue in a Denial Of Service 1635 attack. 1636 (CVE-2014-3567) 1637 [Steve Henson] 1638 1639 *) Build option no-ssl3 is incomplete. 1640 1641 When OpenSSL is configured with "no-ssl3" as a build option, servers 1642 could accept and complete a SSL 3.0 handshake, and clients could be 1643 configured to send them. 1644 (CVE-2014-3568) 1645 [Akamai and the OpenSSL team] 1646 1647 *) Add support for TLS_FALLBACK_SCSV. 1648 Client applications doing fallback retries should call 1649 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV). 1650 (CVE-2014-3566) 1651 [Adam Langley, Bodo Moeller] 1652 1653 *) Add additional DigestInfo checks. 1654 1655 Reencode DigestInto in DER and check against the original when 1656 verifying RSA signature: this will reject any improperly encoded 1657 DigestInfo structures. 1658 1659 Note: this is a precautionary measure and no attacks are currently known. 1660 1661 [Steve Henson] 1662 1663 Changes between 1.0.1h and 1.0.1i [6 Aug 2014] 1664 1665 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the 1666 SRP code can be overrun an internal buffer. Add sanity check that 1667 g, A, B < N to SRP code. 1668 1669 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC 1670 Group for discovering this issue. 1671 (CVE-2014-3512) 1672 [Steve Henson] 1673 1674 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate 1675 TLS 1.0 instead of higher protocol versions when the ClientHello message 1676 is badly fragmented. This allows a man-in-the-middle attacker to force a 1677 downgrade to TLS 1.0 even if both the server and the client support a 1678 higher protocol version, by modifying the client's TLS records. 1679 1680 Thanks to David Benjamin and Adam Langley (Google) for discovering and 1681 researching this issue. 1682 (CVE-2014-3511) 1683 [David Benjamin] 1684 1685 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject 1686 to a denial of service attack. A malicious server can crash the client 1687 with a null pointer dereference (read) by specifying an anonymous (EC)DH 1688 ciphersuite and sending carefully crafted handshake messages. 1689 1690 Thanks to Felix Gr��bert (Google) for discovering and researching this 1691 issue. 1692 (CVE-2014-3510) 1693 [Emilia K��sper] 1694 1695 *) By sending carefully crafted DTLS packets an attacker could cause openssl 1696 to leak memory. This can be exploited through a Denial of Service attack. 1697 Thanks to Adam Langley for discovering and researching this issue. 1698 (CVE-2014-3507) 1699 [Adam Langley] 1700 1701 *) An attacker can force openssl to consume large amounts of memory whilst 1702 processing DTLS handshake messages. This can be exploited through a 1703 Denial of Service attack. 1704 Thanks to Adam Langley for discovering and researching this issue. 1705 (CVE-2014-3506) 1706 [Adam Langley] 1707 1708 *) An attacker can force an error condition which causes openssl to crash 1709 whilst processing DTLS packets due to memory being freed twice. This 1710 can be exploited through a Denial of Service attack. 1711 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching 1712 this issue. 1713 (CVE-2014-3505) 1714 [Adam Langley] 1715 1716 *) If a multithreaded client connects to a malicious server using a resumed 1717 session and the server sends an ec point format extension it could write 1718 up to 255 bytes to freed memory. 1719 1720 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this 1721 issue. 1722 (CVE-2014-3509) 1723 [Gabor Tyukasz] 1724 1725 *) A malicious server can crash an OpenSSL client with a null pointer 1726 dereference (read) by specifying an SRP ciphersuite even though it was not 1727 properly negotiated with the client. This can be exploited through a 1728 Denial of Service attack. 1729 1730 Thanks to Joonas Kuorilehto and Riku Hietam��ki (Codenomicon) for 1731 discovering and researching this issue. 1732 (CVE-2014-5139) 1733 [Steve Henson] 1734 1735 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as 1736 X509_name_oneline, X509_name_print_ex et al. to leak some information 1737 from the stack. Applications may be affected if they echo pretty printing 1738 output to the attacker. 1739 1740 Thanks to Ivan Fratric (Google) for discovering this issue. 1741 (CVE-2014-3508) 1742 [Emilia K��sper, and Steve Henson] 1743 1744 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) 1745 for corner cases. (Certain input points at infinity could lead to 1746 bogus results, with non-infinity inputs mapped to infinity too.) 1747 [Bodo Moeller] 1748 1749 Changes between 1.0.1g and 1.0.1h [5 Jun 2014] 1750 1751 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted 1752 handshake can force the use of weak keying material in OpenSSL 1753 SSL/TLS clients and servers. 1754 1755 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and 1756 researching this issue. (CVE-2014-0224) 1757 [KIKUCHI Masashi, Steve Henson] 1758 1759 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an 1760 OpenSSL DTLS client the code can be made to recurse eventually crashing 1761 in a DoS attack. 1762 1763 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue. 1764 (CVE-2014-0221) 1765 [Imre Rad, Steve Henson] 1766 1767 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can 1768 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS 1769 client or server. This is potentially exploitable to run arbitrary 1770 code on a vulnerable client or server. 1771 1772 Thanks to J��ri Aedla for reporting this issue. (CVE-2014-0195) 1773 [J��ri Aedla, Steve Henson] 1774 1775 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites 1776 are subject to a denial of service attack. 1777 1778 Thanks to Felix Gr��bert and Ivan Fratric at Google for discovering 1779 this issue. (CVE-2014-3470) 1780 [Felix Gr��bert, Ivan Fratric, Steve Henson] 1781 1782 *) Harmonize version and its documentation. -f flag is used to display 1783 compilation flags. 1784 [mancha <mancha1@zoho.com>] 1785 1786 *) Fix eckey_priv_encode so it immediately returns an error upon a failure 1787 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue. 1788 [mancha <mancha1@zoho.com>] 1789 1790 *) Fix some double frees. These are not thought to be exploitable. 1791 [mancha <mancha1@zoho.com>] 1792 1793 Changes between 1.0.1f and 1.0.1g [7 Apr 2014] 1794 1795 *) A missing bounds check in the handling of the TLS heartbeat extension 1796 can be used to reveal up to 64k of memory to a connected client or 1797 server. 1798 1799 Thanks for Neel Mehta of Google Security for discovering this bug and to 1800 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for 1801 preparing the fix (CVE-2014-0160) 1802 [Adam Langley, Bodo Moeller] 1803 1804 *) Fix for the attack described in the paper "Recovering OpenSSL 1805 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack" 1806 by Yuval Yarom and Naomi Benger. Details can be obtained from: 1807 http://eprint.iacr.org/2014/140 1808 1809 Thanks to Yuval Yarom and Naomi Benger for discovering this 1810 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076) 1811 [Yuval Yarom and Naomi Benger] 1812 1813 *) TLS pad extension: draft-agl-tls-padding-03 1814 1815 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the 1816 TLS client Hello record length value would otherwise be > 255 and 1817 less that 512 pad with a dummy extension containing zeroes so it 1818 is at least 512 bytes long. 1819 1820 [Adam Langley, Steve Henson] 1821 1822 Changes between 1.0.1e and 1.0.1f [6 Jan 2014] 1823 1824 *) Fix for TLS record tampering bug. A carefully crafted invalid 1825 handshake could crash OpenSSL with a NULL pointer exception. 1826 Thanks to Anton Johansson for reporting this issues. 1827 (CVE-2013-4353) 1828 1829 *) Keep original DTLS digest and encryption contexts in retransmission 1830 structures so we can use the previous session parameters if they need 1831 to be resent. (CVE-2013-6450) 1832 [Steve Henson] 1833 1834 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which 1835 avoids preferring ECDHE-ECDSA ciphers when the client appears to be 1836 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for 1837 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug 1838 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing 1839 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer. 1840 [Rob Stradling, Adam Langley] 1841 1842 Changes between 1.0.1d and 1.0.1e [11 Feb 2013] 1843 1844 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI 1845 supporting platforms or when small records were transferred. 1846 [Andy Polyakov, Steve Henson] 1847 1848 Changes between 1.0.1c and 1.0.1d [5 Feb 2013] 1849 1850 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 1851 1852 This addresses the flaw in CBC record processing discovered by 1853 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 1854 at: http://www.isg.rhul.ac.uk/tls/ 1855 1856 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 1857 Security Group at Royal Holloway, University of London 1858 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 1859 Emilia K��sper for the initial patch. 1860 (CVE-2013-0169) 1861 [Emilia K��sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 1862 1863 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode 1864 ciphersuites which can be exploited in a denial of service attack. 1865 Thanks go to and to Adam Langley <agl@chromium.org> for discovering 1866 and detecting this bug and to Wolfgang Ettlinger 1867 <wolfgang.ettlinger@gmail.com> for independently discovering this issue. 1868 (CVE-2012-2686) 1869 [Adam Langley] 1870 1871 *) Return an error when checking OCSP signatures when key is NULL. 1872 This fixes a DoS attack. (CVE-2013-0166) 1873 [Steve Henson] 1874 1875 *) Make openssl verify return errors. 1876 [Chris Palmer <palmer@google.com> and Ben Laurie] 1877 1878 *) Call OCSP Stapling callback after ciphersuite has been chosen, so 1879 the right response is stapled. Also change SSL_get_certificate() 1880 so it returns the certificate actually sent. 1881 See http://rt.openssl.org/Ticket/Display.html?id=2836. 1882 [Rob Stradling <rob.stradling@comodo.com>] 1883 1884 *) Fix possible deadlock when decoding public keys. 1885 [Steve Henson] 1886 1887 *) Don't use TLS 1.0 record version number in initial client hello 1888 if renegotiating. 1889 [Steve Henson] 1890 1891 Changes between 1.0.1b and 1.0.1c [10 May 2012] 1892 1893 *) Sanity check record length before skipping explicit IV in TLS 1894 1.2, 1.1 and DTLS to fix DoS attack. 1895 1896 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 1897 fuzzing as a service testing platform. 1898 (CVE-2012-2333) 1899 [Steve Henson] 1900 1901 *) Initialise tkeylen properly when encrypting CMS messages. 1902 Thanks to Solar Designer of Openwall for reporting this issue. 1903 [Steve Henson] 1904 1905 *) In FIPS mode don't try to use composite ciphers as they are not 1906 approved. 1907 [Steve Henson] 1908 1909 Changes between 1.0.1a and 1.0.1b [26 Apr 2012] 1910 1911 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and 1912 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately 1913 mean any application compiled against OpenSSL 1.0.0 headers setting 1914 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng 1915 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to 1916 0x10000000L Any application which was previously compiled against 1917 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1 1918 will need to be recompiled as a result. Letting be results in 1919 inability to disable specifically TLS 1.1 and in client context, 1920 in unlike event, limit maximum offered version to TLS 1.0 [see below]. 1921 [Steve Henson] 1922 1923 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not 1924 disable just protocol X, but all protocols above X *if* there are 1925 protocols *below* X still enabled. In more practical terms it means 1926 that if application wants to disable TLS1.0 in favor of TLS1.1 and 1927 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass 1928 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to 1929 client side. 1930 [Andy Polyakov] 1931 1932 Changes between 1.0.1 and 1.0.1a [19 Apr 2012] 1933 1934 *) Check for potentially exploitable overflows in asn1_d2i_read_bio 1935 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 1936 in CRYPTO_realloc_clean. 1937 1938 Thanks to Tavis Ormandy, Google Security Team, for discovering this 1939 issue and to Adam Langley <agl@chromium.org> for fixing it. 1940 (CVE-2012-2110) 1941 [Adam Langley (Google), Tavis Ormandy, Google Security Team] 1942 1943 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections. 1944 [Adam Langley] 1945 1946 *) Workarounds for some broken servers that "hang" if a client hello 1947 record length exceeds 255 bytes. 1948 1949 1. Do not use record version number > TLS 1.0 in initial client 1950 hello: some (but not all) hanging servers will now work. 1951 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate 1952 the number of ciphers sent in the client hello. This should be 1953 set to an even number, such as 50, for example by passing: 1954 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure. 1955 Most broken servers should now work. 1956 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable 1957 TLS 1.2 client support entirely. 1958 [Steve Henson] 1959 1960 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH. 1961 [Andy Polyakov] 1962 1963 Changes between 1.0.0h and 1.0.1 [14 Mar 2012] 1964 1965 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET 1966 STRING form instead of a DigestInfo. 1967 [Steve Henson] 1968 1969 *) The format used for MDC2 RSA signatures is inconsistent between EVP 1970 and the RSA_sign/RSA_verify functions. This was made more apparent when 1971 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular 1972 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect 1973 the correct format in RSA_verify so both forms transparently work. 1974 [Steve Henson] 1975 1976 *) Some servers which support TLS 1.0 can choke if we initially indicate 1977 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA 1978 encrypted premaster secret. As a workaround use the maximum pemitted 1979 client version in client hello, this should keep such servers happy 1980 and still work with previous versions of OpenSSL. 1981 [Steve Henson] 1982 1983 *) Add support for TLS/DTLS heartbeats. 1984 [Robin Seggelmann <seggelmann@fh-muenster.de>] 1985 1986 *) Add support for SCTP. 1987 [Robin Seggelmann <seggelmann@fh-muenster.de>] 1988 1989 *) Improved PRNG seeding for VOS. 1990 [Paul Green <Paul.Green@stratus.com>] 1991 1992 *) Extensive assembler packs updates, most notably: 1993 1994 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support; 1995 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES); 1996 - x86_64: bit-sliced AES implementation; 1997 - ARM: NEON support, contemporary platforms optimizations; 1998 - s390x: z196 support; 1999 - *: GHASH and GF(2^m) multiplication implementations; 2000 2001 [Andy Polyakov] 2002 2003 *) Make TLS-SRP code conformant with RFC 5054 API cleanup 2004 (removal of unnecessary code) 2005 [Peter Sylvester <peter.sylvester@edelweb.fr>] 2006 2007 *) Add TLS key material exporter from RFC 5705. 2008 [Eric Rescorla] 2009 2010 *) Add DTLS-SRTP negotiation from RFC 5764. 2011 [Eric Rescorla] 2012 2013 *) Add Next Protocol Negotiation, 2014 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be 2015 disabled with a no-npn flag to config or Configure. Code donated 2016 by Google. 2017 [Adam Langley <agl@google.com> and Ben Laurie] 2018 2019 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224, 2020 NIST-P256, NIST-P521, with constant-time single point multiplication on 2021 typical inputs. Compiler support for the nonstandard type __uint128_t is 2022 required to use this (present in gcc 4.4 and later, for 64-bit builds). 2023 Code made available under Apache License version 2.0. 2024 2025 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command 2026 line to include this in your build of OpenSSL, and run "make depend" (or 2027 "make update"). This enables the following EC_METHODs: 2028 2029 EC_GFp_nistp224_method() 2030 EC_GFp_nistp256_method() 2031 EC_GFp_nistp521_method() 2032 2033 EC_GROUP_new_by_curve_name() will automatically use these (while 2034 EC_GROUP_new_curve_GFp() currently prefers the more flexible 2035 implementations). 2036 [Emilia K��sper, Adam Langley, Bodo Moeller (Google)] 2037 2038 *) Use type ossl_ssize_t instad of ssize_t which isn't available on 2039 all platforms. Move ssize_t definition from e_os.h to the public 2040 header file e_os2.h as it now appears in public header file cms.h 2041 [Steve Henson] 2042 2043 *) New -sigopt option to the ca, req and x509 utilities. Additional 2044 signature parameters can be passed using this option and in 2045 particular PSS. 2046 [Steve Henson] 2047 2048 *) Add RSA PSS signing function. This will generate and set the 2049 appropriate AlgorithmIdentifiers for PSS based on those in the 2050 corresponding EVP_MD_CTX structure. No application support yet. 2051 [Steve Henson] 2052 2053 *) Support for companion algorithm specific ASN1 signing routines. 2054 New function ASN1_item_sign_ctx() signs a pre-initialised 2055 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on 2056 the appropriate parameters. 2057 [Steve Henson] 2058 2059 *) Add new algorithm specific ASN1 verification initialisation function 2060 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1 2061 handling will be the same no matter what EVP_PKEY_METHOD is used. 2062 Add a PSS handler to support verification of PSS signatures: checked 2063 against a number of sample certificates. 2064 [Steve Henson] 2065 2066 *) Add signature printing for PSS. Add PSS OIDs. 2067 [Steve Henson, Martin Kaiser <lists@kaiser.cx>] 2068 2069 *) Add algorithm specific signature printing. An individual ASN1 method 2070 can now print out signatures instead of the standard hex dump. 2071 2072 More complex signatures (e.g. PSS) can print out more meaningful 2073 information. Include DSA version that prints out the signature 2074 parameters r, s. 2075 [Steve Henson] 2076 2077 *) Password based recipient info support for CMS library: implementing 2078 RFC3211. 2079 [Steve Henson] 2080 2081 *) Split password based encryption into PBES2 and PBKDF2 functions. This 2082 neatly separates the code into cipher and PBE sections and is required 2083 for some algorithms that split PBES2 into separate pieces (such as 2084 password based CMS). 2085 [Steve Henson] 2086 2087 *) Session-handling fixes: 2088 - Fix handling of connections that are resuming with a session ID, 2089 but also support Session Tickets. 2090 - Fix a bug that suppressed issuing of a new ticket if the client 2091 presented a ticket with an expired session. 2092 - Try to set the ticket lifetime hint to something reasonable. 2093 - Make tickets shorter by excluding irrelevant information. 2094 - On the client side, don't ignore renewed tickets. 2095 [Adam Langley, Bodo Moeller (Google)] 2096 2097 *) Fix PSK session representation. 2098 [Bodo Moeller] 2099 2100 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations. 2101 2102 This work was sponsored by Intel. 2103 [Andy Polyakov] 2104 2105 *) Add GCM support to TLS library. Some custom code is needed to split 2106 the IV between the fixed (from PRF) and explicit (from TLS record) 2107 portions. This adds all GCM ciphersuites supported by RFC5288 and 2108 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and 2109 add a special AESGCM string for GCM only. 2110 [Steve Henson] 2111 2112 *) Expand range of ctrls for AES GCM. Permit setting invocation 2113 field on decrypt and retrieval of invocation field only on encrypt. 2114 [Steve Henson] 2115 2116 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support. 2117 As required by RFC5289 these ciphersuites cannot be used if for 2118 versions of TLS earlier than 1.2. 2119 [Steve Henson] 2120 2121 *) For FIPS capable OpenSSL interpret a NULL default public key method 2122 as unset and return the appopriate default but do *not* set the default. 2123 This means we can return the appopriate method in applications that 2124 swicth between FIPS and non-FIPS modes. 2125 [Steve Henson] 2126 2127 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an 2128 ENGINE is used then we cannot handle that in the FIPS module so we 2129 keep original code iff non-FIPS operations are allowed. 2130 [Steve Henson] 2131 2132 *) Add -attime option to openssl utilities. 2133 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson] 2134 2135 *) Redirect DSA and DH operations to FIPS module in FIPS mode. 2136 [Steve Henson] 2137 2138 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use 2139 FIPS EC methods unconditionally for now. 2140 [Steve Henson] 2141 2142 *) New build option no-ec2m to disable characteristic 2 code. 2143 [Steve Henson] 2144 2145 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not 2146 all cases can be covered as some introduce binary incompatibilities. 2147 [Steve Henson] 2148 2149 *) Redirect RSA operations to FIPS module including keygen, 2150 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods. 2151 [Steve Henson] 2152 2153 *) Add similar low level API blocking to ciphers. 2154 [Steve Henson] 2155 2156 *) Low level digest APIs are not approved in FIPS mode: any attempt 2157 to use these will cause a fatal error. Applications that *really* want 2158 to use them can use the private_* version instead. 2159 [Steve Henson] 2160 2161 *) Redirect cipher operations to FIPS module for FIPS builds. 2162 [Steve Henson] 2163 2164 *) Redirect digest operations to FIPS module for FIPS builds. 2165 [Steve Henson] 2166 2167 *) Update build system to add "fips" flag which will link in fipscanister.o 2168 for static and shared library builds embedding a signature if needed. 2169 [Steve Henson] 2170 2171 *) Output TLS supported curves in preference order instead of numerical 2172 order. This is currently hardcoded for the highest order curves first. 2173 This should be configurable so applications can judge speed vs strength. 2174 [Steve Henson] 2175 2176 *) Add TLS v1.2 server support for client authentication. 2177 [Steve Henson] 2178 2179 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers 2180 and enable MD5. 2181 [Steve Henson] 2182 2183 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying 2184 FIPS modules versions. 2185 [Steve Henson] 2186 2187 *) Add TLS v1.2 client side support for client authentication. Keep cache 2188 of handshake records longer as we don't know the hash algorithm to use 2189 until after the certificate request message is received. 2190 [Steve Henson] 2191 2192 *) Initial TLS v1.2 client support. Add a default signature algorithms 2193 extension including all the algorithms we support. Parse new signature 2194 format in client key exchange. Relax some ECC signing restrictions for 2195 TLS v1.2 as indicated in RFC5246. 2196 [Steve Henson] 2197 2198 *) Add server support for TLS v1.2 signature algorithms extension. Switch 2199 to new signature format when needed using client digest preference. 2200 All server ciphersuites should now work correctly in TLS v1.2. No client 2201 support yet and no support for client certificates. 2202 [Steve Henson] 2203 2204 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch 2205 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based 2206 ciphersuites. At present only RSA key exchange ciphersuites work with 2207 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete 2208 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods 2209 and version checking. 2210 [Steve Henson] 2211 2212 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled 2213 with this defined it will not be affected by any changes to ssl internal 2214 structures. Add several utility functions to allow openssl application 2215 to work with OPENSSL_NO_SSL_INTERN defined. 2216 [Steve Henson] 2217 2218 *) A long standing patch to add support for SRP from EdelWeb (Peter 2219 Sylvester and Christophe Renou) was integrated. 2220 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester 2221 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and 2222 Ben Laurie] 2223 2224 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id. 2225 [Steve Henson] 2226 2227 *) Permit abbreviated handshakes when renegotiating using the function 2228 SSL_renegotiate_abbreviated(). 2229 [Robin Seggelmann <seggelmann@fh-muenster.de>] 2230 2231 *) Add call to ENGINE_register_all_complete() to 2232 ENGINE_load_builtin_engines(), so some implementations get used 2233 automatically instead of needing explicit application support. 2234 [Steve Henson] 2235 2236 *) Add support for TLS key exporter as described in RFC5705. 2237 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson] 2238 2239 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only 2240 a few changes are required: 2241 2242 Add SSL_OP_NO_TLSv1_1 flag. 2243 Add TLSv1_1 methods. 2244 Update version checking logic to handle version 1.1. 2245 Add explicit IV handling (ported from DTLS code). 2246 Add command line options to s_client/s_server. 2247 [Steve Henson] 2248 2249 Changes between 1.0.0g and 1.0.0h [12 Mar 2012] 2250 2251 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 2252 in CMS and PKCS7 code. When RSA decryption fails use a random key for 2253 content decryption and always return the same error. Note: this attack 2254 needs on average 2^20 messages so it only affects automated senders. The 2255 old behaviour can be reenabled in the CMS code by setting the 2256 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 2257 an MMA defence is not necessary. 2258 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 2259 this issue. (CVE-2012-0884) 2260 [Steve Henson] 2261 2262 *) Fix CVE-2011-4619: make sure we really are receiving a 2263 client hello before rejecting multiple SGC restarts. Thanks to 2264 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 2265 [Steve Henson] 2266 2267 Changes between 1.0.0f and 1.0.0g [18 Jan 2012] 2268 2269 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 2270 Thanks to Antonio Martin, Enterprise Secure Access Research and 2271 Development, Cisco Systems, Inc. for discovering this bug and 2272 preparing a fix. (CVE-2012-0050) 2273 [Antonio Martin] 2274 2275 Changes between 1.0.0e and 1.0.0f [4 Jan 2012] 2276 2277 *) Nadhem Alfardan and Kenny Paterson have discovered an extension 2278 of the Vaudenay padding oracle attack on CBC mode encryption 2279 which enables an efficient plaintext recovery attack against 2280 the OpenSSL implementation of DTLS. Their attack exploits timing 2281 differences arising during decryption processing. A research 2282 paper describing this attack can be found at: 2283 http://www.isg.rhul.ac.uk/~kp/dtls.pdf 2284 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 2285 Security Group at Royal Holloway, University of London 2286 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 2287 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 2288 for preparing the fix. (CVE-2011-4108) 2289 [Robin Seggelmann, Michael Tuexen] 2290 2291 *) Clear bytes used for block padding of SSL 3.0 records. 2292 (CVE-2011-4576) 2293 [Adam Langley (Google)] 2294 2295 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 2296 Kadianakis <desnacked@gmail.com> for discovering this issue and 2297 Adam Langley for preparing the fix. (CVE-2011-4619) 2298 [Adam Langley (Google)] 2299 2300 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027) 2301 [Andrey Kulikov <amdeich@gmail.com>] 2302 2303 *) Prevent malformed RFC3779 data triggering an assertion failure. 2304 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 2305 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 2306 [Rob Austein <sra@hactrn.net>] 2307 2308 *) Improved PRNG seeding for VOS. 2309 [Paul Green <Paul.Green@stratus.com>] 2310 2311 *) Fix ssl_ciph.c set-up race. 2312 [Adam Langley (Google)] 2313 2314 *) Fix spurious failures in ecdsatest.c. 2315 [Emilia K��sper (Google)] 2316 2317 *) Fix the BIO_f_buffer() implementation (which was mixing different 2318 interpretations of the '..._len' fields). 2319 [Adam Langley (Google)] 2320 2321 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 2322 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 2323 threads won't reuse the same blinding coefficients. 2324 2325 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 2326 lock to call BN_BLINDING_invert_ex, and avoids one use of 2327 BN_BLINDING_update for each BN_BLINDING structure (previously, 2328 the last update always remained unused). 2329 [Emilia K��sper (Google)] 2330 2331 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf. 2332 [Bob Buckholz (Google)] 2333 2334 Changes between 1.0.0d and 1.0.0e [6 Sep 2011] 2335 2336 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted 2337 by initialising X509_STORE_CTX properly. (CVE-2011-3207) 2338 [Kaspar Brand <ossl@velox.ch>] 2339 2340 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 2341 for multi-threaded use of ECDH. (CVE-2011-3210) 2342 [Adam Langley (Google)] 2343 2344 *) Fix x509_name_ex_d2i memory leak on bad inputs. 2345 [Bodo Moeller] 2346 2347 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check 2348 signature public key algorithm by using OID xref utilities instead. 2349 Before this you could only use some ECC ciphersuites with SHA1 only. 2350 [Steve Henson] 2351 2352 *) Add protection against ECDSA timing attacks as mentioned in the paper 2353 by Billy Bob Brumley and Nicola Tuveri, see: 2354 2355 http://eprint.iacr.org/2011/232.pdf 2356 2357 [Billy Bob Brumley and Nicola Tuveri] 2358 2359 Changes between 1.0.0c and 1.0.0d [8 Feb 2011] 2360 2361 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 2362 [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 2363 2364 *) Fix bug in string printing code: if *any* escaping is enabled we must 2365 escape the escape character (backslash) or the resulting string is 2366 ambiguous. 2367 [Steve Henson] 2368 2369 Changes between 1.0.0b and 1.0.0c [2 Dec 2010] 2370 2371 *) Disable code workaround for ancient and obsolete Netscape browsers 2372 and servers: an attacker can use it in a ciphersuite downgrade attack. 2373 Thanks to Martin Rex for discovering this bug. CVE-2010-4180 2374 [Steve Henson] 2375 2376 *) Fixed J-PAKE implementation error, originally discovered by 2377 Sebastien Martini, further info and confirmation from Stefan 2378 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 2379 [Ben Laurie] 2380 2381 Changes between 1.0.0a and 1.0.0b [16 Nov 2010] 2382 2383 *) Fix extension code to avoid race conditions which can result in a buffer 2384 overrun vulnerability: resumed sessions must not be modified as they can 2385 be shared by multiple threads. CVE-2010-3864 2386 [Steve Henson] 2387 2388 *) Fix WIN32 build system to correctly link an ENGINE directory into 2389 a DLL. 2390 [Steve Henson] 2391 2392 Changes between 1.0.0 and 1.0.0a [01 Jun 2010] 2393 2394 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover 2395 (CVE-2010-1633) 2396 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>] 2397 2398 Changes between 0.9.8n and 1.0.0 [29 Mar 2010] 2399 2400 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher 2401 context. The operation can be customised via the ctrl mechanism in 2402 case ENGINEs want to include additional functionality. 2403 [Steve Henson] 2404 2405 *) Tolerate yet another broken PKCS#8 key format: private key value negative. 2406 [Steve Henson] 2407 2408 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to 2409 output hashes compatible with older versions of OpenSSL. 2410 [Willy Weisz <weisz@vcpc.univie.ac.at>] 2411 2412 *) Fix compression algorithm handling: if resuming a session use the 2413 compression algorithm of the resumed session instead of determining 2414 it from client hello again. Don't allow server to change algorithm. 2415 [Steve Henson] 2416 2417 *) Add load_crls() function to apps tidying load_certs() too. Add option 2418 to verify utility to allow additional CRLs to be included. 2419 [Steve Henson] 2420 2421 *) Update OCSP request code to permit adding custom headers to the request: 2422 some responders need this. 2423 [Steve Henson] 2424 2425 *) The function EVP_PKEY_sign() returns <=0 on error: check return code 2426 correctly. 2427 [Julia Lawall <julia@diku.dk>] 2428 2429 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it 2430 needlessly dereferenced structures, used obsolete functions and 2431 didn't handle all updated verify codes correctly. 2432 [Steve Henson] 2433 2434 *) Disable MD2 in the default configuration. 2435 [Steve Henson] 2436 2437 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to 2438 indicate the initial BIO being pushed or popped. This makes it possible 2439 to determine whether the BIO is the one explicitly called or as a result 2440 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so 2441 it handles reference counts correctly and doesn't zero out the I/O bio 2442 when it is not being explicitly popped. WARNING: applications which 2443 included workarounds for the old buggy behaviour will need to be modified 2444 or they could free up already freed BIOs. 2445 [Steve Henson] 2446 2447 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni 2448 renaming to all platforms (within the 0.9.8 branch, this was 2449 done conditionally on Netware platforms to avoid a name clash). 2450 [Guenter <lists@gknw.net>] 2451 2452 *) Add ECDHE and PSK support to DTLS. 2453 [Michael Tuexen <tuexen@fh-muenster.de>] 2454 2455 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't 2456 be used on C++. 2457 [Steve Henson] 2458 2459 *) Add "missing" function EVP_MD_flags() (without this the only way to 2460 retrieve a digest flags is by accessing the structure directly. Update 2461 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest 2462 or cipher is registered as in the "from" argument. Print out all 2463 registered digests in the dgst usage message instead of manually 2464 attempting to work them out. 2465 [Steve Henson] 2466 2467 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello: 2468 this allows the use of compression and extensions. Change default cipher 2469 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2 2470 by default unless an application cipher string requests it. 2471 [Steve Henson] 2472 2473 *) Alter match criteria in PKCS12_parse(). It used to try to use local 2474 key ids to find matching certificates and keys but some PKCS#12 files 2475 don't follow the (somewhat unwritten) rules and this strategy fails. 2476 Now just gather all certificates together and the first private key 2477 then look for the first certificate that matches the key. 2478 [Steve Henson] 2479 2480 *) Support use of registered digest and cipher names for dgst and cipher 2481 commands instead of having to add each one as a special case. So now 2482 you can do: 2483 2484 openssl sha256 foo 2485 2486 as well as: 2487 2488 openssl dgst -sha256 foo 2489 2490 and this works for ENGINE based algorithms too. 2491 2492 [Steve Henson] 2493 2494 *) Update Gost ENGINE to support parameter files. 2495 [Victor B. Wagner <vitus@cryptocom.ru>] 2496 2497 *) Support GeneralizedTime in ca utility. 2498 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson] 2499 2500 *) Enhance the hash format used for certificate directory links. The new 2501 form uses the canonical encoding (meaning equivalent names will work 2502 even if they aren't identical) and uses SHA1 instead of MD5. This form 2503 is incompatible with the older format and as a result c_rehash should 2504 be used to rebuild symbolic links. 2505 [Steve Henson] 2506 2507 *) Make PKCS#8 the default write format for private keys, replacing the 2508 traditional format. This form is standardised, more secure and doesn't 2509 include an implicit MD5 dependency. 2510 [Steve Henson] 2511 2512 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code 2513 committed to OpenSSL should pass this lot as a minimum. 2514 [Steve Henson] 2515 2516 *) Add session ticket override functionality for use by EAP-FAST. 2517 [Jouni Malinen <j@w1.fi>] 2518 2519 *) Modify HMAC functions to return a value. Since these can be implemented 2520 in an ENGINE errors can occur. 2521 [Steve Henson] 2522 2523 *) Type-checked OBJ_bsearch_ex. 2524 [Ben Laurie] 2525 2526 *) Type-checked OBJ_bsearch. Also some constification necessitated 2527 by type-checking. Still to come: TXT_DB, bsearch(?), 2528 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING, 2529 CONF_VALUE. 2530 [Ben Laurie] 2531 2532 *) New function OPENSSL_gmtime_adj() to add a specific number of days and 2533 seconds to a tm structure directly, instead of going through OS 2534 specific date routines. This avoids any issues with OS routines such 2535 as the year 2038 bug. New *_adj() functions for ASN1 time structures 2536 and X509_time_adj_ex() to cover the extended range. The existing 2537 X509_time_adj() is still usable and will no longer have any date issues. 2538 [Steve Henson] 2539 2540 *) Delta CRL support. New use deltas option which will attempt to locate 2541 and search any appropriate delta CRLs available. 2542 2543 This work was sponsored by Google. 2544 [Steve Henson] 2545 2546 *) Support for CRLs partitioned by reason code. Reorganise CRL processing 2547 code and add additional score elements. Validate alternate CRL paths 2548 as part of the CRL checking and indicate a new error "CRL path validation 2549 error" in this case. Applications wanting additional details can use 2550 the verify callback and check the new "parent" field. If this is not 2551 NULL CRL path validation is taking place. Existing applications wont 2552 see this because it requires extended CRL support which is off by 2553 default. 2554 2555 This work was sponsored by Google. 2556 [Steve Henson] 2557 2558 *) Support for freshest CRL extension. 2559 2560 This work was sponsored by Google. 2561 [Steve Henson] 2562 2563 *) Initial indirect CRL support. Currently only supported in the CRLs 2564 passed directly and not via lookup. Process certificate issuer 2565 CRL entry extension and lookup CRL entries by bother issuer name 2566 and serial number. Check and process CRL issuer entry in IDP extension. 2567 2568 This work was sponsored by Google. 2569 [Steve Henson] 2570 2571 *) Add support for distinct certificate and CRL paths. The CRL issuer 2572 certificate is validated separately in this case. Only enabled if 2573 an extended CRL support flag is set: this flag will enable additional 2574 CRL functionality in future. 2575 2576 This work was sponsored by Google. 2577 [Steve Henson] 2578 2579 *) Add support for policy mappings extension. 2580 2581 This work was sponsored by Google. 2582 [Steve Henson] 2583 2584 *) Fixes to pathlength constraint, self issued certificate handling, 2585 policy processing to align with RFC3280 and PKITS tests. 2586 2587 This work was sponsored by Google. 2588 [Steve Henson] 2589 2590 *) Support for name constraints certificate extension. DN, email, DNS 2591 and URI types are currently supported. 2592 2593 This work was sponsored by Google. 2594 [Steve Henson] 2595 2596 *) To cater for systems that provide a pointer-based thread ID rather 2597 than numeric, deprecate the current numeric thread ID mechanism and 2598 replace it with a structure and associated callback type. This 2599 mechanism allows a numeric "hash" to be extracted from a thread ID in 2600 either case, and on platforms where pointers are larger than 'long', 2601 mixing is done to help ensure the numeric 'hash' is usable even if it 2602 can't be guaranteed unique. The default mechanism is to use "&errno" 2603 as a pointer-based thread ID to distinguish between threads. 2604 2605 Applications that want to provide their own thread IDs should now use 2606 CRYPTO_THREADID_set_callback() to register a callback that will call 2607 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer(). 2608 2609 Note that ERR_remove_state() is now deprecated, because it is tied 2610 to the assumption that thread IDs are numeric. ERR_remove_state(0) 2611 to free the current thread's error state should be replaced by 2612 ERR_remove_thread_state(NULL). 2613 2614 (This new approach replaces the functions CRYPTO_set_idptr_callback(), 2615 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in 2616 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an 2617 application was previously providing a numeric thread callback that 2618 was inappropriate for distinguishing threads, then uniqueness might 2619 have been obtained with &errno that happened immediately in the 2620 intermediate development versions of OpenSSL; this is no longer the 2621 case, the numeric thread callback will now override the automatic use 2622 of &errno.) 2623 [Geoff Thorpe, with help from Bodo Moeller] 2624 2625 *) Initial support for different CRL issuing certificates. This covers a 2626 simple case where the self issued certificates in the chain exist and 2627 the real CRL issuer is higher in the existing chain. 2628 2629 This work was sponsored by Google. 2630 [Steve Henson] 2631 2632 *) Removed effectively defunct crypto/store from the build. 2633 [Ben Laurie] 2634 2635 *) Revamp of STACK to provide stronger type-checking. Still to come: 2636 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE, 2637 ASN1_STRING, CONF_VALUE. 2638 [Ben Laurie] 2639 2640 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer 2641 RAM on SSL connections. This option can save about 34k per idle SSL. 2642 [Nick Mathewson] 2643 2644 *) Revamp of LHASH to provide stronger type-checking. Still to come: 2645 STACK, TXT_DB, bsearch, qsort. 2646 [Ben Laurie] 2647 2648 *) Initial support for Cryptographic Message Syntax (aka CMS) based 2649 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility, 2650 support for data, signedData, compressedData, digestedData and 2651 encryptedData, envelopedData types included. Scripts to check against 2652 RFC4134 examples draft and interop and consistency checks of many 2653 content types and variants. 2654 [Steve Henson] 2655 2656 *) Add options to enc utility to support use of zlib compression BIO. 2657 [Steve Henson] 2658 2659 *) Extend mk1mf to support importing of options and assembly language 2660 files from Configure script, currently only included in VC-WIN32. 2661 The assembly language rules can now optionally generate the source 2662 files from the associated perl scripts. 2663 [Steve Henson] 2664 2665 *) Implement remaining functionality needed to support GOST ciphersuites. 2666 Interop testing has been performed using CryptoPro implementations. 2667 [Victor B. Wagner <vitus@cryptocom.ru>] 2668 2669 *) s390x assembler pack. 2670 [Andy Polyakov] 2671 2672 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU 2673 "family." 2674 [Andy Polyakov] 2675 2676 *) Implement Opaque PRF Input TLS extension as specified in 2677 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an 2678 official specification yet and no extension type assignment by 2679 IANA exists, this extension (for now) will have to be explicitly 2680 enabled when building OpenSSL by providing the extension number 2681 to use. For example, specify an option 2682 2683 -DTLSEXT_TYPE_opaque_prf_input=0x9527 2684 2685 to the "config" or "Configure" script to enable the extension, 2686 assuming extension number 0x9527 (which is a completely arbitrary 2687 and unofficial assignment based on the MD5 hash of the Internet 2688 Draft). Note that by doing so, you potentially lose 2689 interoperability with other TLS implementations since these might 2690 be using the same extension number for other purposes. 2691 2692 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the 2693 opaque PRF input value to use in the handshake. This will create 2694 an interal copy of the length-'len' string at 'src', and will 2695 return non-zero for success. 2696 2697 To get more control and flexibility, provide a callback function 2698 by using 2699 2700 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb) 2701 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg) 2702 2703 where 2704 2705 int (*cb)(SSL *, void *peerinput, size_t len, void *arg); 2706 void *arg; 2707 2708 Callback function 'cb' will be called in handshakes, and is 2709 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate. 2710 Argument 'arg' is for application purposes (the value as given to 2711 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly 2712 be provided to the callback function). The callback function 2713 has to return non-zero to report success: usually 1 to use opaque 2714 PRF input just if possible, or 2 to enforce use of the opaque PRF 2715 input. In the latter case, the library will abort the handshake 2716 if opaque PRF input is not successfully negotiated. 2717 2718 Arguments 'peerinput' and 'len' given to the callback function 2719 will always be NULL and 0 in the case of a client. A server will 2720 see the client's opaque PRF input through these variables if 2721 available (NULL and 0 otherwise). Note that if the server 2722 provides an opaque PRF input, the length must be the same as the 2723 length of the client's opaque PRF input. 2724 2725 Note that the callback function will only be called when creating 2726 a new session (session resumption can resume whatever was 2727 previously negotiated), and will not be called in SSL 2.0 2728 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or 2729 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended 2730 for applications that need to enforce opaque PRF input. 2731 2732 [Bodo Moeller] 2733 2734 *) Update ssl code to support digests other than SHA1+MD5 for handshake 2735 MAC. 2736 2737 [Victor B. Wagner <vitus@cryptocom.ru>] 2738 2739 *) Add RFC4507 support to OpenSSL. This includes the corrections in 2740 RFC4507bis. The encrypted ticket format is an encrypted encoded 2741 SSL_SESSION structure, that way new session features are automatically 2742 supported. 2743 2744 If a client application caches session in an SSL_SESSION structure 2745 support is transparent because tickets are now stored in the encoded 2746 SSL_SESSION. 2747 2748 The SSL_CTX structure automatically generates keys for ticket 2749 protection in servers so again support should be possible 2750 with no application modification. 2751 2752 If a client or server wishes to disable RFC4507 support then the option 2753 SSL_OP_NO_TICKET can be set. 2754 2755 Add a TLS extension debugging callback to allow the contents of any client 2756 or server extensions to be examined. 2757 2758 This work was sponsored by Google. 2759 [Steve Henson] 2760 2761 *) Final changes to avoid use of pointer pointer casts in OpenSSL. 2762 OpenSSL should now compile cleanly on gcc 4.2 2763 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson] 2764 2765 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC 2766 support including streaming MAC support: this is required for GOST 2767 ciphersuite support. 2768 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson] 2769 2770 *) Add option -stream to use PKCS#7 streaming in smime utility. New 2771 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream() 2772 to output in BER and PEM format. 2773 [Steve Henson] 2774 2775 *) Experimental support for use of HMAC via EVP_PKEY interface. This 2776 allows HMAC to be handled via the EVP_DigestSign*() interface. The 2777 EVP_PKEY "key" in this case is the HMAC key, potentially allowing 2778 ENGINE support for HMAC keys which are unextractable. New -mac and 2779 -macopt options to dgst utility. 2780 [Steve Henson] 2781 2782 *) New option -sigopt to dgst utility. Update dgst to use 2783 EVP_Digest{Sign,Verify}*. These two changes make it possible to use 2784 alternative signing paramaters such as X9.31 or PSS in the dgst 2785 utility. 2786 [Steve Henson] 2787 2788 *) Change ssl_cipher_apply_rule(), the internal function that does 2789 the work each time a ciphersuite string requests enabling 2790 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or 2791 removing ("!foo+bar") a class of ciphersuites: Now it maintains 2792 the order of disabled ciphersuites such that those ciphersuites 2793 that most recently went from enabled to disabled not only stay 2794 in order with respect to each other, but also have higher priority 2795 than other disabled ciphersuites the next time ciphersuites are 2796 enabled again. 2797 2798 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable 2799 the same ciphersuites as with "HIGH" alone, but in a specific 2800 order where the PSK ciphersuites come first (since they are the 2801 most recently disabled ciphersuites when "HIGH" is parsed). 2802 2803 Also, change ssl_create_cipher_list() (using this new 2804 funcionality) such that between otherwise identical 2805 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in 2806 the default order. 2807 [Bodo Moeller] 2808 2809 *) Change ssl_create_cipher_list() so that it automatically 2810 arranges the ciphersuites in reasonable order before starting 2811 to process the rule string. Thus, the definition for "DEFAULT" 2812 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but 2813 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH". 2814 This makes it much easier to arrive at a reasonable default order 2815 in applications for which anonymous ciphers are OK (meaning 2816 that you can't actually use DEFAULT). 2817 [Bodo Moeller; suggested by Victor Duchovni] 2818 2819 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string 2820 processing) into multiple integers instead of setting 2821 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK", 2822 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer. 2823 (These masks as well as the individual bit definitions are hidden 2824 away into the non-exported interface ssl/ssl_locl.h, so this 2825 change to the definition of the SSL_CIPHER structure shouldn't 2826 affect applications.) This give us more bits for each of these 2827 categories, so there is no longer a need to coagulate AES128 and 2828 AES256 into a single algorithm bit, and to coagulate Camellia128 2829 and Camellia256 into a single algorithm bit, which has led to all 2830 kinds of kludges. 2831 2832 Thus, among other things, the kludge introduced in 0.9.7m and 2833 0.9.8e for masking out AES256 independently of AES128 or masking 2834 out Camellia256 independently of AES256 is not needed here in 0.9.9. 2835 2836 With the change, we also introduce new ciphersuite aliases that 2837 so far were missing: "AES128", "AES256", "CAMELLIA128", and 2838 "CAMELLIA256". 2839 [Bodo Moeller] 2840 2841 *) Add support for dsa-with-SHA224 and dsa-with-SHA256. 2842 Use the leftmost N bytes of the signature input if the input is 2843 larger than the prime q (with N being the size in bytes of q). 2844 [Nils Larsch] 2845 2846 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses 2847 it yet and it is largely untested. 2848 [Steve Henson] 2849 2850 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types. 2851 [Nils Larsch] 2852 2853 *) Initial incomplete changes to avoid need for function casts in OpenSSL 2854 some compilers (gcc 4.2 and later) reject their use. Safestack is 2855 reimplemented. Update ASN1 to avoid use of legacy functions. 2856 [Steve Henson] 2857 2858 *) Win32/64 targets are linked with Winsock2. 2859 [Andy Polyakov] 2860 2861 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected 2862 to external functions. This can be used to increase CRL handling 2863 efficiency especially when CRLs are very large by (for example) storing 2864 the CRL revoked certificates in a database. 2865 [Steve Henson] 2866 2867 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so 2868 new CRLs added to a directory can be used. New command line option 2869 -verify_return_error to s_client and s_server. This causes real errors 2870 to be returned by the verify callback instead of carrying on no matter 2871 what. This reflects the way a "real world" verify callback would behave. 2872 [Steve Henson] 2873 2874 *) GOST engine, supporting several GOST algorithms and public key formats. 2875 Kindly donated by Cryptocom. 2876 [Cryptocom] 2877 2878 *) Partial support for Issuing Distribution Point CRL extension. CRLs 2879 partitioned by DP are handled but no indirect CRL or reason partitioning 2880 (yet). Complete overhaul of CRL handling: now the most suitable CRL is 2881 selected via a scoring technique which handles IDP and AKID in CRLs. 2882 [Steve Henson] 2883 2884 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which 2885 will ultimately be used for all verify operations: this will remove the 2886 X509_STORE dependency on certificate verification and allow alternative 2887 lookup methods. X509_STORE based implementations of these two callbacks. 2888 [Steve Henson] 2889 2890 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names. 2891 Modify get_crl() to find a valid (unexpired) CRL if possible. 2892 [Steve Henson] 2893 2894 *) New function X509_CRL_match() to check if two CRLs are identical. Normally 2895 this would be called X509_CRL_cmp() but that name is already used by 2896 a function that just compares CRL issuer names. Cache several CRL 2897 extensions in X509_CRL structure and cache CRLDP in X509. 2898 [Steve Henson] 2899 2900 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name) 2901 this maps equivalent X509_NAME structures into a consistent structure. 2902 Name comparison can then be performed rapidly using memcmp(). 2903 [Steve Henson] 2904 2905 *) Non-blocking OCSP request processing. Add -timeout option to ocsp 2906 utility. 2907 [Steve Henson] 2908 2909 *) Allow digests to supply their own micalg string for S/MIME type using 2910 the ctrl EVP_MD_CTRL_MICALG. 2911 [Steve Henson] 2912 2913 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the 2914 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN 2915 ctrl. It can then customise the structure before and/or after signing 2916 if necessary. 2917 [Steve Henson] 2918 2919 *) New function OBJ_add_sigid() to allow application defined signature OIDs 2920 to be added to OpenSSLs internal tables. New function OBJ_sigid_free() 2921 to free up any added signature OIDs. 2922 [Steve Henson] 2923 2924 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(), 2925 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal 2926 digest and cipher tables. New options added to openssl utility: 2927 list-message-digest-algorithms and list-cipher-algorithms. 2928 [Steve Henson] 2929 2930 *) Change the array representation of binary polynomials: the list 2931 of degrees of non-zero coefficients is now terminated with -1. 2932 Previously it was terminated with 0, which was also part of the 2933 value; thus, the array representation was not applicable to 2934 polynomials where t^0 has coefficient zero. This change makes 2935 the array representation useful in a more general context. 2936 [Douglas Stebila] 2937 2938 *) Various modifications and fixes to SSL/TLS cipher string 2939 handling. For ECC, the code now distinguishes between fixed ECDH 2940 with RSA certificates on the one hand and with ECDSA certificates 2941 on the other hand, since these are separate ciphersuites. The 2942 unused code for Fortezza ciphersuites has been removed. 2943 2944 For consistency with EDH, ephemeral ECDH is now called "EECDH" 2945 (not "ECDHE"). For consistency with the code for DH 2946 certificates, use of ECDH certificates is now considered ECDH 2947 authentication, not RSA or ECDSA authentication (the latter is 2948 merely the CA's signing algorithm and not actively used in the 2949 protocol). 2950 2951 The temporary ciphersuite alias "ECCdraft" is no longer 2952 available, and ECC ciphersuites are no longer excluded from "ALL" 2953 and "DEFAULT". The following aliases now exist for RFC 4492 2954 ciphersuites, most of these by analogy with the DH case: 2955 2956 kECDHr - ECDH cert, signed with RSA 2957 kECDHe - ECDH cert, signed with ECDSA 2958 kECDH - ECDH cert (signed with either RSA or ECDSA) 2959 kEECDH - ephemeral ECDH 2960 ECDH - ECDH cert or ephemeral ECDH 2961 2962 aECDH - ECDH cert 2963 aECDSA - ECDSA cert 2964 ECDSA - ECDSA cert 2965 2966 AECDH - anonymous ECDH 2967 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH") 2968 2969 [Bodo Moeller] 2970 2971 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported. 2972 Use correct micalg parameters depending on digest(s) in signed message. 2973 [Steve Henson] 2974 2975 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process 2976 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code. 2977 [Steve Henson] 2978 2979 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit 2980 an engine to register a method. Add ENGINE lookups for methods and 2981 functional reference processing. 2982 [Steve Henson] 2983 2984 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of 2985 EVP_{Sign,Verify}* which allow an application to customise the signature 2986 process. 2987 [Steve Henson] 2988 2989 *) New -resign option to smime utility. This adds one or more signers 2990 to an existing PKCS#7 signedData structure. Also -md option to use an 2991 alternative message digest algorithm for signing. 2992 [Steve Henson] 2993 2994 *) Tidy up PKCS#7 routines and add new functions to make it easier to 2995 create PKCS7 structures containing multiple signers. Update smime 2996 application to support multiple signers. 2997 [Steve Henson] 2998 2999 *) New -macalg option to pkcs12 utility to allow setting of an alternative 3000 digest MAC. 3001 [Steve Henson] 3002 3003 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC. 3004 Reorganize PBE internals to lookup from a static table using NIDs, 3005 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl: 3006 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative 3007 PRF which will be automatically used with PBES2. 3008 [Steve Henson] 3009 3010 *) Replace the algorithm specific calls to generate keys in "req" with the 3011 new API. 3012 [Steve Henson] 3013 3014 *) Update PKCS#7 enveloped data routines to use new API. This is now 3015 supported by any public key method supporting the encrypt operation. A 3016 ctrl is added to allow the public key algorithm to examine or modify 3017 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is 3018 a no op. 3019 [Steve Henson] 3020 3021 *) Add a ctrl to asn1 method to allow a public key algorithm to express 3022 a default digest type to use. In most cases this will be SHA1 but some 3023 algorithms (such as GOST) need to specify an alternative digest. The 3024 return value indicates how strong the prefernce is 1 means optional and 3025 2 is mandatory (that is it is the only supported type). Modify 3026 ASN1_item_sign() to accept a NULL digest argument to indicate it should 3027 use the default md. Update openssl utilities to use the default digest 3028 type for signing if it is not explicitly indicated. 3029 [Steve Henson] 3030 3031 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New 3032 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant 3033 signing method from the key type. This effectively removes the link 3034 between digests and public key types. 3035 [Steve Henson] 3036 3037 *) Add an OID cross reference table and utility functions. Its purpose is to 3038 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1, 3039 rsaEncryption. This will allow some of the algorithm specific hackery 3040 needed to use the correct OID to be removed. 3041 [Steve Henson] 3042 3043 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO 3044 structures for PKCS7_sign(). They are now set up by the relevant public 3045 key ASN1 method. 3046 [Steve Henson] 3047 3048 *) Add provisional EC pkey method with support for ECDSA and ECDH. 3049 [Steve Henson] 3050 3051 *) Add support for key derivation (agreement) in the API, DH method and 3052 pkeyutl. 3053 [Steve Henson] 3054 3055 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support 3056 public and private key formats. As a side effect these add additional 3057 command line functionality not previously available: DSA signatures can be 3058 generated and verified using pkeyutl and DH key support and generation in 3059 pkey, genpkey. 3060 [Steve Henson] 3061 3062 *) BeOS support. 3063 [Oliver Tappe <zooey@hirschkaefer.de>] 3064 3065 *) New make target "install_html_docs" installs HTML renditions of the 3066 manual pages. 3067 [Oliver Tappe <zooey@hirschkaefer.de>] 3068 3069 *) New utility "genpkey" this is analagous to "genrsa" etc except it can 3070 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to 3071 support key and parameter generation and add initial key generation 3072 functionality for RSA. 3073 [Steve Henson] 3074 3075 *) Add functions for main EVP_PKEY_method operations. The undocumented 3076 functions EVP_PKEY_{encrypt,decrypt} have been renamed to 3077 EVP_PKEY_{encrypt,decrypt}_old. 3078 [Steve Henson] 3079 3080 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public 3081 key API, doesn't do much yet. 3082 [Steve Henson] 3083 3084 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about 3085 public key algorithms. New option to openssl utility: 3086 "list-public-key-algorithms" to print out info. 3087 [Steve Henson] 3088 3089 *) Implement the Supported Elliptic Curves Extension for 3090 ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 3091 [Douglas Stebila] 3092 3093 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or 3094 EVP_CIPHER structures to avoid later problems in EVP_cleanup(). 3095 [Steve Henson] 3096 3097 *) New utilities pkey and pkeyparam. These are similar to algorithm specific 3098 utilities such as rsa, dsa, dsaparam etc except they process any key 3099 type. 3100 [Steve Henson] 3101 3102 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New 3103 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(), 3104 EVP_PKEY_print_param() to print public key data from an EVP_PKEY 3105 structure. 3106 [Steve Henson] 3107 3108 *) Initial support for pluggable public key ASN1. 3109 De-spaghettify the public key ASN1 handling. Move public and private 3110 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate 3111 algorithm specific handling to a single module within the relevant 3112 algorithm directory. Add functions to allow (near) opaque processing 3113 of public and private key structures. 3114 [Steve Henson] 3115 3116 *) Implement the Supported Point Formats Extension for 3117 ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 3118 [Douglas Stebila] 3119 3120 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members 3121 for the psk identity [hint] and the psk callback functions to the 3122 SSL_SESSION, SSL and SSL_CTX structure. 3123 3124 New ciphersuites: 3125 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA, 3126 PSK-AES256-CBC-SHA 3127 3128 New functions: 3129 SSL_CTX_use_psk_identity_hint 3130 SSL_get_psk_identity_hint 3131 SSL_get_psk_identity 3132 SSL_use_psk_identity_hint 3133 3134 [Mika Kousa and Pasi Eronen of Nokia Corporation] 3135 3136 *) Add RFC 3161 compliant time stamp request creation, response generation 3137 and response verification functionality. 3138 [Zolt��n Gl��zik <zglozik@opentsa.org>, The OpenTSA Project] 3139 3140 *) Add initial support for TLS extensions, specifically for the server_name 3141 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 3142 have new members for a host name. The SSL data structure has an 3143 additional member SSL_CTX *initial_ctx so that new sessions can be 3144 stored in that context to allow for session resumption, even after the 3145 SSL has been switched to a new SSL_CTX in reaction to a client's 3146 server_name extension. 3147 3148 New functions (subject to change): 3149 3150 SSL_get_servername() 3151 SSL_get_servername_type() 3152 SSL_set_SSL_CTX() 3153 3154 New CTRL codes and macros (subject to change): 3155 3156 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 3157 - SSL_CTX_set_tlsext_servername_callback() 3158 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 3159 - SSL_CTX_set_tlsext_servername_arg() 3160 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 3161 3162 openssl s_client has a new '-servername ...' option. 3163 3164 openssl s_server has new options '-servername_host ...', '-cert2 ...', 3165 '-key2 ...', '-servername_fatal' (subject to change). This allows 3166 testing the HostName extension for a specific single host name ('-cert' 3167 and '-key' remain fallbacks for handshakes without HostName 3168 negotiation). If the unrecogninzed_name alert has to be sent, this by 3169 default is a warning; it becomes fatal with the '-servername_fatal' 3170 option. 3171 3172 [Peter Sylvester, Remy Allais, Christophe Renou] 3173 3174 *) Whirlpool hash implementation is added. 3175 [Andy Polyakov] 3176 3177 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to 3178 bn(64,32). Because of instruction set limitations it doesn't have 3179 any negative impact on performance. This was done mostly in order 3180 to make it possible to share assembler modules, such as bn_mul_mont 3181 implementations, between 32- and 64-bit builds without hassle. 3182 [Andy Polyakov] 3183 3184 *) Move code previously exiled into file crypto/ec/ec2_smpt.c 3185 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP 3186 macro. 3187 [Bodo Moeller] 3188 3189 *) New candidate for BIGNUM assembler implementation, bn_mul_mont, 3190 dedicated Montgomery multiplication procedure, is introduced. 3191 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher 3192 "64-bit" performance on certain 32-bit targets. 3193 [Andy Polyakov] 3194 3195 *) New option SSL_OP_NO_COMP to disable use of compression selectively 3196 in SSL structures. New SSL ctrl to set maximum send fragment size. 3197 Save memory by seeting the I/O buffer sizes dynamically instead of 3198 using the maximum available value. 3199 [Steve Henson] 3200 3201 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code 3202 in addition to the text details. 3203 [Bodo Moeller] 3204 3205 *) Very, very preliminary EXPERIMENTAL support for printing of general 3206 ASN1 structures. This currently produces rather ugly output and doesn't 3207 handle several customised structures at all. 3208 [Steve Henson] 3209 3210 *) Integrated support for PVK file format and some related formats such 3211 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support 3212 these in the 'rsa' and 'dsa' utilities. 3213 [Steve Henson] 3214 3215 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line. 3216 [Steve Henson] 3217 3218 *) Remove the ancient ASN1_METHOD code. This was only ever used in one 3219 place for the (very old) "NETSCAPE" format certificates which are now 3220 handled using new ASN1 code equivalents. 3221 [Steve Henson] 3222 3223 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD 3224 pointer and make the SSL_METHOD parameter in SSL_CTX_new, 3225 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'. 3226 [Nils Larsch] 3227 3228 *) Modify CRL distribution points extension code to print out previously 3229 unsupported fields. Enhance extension setting code to allow setting of 3230 all fields. 3231 [Steve Henson] 3232 3233 *) Add print and set support for Issuing Distribution Point CRL extension. 3234 [Steve Henson] 3235 3236 *) Change 'Configure' script to enable Camellia by default. 3237 [NTT] 3238 3239 Changes between 0.9.8m and 0.9.8n [24 Mar 2010] 3240 3241 *) When rejecting SSL/TLS records due to an incorrect version number, never 3242 update s->server with a new major version number. As of 3243 - OpenSSL 0.9.8m if 'short' is a 16-bit type, 3244 - OpenSSL 0.9.8f if 'short' is longer than 16 bits, 3245 the previous behavior could result in a read attempt at NULL when 3246 receiving specific incorrect SSL/TLS records once record payload 3247 protection is active. (CVE-2010-0740) 3248 [Bodo Moeller, Adam Langley <agl@chromium.org>] 3249 3250 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL 3251 could be crashed if the relevant tables were not present (e.g. chrooted). 3252 [Tomas Hoger <thoger@redhat.com>] 3253 3254 Changes between 0.9.8l and 0.9.8m [25 Feb 2010] 3255 3256 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245) 3257 [Martin Olsson, Neel Mehta] 3258 3259 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to 3260 accommodate for stack sorting, always a write lock!). 3261 [Bodo Moeller] 3262 3263 *) On some versions of WIN32 Heap32Next is very slow. This can cause 3264 excessive delays in the RAND_poll(): over a minute. As a workaround 3265 include a time check in the inner Heap32Next loop too. 3266 [Steve Henson] 3267 3268 *) The code that handled flushing of data in SSL/TLS originally used the 3269 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused 3270 the problem outlined in PR#1949. The fix suggested there however can 3271 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions 3272 of Apache). So instead simplify the code to flush unconditionally. 3273 This should be fine since flushing with no data to flush is a no op. 3274 [Steve Henson] 3275 3276 *) Handle TLS versions 2.0 and later properly and correctly use the 3277 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way 3278 off ancient servers have a habit of sticking around for a while... 3279 [Steve Henson] 3280 3281 *) Modify compression code so it frees up structures without using the 3282 ex_data callbacks. This works around a problem where some applications 3283 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when 3284 restarting) then use compression (e.g. SSL with compression) later. 3285 This results in significant per-connection memory leaks and 3286 has caused some security issues including CVE-2008-1678 and 3287 CVE-2009-4355. 3288 [Steve Henson] 3289 3290 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't 3291 change when encrypting or decrypting. 3292 [Bodo Moeller] 3293 3294 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to 3295 connect and renegotiate with servers which do not support RI. 3296 Until RI is more widely deployed this option is enabled by default. 3297 [Steve Henson] 3298 3299 *) Add "missing" ssl ctrls to clear options and mode. 3300 [Steve Henson] 3301 3302 *) If client attempts to renegotiate and doesn't support RI respond with 3303 a no_renegotiation alert as required by RFC5746. Some renegotiating 3304 TLS clients will continue a connection gracefully when they receive 3305 the alert. Unfortunately OpenSSL mishandled this alert and would hang 3306 waiting for a server hello which it will never receive. Now we treat a 3307 received no_renegotiation alert as a fatal error. This is because 3308 applications requesting a renegotiation might well expect it to succeed 3309 and would have no code in place to handle the server denying it so the 3310 only safe thing to do is to terminate the connection. 3311 [Steve Henson] 3312 3313 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if 3314 peer supports secure renegotiation and 0 otherwise. Print out peer 3315 renegotiation support in s_client/s_server. 3316 [Steve Henson] 3317 3318 *) Replace the highly broken and deprecated SPKAC certification method with 3319 the updated NID creation version. This should correctly handle UTF8. 3320 [Steve Henson] 3321 3322 *) Implement RFC5746. Re-enable renegotiation but require the extension 3323 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION 3324 turns out to be a bad idea. It has been replaced by 3325 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with 3326 SSL_CTX_set_options(). This is really not recommended unless you 3327 know what you are doing. 3328 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson] 3329 3330 *) Fixes to stateless session resumption handling. Use initial_ctx when 3331 issuing and attempting to decrypt tickets in case it has changed during 3332 servername handling. Use a non-zero length session ID when attempting 3333 stateless session resumption: this makes it possible to determine if 3334 a resumption has occurred immediately after receiving server hello 3335 (several places in OpenSSL subtly assume this) instead of later in 3336 the handshake. 3337 [Steve Henson] 3338 3339 *) The functions ENGINE_ctrl(), OPENSSL_isservice(), 3340 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error 3341 fixes for a few places where the return code is not checked 3342 correctly. 3343 [Julia Lawall <julia@diku.dk>] 3344 3345 *) Add --strict-warnings option to Configure script to include devteam 3346 warnings in other configurations. 3347 [Steve Henson] 3348 3349 *) Add support for --libdir option and LIBDIR variable in makefiles. This 3350 makes it possible to install openssl libraries in locations which 3351 have names other than "lib", for example "/usr/lib64" which some 3352 systems need. 3353 [Steve Henson, based on patch from Jeremy Utley] 3354 3355 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of 3356 X690 8.9.12 and can produce some misleading textual output of OIDs. 3357 [Steve Henson, reported by Dan Kaminsky] 3358 3359 *) Delete MD2 from algorithm tables. This follows the recommendation in 3360 several standards that it is not used in new applications due to 3361 several cryptographic weaknesses. For binary compatibility reasons 3362 the MD2 API is still compiled in by default. 3363 [Steve Henson] 3364 3365 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved 3366 and restored. 3367 [Steve Henson] 3368 3369 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and 3370 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name 3371 clash. 3372 [Guenter <lists@gknw.net>] 3373 3374 *) Fix the server certificate chain building code to use X509_verify_cert(), 3375 it used to have an ad-hoc builder which was unable to cope with anything 3376 other than a simple chain. 3377 [David Woodhouse <dwmw2@infradead.org>, Steve Henson] 3378 3379 *) Don't check self signed certificate signatures in X509_verify_cert() 3380 by default (a flag can override this): it just wastes time without 3381 adding any security. As a useful side effect self signed root CAs 3382 with non-FIPS digests are now usable in FIPS mode. 3383 [Steve Henson] 3384 3385 *) In dtls1_process_out_of_seq_message() the check if the current message 3386 is already buffered was missing. For every new message was memory 3387 allocated, allowing an attacker to perform an denial of service attack 3388 with sending out of seq handshake messages until there is no memory 3389 left. Additionally every future messege was buffered, even if the 3390 sequence number made no sense and would be part of another handshake. 3391 So only messages with sequence numbers less than 10 in advance will be 3392 buffered. (CVE-2009-1378) 3393 [Robin Seggelmann, discovered by Daniel Mentz] 3394 3395 *) Records are buffered if they arrive with a future epoch to be 3396 processed after finishing the corresponding handshake. There is 3397 currently no limitation to this buffer allowing an attacker to perform 3398 a DOS attack with sending records with future epochs until there is no 3399 memory left. This patch adds the pqueue_size() function to detemine 3400 the size of a buffer and limits the record buffer to 100 entries. 3401 (CVE-2009-1377) 3402 [Robin Seggelmann, discovered by Daniel Mentz] 3403 3404 *) Keep a copy of frag->msg_header.frag_len so it can be used after the 3405 parent structure is freed. (CVE-2009-1379) 3406 [Daniel Mentz] 3407 3408 *) Handle non-blocking I/O properly in SSL_shutdown() call. 3409 [Darryl Miles <darryl-mailinglists@netbauds.net>] 3410 3411 *) Add 2.5.4.* OIDs 3412 [Ilya O. <vrghost@gmail.com>] 3413 3414 Changes between 0.9.8k and 0.9.8l [5 Nov 2009] 3415 3416 *) Disable renegotiation completely - this fixes a severe security 3417 problem (CVE-2009-3555) at the cost of breaking all 3418 renegotiation. Renegotiation can be re-enabled by setting 3419 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at 3420 run-time. This is really not recommended unless you know what 3421 you're doing. 3422 [Ben Laurie] 3423 3424 Changes between 0.9.8j and 0.9.8k [25 Mar 2009] 3425 3426 *) Don't set val to NULL when freeing up structures, it is freed up by 3427 underlying code. If sizeof(void *) > sizeof(long) this can result in 3428 zeroing past the valid field. (CVE-2009-0789) 3429 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>] 3430 3431 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not 3432 checked correctly. This would allow some invalid signed attributes to 3433 appear to verify correctly. (CVE-2009-0591) 3434 [Ivan Nestlerode <inestlerode@us.ibm.com>] 3435 3436 *) Reject UniversalString and BMPString types with invalid lengths. This 3437 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have 3438 a legal length. (CVE-2009-0590) 3439 [Steve Henson] 3440 3441 *) Set S/MIME signing as the default purpose rather than setting it 3442 unconditionally. This allows applications to override it at the store 3443 level. 3444 [Steve Henson] 3445 3446 *) Permit restricted recursion of ASN1 strings. This is needed in practice 3447 to handle some structures. 3448 [Steve Henson] 3449 3450 *) Improve efficiency of mem_gets: don't search whole buffer each time 3451 for a '\n' 3452 [Jeremy Shapiro <jnshapir@us.ibm.com>] 3453 3454 *) New -hex option for openssl rand. 3455 [Matthieu Herrb] 3456 3457 *) Print out UTF8String and NumericString when parsing ASN1. 3458 [Steve Henson] 3459 3460 *) Support NumericString type for name components. 3461 [Steve Henson] 3462 3463 *) Allow CC in the environment to override the automatically chosen 3464 compiler. Note that nothing is done to ensure flags work with the 3465 chosen compiler. 3466 [Ben Laurie] 3467 3468 Changes between 0.9.8i and 0.9.8j [07 Jan 2009] 3469 3470 *) Properly check EVP_VerifyFinal() and similar return values 3471 (CVE-2008-5077). 3472 [Ben Laurie, Bodo Moeller, Google Security Team] 3473 3474 *) Enable TLS extensions by default. 3475 [Ben Laurie] 3476 3477 *) Allow the CHIL engine to be loaded, whether the application is 3478 multithreaded or not. (This does not release the developer from the 3479 obligation to set up the dynamic locking callbacks.) 3480 [Sander Temme <sander@temme.net>] 3481 3482 *) Use correct exit code if there is an error in dgst command. 3483 [Steve Henson; problem pointed out by Roland Dirlewanger] 3484 3485 *) Tweak Configure so that you need to say "experimental-jpake" to enable 3486 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. 3487 [Bodo Moeller] 3488 3489 *) Add experimental JPAKE support, including demo authentication in 3490 s_client and s_server. 3491 [Ben Laurie] 3492 3493 *) Set the comparison function in v3_addr_canonize(). 3494 [Rob Austein <sra@hactrn.net>] 3495 3496 *) Add support for XMPP STARTTLS in s_client. 3497 [Philip Paeps <philip@freebsd.org>] 3498 3499 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior 3500 to ensure that even with this option, only ciphersuites in the 3501 server's preference list will be accepted. (Note that the option 3502 applies only when resuming a session, so the earlier behavior was 3503 just about the algorithm choice for symmetric cryptography.) 3504 [Bodo Moeller] 3505 3506 Changes between 0.9.8h and 0.9.8i [15 Sep 2008] 3507 3508 *) Fix NULL pointer dereference if a DTLS server received 3509 ChangeCipherSpec as first record (CVE-2009-1386). 3510 [PR #1679] 3511 3512 *) Fix a state transitition in s3_srvr.c and d1_srvr.c 3513 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...). 3514 [Nagendra Modadugu] 3515 3516 *) The fix in 0.9.8c that supposedly got rid of unsafe 3517 double-checked locking was incomplete for RSA blinding, 3518 addressing just one layer of what turns out to have been 3519 doubly unsafe triple-checked locking. 3520 3521 So now fix this for real by retiring the MONT_HELPER macro 3522 in crypto/rsa/rsa_eay.c. 3523 3524 [Bodo Moeller; problem pointed out by Marius Schilder] 3525 3526 *) Various precautionary measures: 3527 3528 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). 3529 3530 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). 3531 (NB: This would require knowledge of the secret session ticket key 3532 to exploit, in which case you'd be SOL either way.) 3533 3534 - Change bn_nist.c so that it will properly handle input BIGNUMs 3535 outside the expected range. 3536 3537 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG 3538 builds. 3539 3540 [Neel Mehta, Bodo Moeller] 3541 3542 *) Allow engines to be "soft loaded" - i.e. optionally don't die if 3543 the load fails. Useful for distros. 3544 [Ben Laurie and the FreeBSD team] 3545 3546 *) Add support for Local Machine Keyset attribute in PKCS#12 files. 3547 [Steve Henson] 3548 3549 *) Fix BN_GF2m_mod_arr() top-bit cleanup code. 3550 [Huang Ying] 3551 3552 *) Expand ENGINE to support engine supplied SSL client certificate functions. 3553 3554 This work was sponsored by Logica. 3555 [Steve Henson] 3556 3557 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows 3558 keystores. Support for SSL/TLS client authentication too. 3559 Not compiled unless enable-capieng specified to Configure. 3560 3561 This work was sponsored by Logica. 3562 [Steve Henson] 3563 3564 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using 3565 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain 3566 attribute creation routines such as certifcate requests and PKCS#12 3567 files. 3568 [Steve Henson] 3569 3570 Changes between 0.9.8g and 0.9.8h [28 May 2008] 3571 3572 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS 3573 handshake which could lead to a cilent crash as found using the 3574 Codenomicon TLS test suite (CVE-2008-1672) 3575 [Steve Henson, Mark Cox] 3576 3577 *) Fix double free in TLS server name extensions which could lead to 3578 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891) 3579 [Joe Orton] 3580 3581 *) Clear error queue in SSL_CTX_use_certificate_chain_file() 3582 3583 Clear the error queue to ensure that error entries left from 3584 older function calls do not interfere with the correct operation. 3585 [Lutz Jaenicke, Erik de Castro Lopo] 3586 3587 *) Remove root CA certificates of commercial CAs: 3588 3589 The OpenSSL project does not recommend any specific CA and does not 3590 have any policy with respect to including or excluding any CA. 3591 Therefore it does not make any sense to ship an arbitrary selection 3592 of root CA certificates with the OpenSSL software. 3593 [Lutz Jaenicke] 3594 3595 *) RSA OAEP patches to fix two separate invalid memory reads. 3596 The first one involves inputs when 'lzero' is greater than 3597 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes 3598 before the beginning of from). The second one involves inputs where 3599 the 'db' section contains nothing but zeroes (there is a one-byte 3600 invalid read after the end of 'db'). 3601 [Ivan Nestlerode <inestlerode@us.ibm.com>] 3602 3603 *) Partial backport from 0.9.9-dev: 3604 3605 Introduce bn_mul_mont (dedicated Montgomery multiplication 3606 procedure) as a candidate for BIGNUM assembler implementation. 3607 While 0.9.9-dev uses assembler for various architectures, only 3608 x86_64 is available by default here in the 0.9.8 branch, and 3609 32-bit x86 is available through a compile-time setting. 3610 3611 To try the 32-bit x86 assembler implementation, use Configure 3612 option "enable-montasm" (which exists only for this backport). 3613 3614 As "enable-montasm" for 32-bit x86 disclaims code stability 3615 anyway, in this constellation we activate additional code 3616 backported from 0.9.9-dev for further performance improvements, 3617 namely BN_from_montgomery_word. (To enable this otherwise, 3618 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".) 3619 3620 [Andy Polyakov (backport partially by Bodo Moeller)] 3621 3622 *) Add TLS session ticket callback. This allows an application to set 3623 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed 3624 values. This is useful for key rollover for example where several key 3625 sets may exist with different names. 3626 [Steve Henson] 3627 3628 *) Reverse ENGINE-internal logic for caching default ENGINE handles. 3629 This was broken until now in 0.9.8 releases, such that the only way 3630 a registered ENGINE could be used (assuming it initialises 3631 successfully on the host) was to explicitly set it as the default 3632 for the relevant algorithms. This is in contradiction with 0.9.7 3633 behaviour and the documentation. With this fix, when an ENGINE is 3634 registered into a given algorithm's table of implementations, the 3635 'uptodate' flag is reset so that auto-discovery will be used next 3636 time a new context for that algorithm attempts to select an 3637 implementation. 3638 [Ian Lister (tweaked by Geoff Thorpe)] 3639 3640 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 3641 implemention in the following ways: 3642 3643 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be 3644 hard coded. 3645 3646 Lack of BER streaming support means one pass streaming processing is 3647 only supported if data is detached: setting the streaming flag is 3648 ignored for embedded content. 3649 3650 CMS support is disabled by default and must be explicitly enabled 3651 with the enable-cms configuration option. 3652 [Steve Henson] 3653 3654 *) Update the GMP engine glue to do direct copies between BIGNUM and 3655 mpz_t when openssl and GMP use the same limb size. Otherwise the 3656 existing "conversion via a text string export" trick is still used. 3657 [Paul Sheer <paulsheer@gmail.com>] 3658 3659 *) Zlib compression BIO. This is a filter BIO which compressed and 3660 uncompresses any data passed through it. 3661 [Steve Henson] 3662 3663 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement 3664 RFC3394 compatible AES key wrapping. 3665 [Steve Henson] 3666 3667 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): 3668 sets string data without copying. X509_ALGOR_set0() and 3669 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) 3670 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data 3671 from an X509_ATTRIBUTE structure optionally checking it occurs only 3672 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied 3673 data. 3674 [Steve Henson] 3675 3676 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() 3677 to get the expected BN_FLG_CONSTTIME behavior. 3678 [Bodo Moeller (Google)] 3679 3680 *) Netware support: 3681 3682 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets 3683 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) 3684 - added some more tests to do_tests.pl 3685 - fixed RunningProcess usage so that it works with newer LIBC NDKs too 3686 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency 3687 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, 3688 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc 3689 - various changes to netware.pl to enable gcc-cross builds on Win32 3690 platform 3691 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) 3692 - various changes to fix missing prototype warnings 3693 - fixed x86nasm.pl to create correct asm files for NASM COFF output 3694 - added AES, WHIRLPOOL and CPUID assembler code to build files 3695 - added missing AES assembler make rules to mk1mf.pl 3696 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply 3697 [Guenter Knauf <eflash@gmx.net>] 3698 3699 *) Implement certificate status request TLS extension defined in RFC3546. 3700 A client can set the appropriate parameters and receive the encoded 3701 OCSP response via a callback. A server can query the supplied parameters 3702 and set the encoded OCSP response in the callback. Add simplified examples 3703 to s_client and s_server. 3704 [Steve Henson] 3705 3706 Changes between 0.9.8f and 0.9.8g [19 Oct 2007] 3707 3708 *) Fix various bugs: 3709 + Binary incompatibility of ssl_ctx_st structure 3710 + DTLS interoperation with non-compliant servers 3711 + Don't call get_session_cb() without proposed session 3712 + Fix ia64 assembler code 3713 [Andy Polyakov, Steve Henson] 3714 3715 Changes between 0.9.8e and 0.9.8f [11 Oct 2007] 3716 3717 *) DTLS Handshake overhaul. There were longstanding issues with 3718 OpenSSL DTLS implementation, which were making it impossible for 3719 RFC 4347 compliant client to communicate with OpenSSL server. 3720 Unfortunately just fixing these incompatibilities would "cut off" 3721 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e 3722 server keeps tolerating non RFC compliant syntax. The opposite is 3723 not true, 0.9.8f client can not communicate with earlier server. 3724 This update even addresses CVE-2007-4995. 3725 [Andy Polyakov] 3726 3727 *) Changes to avoid need for function casts in OpenSSL: some compilers 3728 (gcc 4.2 and later) reject their use. 3729 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, 3730 Steve Henson] 3731 3732 *) Add RFC4507 support to OpenSSL. This includes the corrections in 3733 RFC4507bis. The encrypted ticket format is an encrypted encoded 3734 SSL_SESSION structure, that way new session features are automatically 3735 supported. 3736 3737 If a client application caches session in an SSL_SESSION structure 3738 support is transparent because tickets are now stored in the encoded 3739 SSL_SESSION. 3740 3741 The SSL_CTX structure automatically generates keys for ticket 3742 protection in servers so again support should be possible 3743 with no application modification. 3744 3745 If a client or server wishes to disable RFC4507 support then the option 3746 SSL_OP_NO_TICKET can be set. 3747 3748 Add a TLS extension debugging callback to allow the contents of any client 3749 or server extensions to be examined. 3750 3751 This work was sponsored by Google. 3752 [Steve Henson] 3753 3754 *) Add initial support for TLS extensions, specifically for the server_name 3755 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 3756 have new members for a host name. The SSL data structure has an 3757 additional member SSL_CTX *initial_ctx so that new sessions can be 3758 stored in that context to allow for session resumption, even after the 3759 SSL has been switched to a new SSL_CTX in reaction to a client's 3760 server_name extension. 3761 3762 New functions (subject to change): 3763 3764 SSL_get_servername() 3765 SSL_get_servername_type() 3766 SSL_set_SSL_CTX() 3767 3768 New CTRL codes and macros (subject to change): 3769 3770 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 3771 - SSL_CTX_set_tlsext_servername_callback() 3772 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 3773 - SSL_CTX_set_tlsext_servername_arg() 3774 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 3775 3776 openssl s_client has a new '-servername ...' option. 3777 3778 openssl s_server has new options '-servername_host ...', '-cert2 ...', 3779 '-key2 ...', '-servername_fatal' (subject to change). This allows 3780 testing the HostName extension for a specific single host name ('-cert' 3781 and '-key' remain fallbacks for handshakes without HostName 3782 negotiation). If the unrecogninzed_name alert has to be sent, this by 3783 default is a warning; it becomes fatal with the '-servername_fatal' 3784 option. 3785 3786 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson] 3787 3788 *) Add AES and SSE2 assembly language support to VC++ build. 3789 [Steve Henson] 3790 3791 *) Mitigate attack on final subtraction in Montgomery reduction. 3792 [Andy Polyakov] 3793 3794 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 3795 (which previously caused an internal error). 3796 [Bodo Moeller] 3797 3798 *) Squeeze another 10% out of IGE mode when in != out. 3799 [Ben Laurie] 3800 3801 *) AES IGE mode speedup. 3802 [Dean Gaudet (Google)] 3803 3804 *) Add the Korean symmetric 128-bit cipher SEED (see 3805 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and 3806 add SEED ciphersuites from RFC 4162: 3807 3808 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" 3809 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" 3810 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" 3811 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" 3812 3813 To minimize changes between patchlevels in the OpenSSL 0.9.8 3814 series, SEED remains excluded from compilation unless OpenSSL 3815 is configured with 'enable-seed'. 3816 [KISA, Bodo Moeller] 3817 3818 *) Mitigate branch prediction attacks, which can be practical if a 3819 single processor is shared, allowing a spy process to extract 3820 information. For detailed background information, see 3821 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron, 3822 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL 3823 and Necessary Software Countermeasures"). The core of the change 3824 are new versions BN_div_no_branch() and 3825 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), 3826 respectively, which are slower, but avoid the security-relevant 3827 conditional branches. These are automatically called by BN_div() 3828 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one 3829 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to 3830 remove a conditional branch. 3831 3832 BN_FLG_CONSTTIME is the new name for the previous 3833 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just 3834 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag 3835 in the exponent causes BN_mod_exp_mont() to use the alternative 3836 implementation in BN_mod_exp_mont_consttime().) The old name 3837 remains as a deprecated alias. 3838 3839 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general 3840 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses 3841 constant-time implementations for more than just exponentiation. 3842 Here too the old name is kept as a deprecated alias. 3843 3844 BN_BLINDING_new() will now use BN_dup() for the modulus so that 3845 the BN_BLINDING structure gets an independent copy of the 3846 modulus. This means that the previous "BIGNUM *m" argument to 3847 BN_BLINDING_new() and to BN_BLINDING_create_param() now 3848 essentially becomes "const BIGNUM *m", although we can't actually 3849 change this in the header file before 0.9.9. It allows 3850 RSA_setup_blinding() to use BN_with_flags() on the modulus to 3851 enable BN_FLG_CONSTTIME. 3852 3853 [Matthew D Wood (Intel Corp)] 3854 3855 *) In the SSL/TLS server implementation, be strict about session ID 3856 context matching (which matters if an application uses a single 3857 external cache for different purposes). Previously, 3858 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was 3859 set. This did ensure strict client verification, but meant that, 3860 with applications using a single external cache for quite 3861 different requirements, clients could circumvent ciphersuite 3862 restrictions for a given session ID context by starting a session 3863 in a different context. 3864 [Bodo Moeller] 3865 3866 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 3867 a ciphersuite string such as "DEFAULT:RSA" cannot enable 3868 authentication-only ciphersuites. 3869 [Bodo Moeller] 3870 3871 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was 3872 not complete and could lead to a possible single byte overflow 3873 (CVE-2007-5135) [Ben Laurie] 3874 3875 Changes between 0.9.8d and 0.9.8e [23 Feb 2007] 3876 3877 *) Since AES128 and AES256 (and similarly Camellia128 and 3878 Camellia256) share a single mask bit in the logic of 3879 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 3880 kludge to work properly if AES128 is available and AES256 isn't 3881 (or if Camellia128 is available and Camellia256 isn't). 3882 [Victor Duchovni] 3883 3884 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c 3885 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): 3886 When a point or a seed is encoded in a BIT STRING, we need to 3887 prevent the removal of trailing zero bits to get the proper DER 3888 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case 3889 of a NamedBitList, for which trailing 0 bits need to be removed.) 3890 [Bodo Moeller] 3891 3892 *) Have SSL/TLS server implementation tolerate "mismatched" record 3893 protocol version while receiving ClientHello even if the 3894 ClientHello is fragmented. (The server can't insist on the 3895 particular protocol version it has chosen before the ServerHello 3896 message has informed the client about his choice.) 3897 [Bodo Moeller] 3898 3899 *) Add RFC 3779 support. 3900 [Rob Austein for ARIN, Ben Laurie] 3901 3902 *) Load error codes if they are not already present instead of using a 3903 static variable. This allows them to be cleanly unloaded and reloaded. 3904 Improve header file function name parsing. 3905 [Steve Henson] 3906 3907 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO 3908 or CAPABILITY handshake as required by RFCs. 3909 [Goetz Babin-Ebell] 3910 3911 Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 3912 3913 *) Introduce limits to prevent malicious keys being able to 3914 cause a denial of service. (CVE-2006-2940) 3915 [Steve Henson, Bodo Moeller] 3916 3917 *) Fix ASN.1 parsing of certain invalid structures that can result 3918 in a denial of service. (CVE-2006-2937) [Steve Henson] 3919 3920 *) Fix buffer overflow in SSL_get_shared_ciphers() function. 3921 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 3922 3923 *) Fix SSL client code which could crash if connecting to a 3924 malicious SSLv2 server. (CVE-2006-4343) 3925 [Tavis Ormandy and Will Drewry, Google Security Team] 3926 3927 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 3928 match only those. Before that, "AES256-SHA" would be interpreted 3929 as a pattern and match "AES128-SHA" too (since AES128-SHA got 3930 the same strength classification in 0.9.7h) as we currently only 3931 have a single AES bit in the ciphersuite description bitmap. 3932 That change, however, also applied to ciphersuite strings such as 3933 "RC4-MD5" that intentionally matched multiple ciphersuites -- 3934 namely, SSL 2.0 ciphersuites in addition to the more common ones 3935 from SSL 3.0/TLS 1.0. 3936 3937 So we change the selection algorithm again: Naming an explicit 3938 ciphersuite selects this one ciphersuite, and any other similar 3939 ciphersuite (same bitmap) from *other* protocol versions. 3940 Thus, "RC4-MD5" again will properly select both the SSL 2.0 3941 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 3942 3943 Since SSL 2.0 does not have any ciphersuites for which the 3944 128/256 bit distinction would be relevant, this works for now. 3945 The proper fix will be to use different bits for AES128 and 3946 AES256, which would have avoided the problems from the beginning; 3947 however, bits are scarce, so we can only do this in a new release 3948 (not just a patchlevel) when we can change the SSL_CIPHER 3949 definition to split the single 'unsigned long mask' bitmap into 3950 multiple values to extend the available space. 3951 3952 [Bodo Moeller] 3953 3954 Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 3955 3956 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 3957 (CVE-2006-4339) [Ben Laurie and Google Security Team] 3958 3959 *) Add AES IGE and biIGE modes. 3960 [Ben Laurie] 3961 3962 *) Change the Unix randomness entropy gathering to use poll() when 3963 possible instead of select(), since the latter has some 3964 undesirable limitations. 3965 [Darryl Miles via Richard Levitte and Bodo Moeller] 3966 3967 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special 3968 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 3969 cannot be implicitly activated as part of, e.g., the "AES" alias. 3970 However, please upgrade to OpenSSL 0.9.9[-dev] for 3971 non-experimental use of the ECC ciphersuites to get TLS extension 3972 support, which is required for curve and point format negotiation 3973 to avoid potential handshake problems. 3974 [Bodo Moeller] 3975 3976 *) Disable rogue ciphersuites: 3977 3978 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 3979 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 3980 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 3981 3982 The latter two were purportedly from 3983 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 3984 appear there. 3985 3986 Also deactivate the remaining ciphersuites from 3987 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 3988 unofficial, and the ID has long expired. 3989 [Bodo Moeller] 3990 3991 *) Fix RSA blinding Heisenbug (problems sometimes occured on 3992 dual-core machines) and other potential thread-safety issues. 3993 [Bodo Moeller] 3994 3995 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 3996 versions), which is now available for royalty-free use 3997 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). 3998 Also, add Camellia TLS ciphersuites from RFC 4132. 3999 4000 To minimize changes between patchlevels in the OpenSSL 0.9.8 4001 series, Camellia remains excluded from compilation unless OpenSSL 4002 is configured with 'enable-camellia'. 4003 [NTT] 4004 4005 *) Disable the padding bug check when compression is in use. The padding 4006 bug check assumes the first packet is of even length, this is not 4007 necessarily true if compresssion is enabled and can result in false 4008 positives causing handshake failure. The actual bug test is ancient 4009 code so it is hoped that implementations will either have fixed it by 4010 now or any which still have the bug do not support compression. 4011 [Steve Henson] 4012 4013 Changes between 0.9.8a and 0.9.8b [04 May 2006] 4014 4015 *) When applying a cipher rule check to see if string match is an explicit 4016 cipher suite and only match that one cipher suite if it is. 4017 [Steve Henson] 4018 4019 *) Link in manifests for VC++ if needed. 4020 [Austin Ziegler <halostatue@gmail.com>] 4021 4022 *) Update support for ECC-based TLS ciphersuites according to 4023 draft-ietf-tls-ecc-12.txt with proposed changes (but without 4024 TLS extensions, which are supported starting with the 0.9.9 4025 branch, not in the OpenSSL 0.9.8 branch). 4026 [Douglas Stebila] 4027 4028 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 4029 opaque EVP_CIPHER_CTX handling. 4030 [Steve Henson] 4031 4032 *) Fixes and enhancements to zlib compression code. We now only use 4033 "zlib1.dll" and use the default __cdecl calling convention on Win32 4034 to conform with the standards mentioned here: 4035 http://www.zlib.net/DLL_FAQ.txt 4036 Static zlib linking now works on Windows and the new --with-zlib-include 4037 --with-zlib-lib options to Configure can be used to supply the location 4038 of the headers and library. Gracefully handle case where zlib library 4039 can't be loaded. 4040 [Steve Henson] 4041 4042 *) Several fixes and enhancements to the OID generation code. The old code 4043 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 4044 handle numbers larger than ULONG_MAX, truncated printing and had a 4045 non standard OBJ_obj2txt() behaviour. 4046 [Steve Henson] 4047 4048 *) Add support for building of engines under engine/ as shared libraries 4049 under VC++ build system. 4050 [Steve Henson] 4051 4052 *) Corrected the numerous bugs in the Win32 path splitter in DSO. 4053 Hopefully, we will not see any false combination of paths any more. 4054 [Richard Levitte] 4055 4056 Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 4057 4058 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 4059 (part of SSL_OP_ALL). This option used to disable the 4060 countermeasure against man-in-the-middle protocol-version 4061 rollback in the SSL 2.0 server implementation, which is a bad 4062 idea. (CVE-2005-2969) 4063 4064 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 4065 for Information Security, National Institute of Advanced Industrial 4066 Science and Technology [AIST], Japan)] 4067 4068 *) Add two function to clear and return the verify parameter flags. 4069 [Steve Henson] 4070 4071 *) Keep cipherlists sorted in the source instead of sorting them at 4072 runtime, thus removing the need for a lock. 4073 [Nils Larsch] 4074 4075 *) Avoid some small subgroup attacks in Diffie-Hellman. 4076 [Nick Mathewson and Ben Laurie] 4077 4078 *) Add functions for well-known primes. 4079 [Nick Mathewson] 4080 4081 *) Extended Windows CE support. 4082 [Satoshi Nakamura and Andy Polyakov] 4083 4084 *) Initialize SSL_METHOD structures at compile time instead of during 4085 runtime, thus removing the need for a lock. 4086 [Steve Henson] 4087 4088 *) Make PKCS7_decrypt() work even if no certificate is supplied by 4089 attempting to decrypt each encrypted key in turn. Add support to 4090 smime utility. 4091 [Steve Henson] 4092 4093 Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 4094 4095 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 4096 OpenSSL 0.9.8.] 4097 4098 *) Add libcrypto.pc and libssl.pc for those who feel they need them. 4099 [Richard Levitte] 4100 4101 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 4102 key into the same file any more. 4103 [Richard Levitte] 4104 4105 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 4106 [Andy Polyakov] 4107 4108 *) Add -utf8 command line and config file option to 'ca'. 4109 [Stefan <stf@udoma.org] 4110 4111 *) Removed the macro des_crypt(), as it seems to conflict with some 4112 libraries. Use DES_crypt(). 4113 [Richard Levitte] 4114 4115 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 4116 involves renaming the source and generated shared-libs for 4117 both. The engines will accept the corrected or legacy ids 4118 ('ncipher' and '4758_cca' respectively) when binding. NB, 4119 this only applies when building 'shared'. 4120 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 4121 4122 *) Add attribute functions to EVP_PKEY structure. Modify 4123 PKCS12_create() to recognize a CSP name attribute and 4124 use it. Make -CSP option work again in pkcs12 utility. 4125 [Steve Henson] 4126 4127 *) Add new functionality to the bn blinding code: 4128 - automatic re-creation of the BN_BLINDING parameters after 4129 a fixed number of uses (currently 32) 4130 - add new function for parameter creation 4131 - introduce flags to control the update behaviour of the 4132 BN_BLINDING parameters 4133 - hide BN_BLINDING structure 4134 Add a second BN_BLINDING slot to the RSA structure to improve 4135 performance when a single RSA object is shared among several 4136 threads. 4137 [Nils Larsch] 4138 4139 *) Add support for DTLS. 4140 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 4141 4142 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 4143 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 4144 [Walter Goulet] 4145 4146 *) Remove buggy and incompletet DH cert support from 4147 ssl/ssl_rsa.c and ssl/s3_both.c 4148 [Nils Larsch] 4149 4150 *) Use SHA-1 instead of MD5 as the default digest algorithm for 4151 the apps/openssl applications. 4152 [Nils Larsch] 4153 4154 *) Compile clean with "-Wall -Wmissing-prototypes 4155 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 4156 DEBUG_SAFESTACK must also be set. 4157 [Ben Laurie] 4158 4159 *) Change ./Configure so that certain algorithms can be disabled by default. 4160 The new counterpiece to "no-xxx" is "enable-xxx". 4161 4162 The patented RC5 and MDC2 algorithms will now be disabled unless 4163 "enable-rc5" and "enable-mdc2", respectively, are specified. 4164 4165 (IDEA remains enabled despite being patented. This is because IDEA 4166 is frequently required for interoperability, and there is no license 4167 fee for non-commercial use. As before, "no-idea" can be used to 4168 avoid this algorithm.) 4169 4170 [Bodo Moeller] 4171 4172 *) Add processing of proxy certificates (see RFC 3820). This work was 4173 sponsored by KTH (The Royal Institute of Technology in Stockholm) and 4174 EGEE (Enabling Grids for E-science in Europe). 4175 [Richard Levitte] 4176 4177 *) RC4 performance overhaul on modern architectures/implementations, such 4178 as Intel P4, IA-64 and AMD64. 4179 [Andy Polyakov] 4180 4181 *) New utility extract-section.pl. This can be used specify an alternative 4182 section number in a pod file instead of having to treat each file as 4183 a separate case in Makefile. This can be done by adding two lines to the 4184 pod file: 4185 4186 =for comment openssl_section:XXX 4187 4188 The blank line is mandatory. 4189 4190 [Steve Henson] 4191 4192 *) New arguments -certform, -keyform and -pass for s_client and s_server 4193 to allow alternative format key and certificate files and passphrase 4194 sources. 4195 [Steve Henson] 4196 4197 *) New structure X509_VERIFY_PARAM which combines current verify parameters, 4198 update associated structures and add various utility functions. 4199 4200 Add new policy related verify parameters, include policy checking in 4201 standard verify code. Enhance 'smime' application with extra parameters 4202 to support policy checking and print out. 4203 [Steve Henson] 4204 4205 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 4206 Nehemiah processors. These extensions support AES encryption in hardware 4207 as well as RNG (though RNG support is currently disabled). 4208 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 4209 4210 *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 4211 [Geoff Thorpe] 4212 4213 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 4214 [Andy Polyakov and a number of other people] 4215 4216 *) Improved PowerPC platform support. Most notably BIGNUM assembler 4217 implementation contributed by IBM. 4218 [Suresh Chari, Peter Waltenberg, Andy Polyakov] 4219 4220 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 4221 exponent rather than 'unsigned long'. There is a corresponding change to 4222 the new 'rsa_keygen' element of the RSA_METHOD structure. 4223 [Jelte Jansen, Geoff Thorpe] 4224 4225 *) Functionality for creating the initial serial number file is now 4226 moved from CA.pl to the 'ca' utility with a new option -create_serial. 4227 4228 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 4229 number file to 1, which is bound to cause problems. To avoid 4230 the problems while respecting compatibility between different 0.9.7 4231 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 4232 CA.pl for serial number initialization. With the new release 0.9.8, 4233 we can fix the problem directly in the 'ca' utility.) 4234 [Steve Henson] 4235 4236 *) Reduced header interdepencies by declaring more opaque objects in 4237 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 4238 give fewer recursive includes, which could break lazy source code - so 4239 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 4240 developers should define this symbol when building and using openssl to 4241 ensure they track the recommended behaviour, interfaces, [etc], but 4242 backwards-compatible behaviour prevails when this isn't defined. 4243 [Geoff Thorpe] 4244 4245 *) New function X509_POLICY_NODE_print() which prints out policy nodes. 4246 [Steve Henson] 4247 4248 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 4249 This will generate a random key of the appropriate length based on the 4250 cipher context. The EVP_CIPHER can provide its own random key generation 4251 routine to support keys of a specific form. This is used in the des and 4252 3des routines to generate a key of the correct parity. Update S/MIME 4253 code to use new functions and hence generate correct parity DES keys. 4254 Add EVP_CHECK_DES_KEY #define to return an error if the key is not 4255 valid (weak or incorrect parity). 4256 [Steve Henson] 4257 4258 *) Add a local set of CRLs that can be used by X509_verify_cert() as well 4259 as looking them up. This is useful when the verified structure may contain 4260 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 4261 present unless the new PKCS7_NO_CRL flag is asserted. 4262 [Steve Henson] 4263 4264 *) Extend ASN1 oid configuration module. It now additionally accepts the 4265 syntax: 4266 4267 shortName = some long name, 1.2.3.4 4268 [Steve Henson] 4269 4270 *) Reimplemented the BN_CTX implementation. There is now no more static 4271 limitation on the number of variables it can handle nor the depth of the 4272 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 4273 information can now expand as required, and rather than having a single 4274 static array of bignums, BN_CTX now uses a linked-list of such arrays 4275 allowing it to expand on demand whilst maintaining the usefulness of 4276 BN_CTX's "bundling". 4277 [Geoff Thorpe] 4278 4279 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 4280 to allow all RSA operations to function using a single BN_CTX. 4281 [Geoff Thorpe] 4282 4283 *) Preliminary support for certificate policy evaluation and checking. This 4284 is initially intended to pass the tests outlined in "Conformance Testing 4285 of Relying Party Client Certificate Path Processing Logic" v1.07. 4286 [Steve Henson] 4287 4288 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 4289 remained unused and not that useful. A variety of other little bignum 4290 tweaks and fixes have also been made continuing on from the audit (see 4291 below). 4292 [Geoff Thorpe] 4293 4294 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 4295 associated ASN1, EVP and SSL functions and old ASN1 macros. 4296 [Richard Levitte] 4297 4298 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 4299 and this should never fail. So the return value from the use of 4300 BN_set_word() (which can fail due to needless expansion) is now deprecated; 4301 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 4302 [Geoff Thorpe] 4303 4304 *) BN_CTX_get() should return zero-valued bignums, providing the same 4305 initialised value as BN_new(). 4306 [Geoff Thorpe, suggested by Ulf M��ller] 4307 4308 *) Support for inhibitAnyPolicy certificate extension. 4309 [Steve Henson] 4310 4311 *) An audit of the BIGNUM code is underway, for which debugging code is 4312 enabled when BN_DEBUG is defined. This makes stricter enforcements on what 4313 is considered valid when processing BIGNUMs, and causes execution to 4314 assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 4315 further steps are taken to deliberately pollute unused data in BIGNUM 4316 structures to try and expose faulty code further on. For now, openssl will 4317 (in its default mode of operation) continue to tolerate the inconsistent 4318 forms that it has tolerated in the past, but authors and packagers should 4319 consider trying openssl and their own applications when compiled with 4320 these debugging symbols defined. It will help highlight potential bugs in 4321 their own code, and will improve the test coverage for OpenSSL itself. At 4322 some point, these tighter rules will become openssl's default to improve 4323 maintainability, though the assert()s and other overheads will remain only 4324 in debugging configurations. See bn.h for more details. 4325 [Geoff Thorpe, Nils Larsch, Ulf M��ller] 4326 4327 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 4328 that can only be obtained through BN_CTX_new() (which implicitly 4329 initialises it). The presence of this function only made it possible 4330 to overwrite an existing structure (and cause memory leaks). 4331 [Geoff Thorpe] 4332 4333 *) Because of the callback-based approach for implementing LHASH as a 4334 template type, lh_insert() adds opaque objects to hash-tables and 4335 lh_doall() or lh_doall_arg() are typically used with a destructor callback 4336 to clean up those corresponding objects before destroying the hash table 4337 (and losing the object pointers). So some over-zealous constifications in 4338 LHASH have been relaxed so that lh_insert() does not take (nor store) the 4339 objects as "const" and the lh_doall[_arg] callback wrappers are not 4340 prototyped to have "const" restrictions on the object pointers they are 4341 given (and so aren't required to cast them away any more). 4342 [Geoff Thorpe] 4343 4344 *) The tmdiff.h API was so ugly and minimal that our own timing utility 4345 (speed) prefers to use its own implementation. The two implementations 4346 haven't been consolidated as yet (volunteers?) but the tmdiff API has had 4347 its object type properly exposed (MS_TM) instead of casting to/from "char 4348 *". This may still change yet if someone realises MS_TM and "ms_time_***" 4349 aren't necessarily the greatest nomenclatures - but this is what was used 4350 internally to the implementation so I've used that for now. 4351 [Geoff Thorpe] 4352 4353 *) Ensure that deprecated functions do not get compiled when 4354 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 4355 the self-tests were still using deprecated key-generation functions so 4356 these have been updated also. 4357 [Geoff Thorpe] 4358 4359 *) Reorganise PKCS#7 code to separate the digest location functionality 4360 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 4361 New function PKCS7_set_digest() to set the digest type for PKCS#7 4362 digestedData type. Add additional code to correctly generate the 4363 digestedData type and add support for this type in PKCS7 initialization 4364 functions. 4365 [Steve Henson] 4366 4367 *) New function PKCS7_set0_type_other() this initializes a PKCS7 4368 structure of type "other". 4369 [Steve Henson] 4370 4371 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 4372 sure the loop does correctly stop and breaking ("division by zero") 4373 modulus operations are not performed. The (pre-generated) prime 4374 table crypto/bn/bn_prime.h was already correct, but it could not be 4375 re-generated on some platforms because of the "division by zero" 4376 situation in the script. 4377 [Ralf S. Engelschall] 4378 4379 *) Update support for ECC-based TLS ciphersuites according to 4380 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 4381 SHA-1 now is only used for "small" curves (where the 4382 representation of a field element takes up to 24 bytes); for 4383 larger curves, the field element resulting from ECDH is directly 4384 used as premaster secret. 4385 [Douglas Stebila (Sun Microsystems Laboratories)] 4386 4387 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 4388 curve secp160r1 to the tests. 4389 [Douglas Stebila (Sun Microsystems Laboratories)] 4390 4391 *) Add the possibility to load symbols globally with DSO. 4392 [G��tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 4393 4394 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 4395 control of the error stack. 4396 [Richard Levitte] 4397 4398 *) Add support for STORE in ENGINE. 4399 [Richard Levitte] 4400 4401 *) Add the STORE type. The intention is to provide a common interface 4402 to certificate and key stores, be they simple file-based stores, or 4403 HSM-type store, or LDAP stores, or... 4404 NOTE: The code is currently UNTESTED and isn't really used anywhere. 4405 [Richard Levitte] 4406 4407 *) Add a generic structure called OPENSSL_ITEM. This can be used to 4408 pass a list of arguments to any function as well as provide a way 4409 for a function to pass data back to the caller. 4410 [Richard Levitte] 4411 4412 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 4413 works like BUF_strdup() but can be used to duplicate a portion of 4414 a string. The copy gets NUL-terminated. BUF_memdup() duplicates 4415 a memory area. 4416 [Richard Levitte] 4417 4418 *) Add the function sk_find_ex() which works like sk_find(), but will 4419 return an index to an element even if an exact match couldn't be 4420 found. The index is guaranteed to point at the element where the 4421 searched-for key would be inserted to preserve sorting order. 4422 [Richard Levitte] 4423 4424 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 4425 takes an extra flags argument for optional functionality. Currently, 4426 the following flags are defined: 4427 4428 OBJ_BSEARCH_VALUE_ON_NOMATCH 4429 This one gets OBJ_bsearch_ex() to return a pointer to the first 4430 element where the comparing function returns a negative or zero 4431 number. 4432 4433 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 4434 This one gets OBJ_bsearch_ex() to return a pointer to the first 4435 element where the comparing function returns zero. This is useful 4436 if there are more than one element where the comparing function 4437 returns zero. 4438 [Richard Levitte] 4439 4440 *) Make it possible to create self-signed certificates with 'openssl ca' 4441 in such a way that the self-signed certificate becomes part of the 4442 CA database and uses the same mechanisms for serial number generation 4443 as all other certificate signing. The new flag '-selfsign' enables 4444 this functionality. Adapt CA.sh and CA.pl.in. 4445 [Richard Levitte] 4446 4447 *) Add functionality to check the public key of a certificate request 4448 against a given private. This is useful to check that a certificate 4449 request can be signed by that key (self-signing). 4450 [Richard Levitte] 4451 4452 *) Make it possible to have multiple active certificates with the same 4453 subject in the CA index file. This is done only if the keyword 4454 'unique_subject' is set to 'no' in the main CA section (default 4455 if 'CA_default') of the configuration file. The value is saved 4456 with the database itself in a separate index attribute file, 4457 named like the index file with '.attr' appended to the name. 4458 [Richard Levitte] 4459 4460 *) Generate muti valued AVAs using '+' notation in config files for 4461 req and dirName. 4462 [Steve Henson] 4463 4464 *) Support for nameConstraints certificate extension. 4465 [Steve Henson] 4466 4467 *) Support for policyConstraints certificate extension. 4468 [Steve Henson] 4469 4470 *) Support for policyMappings certificate extension. 4471 [Steve Henson] 4472 4473 *) Make sure the default DSA_METHOD implementation only uses its 4474 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 4475 and change its own handlers to be NULL so as to remove unnecessary 4476 indirection. This lets alternative implementations fallback to the 4477 default implementation more easily. 4478 [Geoff Thorpe] 4479 4480 *) Support for directoryName in GeneralName related extensions 4481 in config files. 4482 [Steve Henson] 4483 4484 *) Make it possible to link applications using Makefile.shared. 4485 Make that possible even when linking against static libraries! 4486 [Richard Levitte] 4487 4488 *) Support for single pass processing for S/MIME signing. This now 4489 means that S/MIME signing can be done from a pipe, in addition 4490 cleartext signing (multipart/signed type) is effectively streaming 4491 and the signed data does not need to be all held in memory. 4492 4493 This is done with a new flag PKCS7_STREAM. When this flag is set 4494 PKCS7_sign() only initializes the PKCS7 structure and the actual signing 4495 is done after the data is output (and digests calculated) in 4496 SMIME_write_PKCS7(). 4497 [Steve Henson] 4498 4499 *) Add full support for -rpath/-R, both in shared libraries and 4500 applications, at least on the platforms where it's known how 4501 to do it. 4502 [Richard Levitte] 4503 4504 *) In crypto/ec/ec_mult.c, implement fast point multiplication with 4505 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 4506 will now compute a table of multiples of the generator that 4507 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 4508 faster (notably in the case of a single point multiplication, 4509 scalar * generator). 4510 [Nils Larsch, Bodo Moeller] 4511 4512 *) IPv6 support for certificate extensions. The various extensions 4513 which use the IP:a.b.c.d can now take IPv6 addresses using the 4514 formats of RFC1884 2.2 . IPv6 addresses are now also displayed 4515 correctly. 4516 [Steve Henson] 4517 4518 *) Added an ENGINE that implements RSA by performing private key 4519 exponentiations with the GMP library. The conversions to and from 4520 GMP's mpz_t format aren't optimised nor are any montgomery forms 4521 cached, and on x86 it appears OpenSSL's own performance has caught up. 4522 However there are likely to be other architectures where GMP could 4523 provide a boost. This ENGINE is not built in by default, but it can be 4524 specified at Configure time and should be accompanied by the necessary 4525 linker additions, eg; 4526 ./config -DOPENSSL_USE_GMP -lgmp 4527 [Geoff Thorpe] 4528 4529 *) "openssl engine" will not display ENGINE/DSO load failure errors when 4530 testing availability of engines with "-t" - the old behaviour is 4531 produced by increasing the feature's verbosity with "-tt". 4532 [Geoff Thorpe] 4533 4534 *) ECDSA routines: under certain error conditions uninitialized BN objects 4535 could be freed. Solution: make sure initialization is performed early 4536 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 4537 via PR#459) 4538 [Lutz Jaenicke] 4539 4540 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 4541 and DH_METHOD (eg. by ENGINE implementations) to override the normal 4542 software implementations. For DSA and DH, parameter generation can 4543 also be overriden by providing the appropriate method callbacks. 4544 [Geoff Thorpe] 4545 4546 *) Change the "progress" mechanism used in key-generation and 4547 primality testing to functions that take a new BN_GENCB pointer in 4548 place of callback/argument pairs. The new API functions have "_ex" 4549 postfixes and the older functions are reimplemented as wrappers for 4550 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 4551 declarations of the old functions to help (graceful) attempts to 4552 migrate to the new functions. Also, the new key-generation API 4553 functions operate on a caller-supplied key-structure and return 4554 success/failure rather than returning a key or NULL - this is to 4555 help make "keygen" another member function of RSA_METHOD etc. 4556 4557 Example for using the new callback interface: 4558 4559 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 4560 void *my_arg = ...; 4561 BN_GENCB my_cb; 4562 4563 BN_GENCB_set(&my_cb, my_callback, my_arg); 4564 4565 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 4566 /* For the meaning of a, b in calls to my_callback(), see the 4567 * documentation of the function that calls the callback. 4568 * cb will point to my_cb; my_arg can be retrieved as cb->arg. 4569 * my_callback should return 1 if it wants BN_is_prime_ex() 4570 * to continue, or 0 to stop. 4571 */ 4572 4573 [Geoff Thorpe] 4574 4575 *) Change the ZLIB compression method to be stateful, and make it 4576 available to TLS with the number defined in 4577 draft-ietf-tls-compression-04.txt. 4578 [Richard Levitte] 4579 4580 *) Add the ASN.1 structures and functions for CertificatePair, which 4581 is defined as follows (according to X.509_4thEditionDraftV6.pdf): 4582 4583 CertificatePair ::= SEQUENCE { 4584 forward [0] Certificate OPTIONAL, 4585 reverse [1] Certificate OPTIONAL, 4586 -- at least one of the pair shall be present -- } 4587 4588 Also implement the PEM functions to read and write certificate 4589 pairs, and defined the PEM tag as "CERTIFICATE PAIR". 4590 4591 This needed to be defined, mostly for the sake of the LDAP 4592 attribute crossCertificatePair, but may prove useful elsewhere as 4593 well. 4594 [Richard Levitte] 4595 4596 *) Make it possible to inhibit symlinking of shared libraries in 4597 Makefile.shared, for Cygwin's sake. 4598 [Richard Levitte] 4599 4600 *) Extend the BIGNUM API by creating a function 4601 void BN_set_negative(BIGNUM *a, int neg); 4602 and a macro that behave like 4603 int BN_is_negative(const BIGNUM *a); 4604 4605 to avoid the need to access 'a->neg' directly in applications. 4606 [Nils Larsch] 4607 4608 *) Implement fast modular reduction for pseudo-Mersenne primes 4609 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 4610 EC_GROUP_new_curve_GFp() will now automatically use this 4611 if applicable. 4612 [Nils Larsch <nla@trustcenter.de>] 4613 4614 *) Add new lock type (CRYPTO_LOCK_BN). 4615 [Bodo Moeller] 4616 4617 *) Change the ENGINE framework to automatically load engines 4618 dynamically from specific directories unless they could be 4619 found to already be built in or loaded. Move all the 4620 current engines except for the cryptodev one to a new 4621 directory engines/. 4622 The engines in engines/ are built as shared libraries if 4623 the "shared" options was given to ./Configure or ./config. 4624 Otherwise, they are inserted in libcrypto.a. 4625 /usr/local/ssl/engines is the default directory for dynamic 4626 engines, but that can be overriden at configure time through 4627 the usual use of --prefix and/or --openssldir, and at run 4628 time with the environment variable OPENSSL_ENGINES. 4629 [Geoff Thorpe and Richard Levitte] 4630 4631 *) Add Makefile.shared, a helper makefile to build shared 4632 libraries. Addapt Makefile.org. 4633 [Richard Levitte] 4634 4635 *) Add version info to Win32 DLLs. 4636 [Peter 'Luna' Runestig" <peter@runestig.com>] 4637 4638 *) Add new 'medium level' PKCS#12 API. Certificates and keys 4639 can be added using this API to created arbitrary PKCS#12 4640 files while avoiding the low level API. 4641 4642 New options to PKCS12_create(), key or cert can be NULL and 4643 will then be omitted from the output file. The encryption 4644 algorithm NIDs can be set to -1 for no encryption, the mac 4645 iteration count can be set to 0 to omit the mac. 4646 4647 Enhance pkcs12 utility by making the -nokeys and -nocerts 4648 options work when creating a PKCS#12 file. New option -nomac 4649 to omit the mac, NONE can be set for an encryption algorithm. 4650 New code is modified to use the enhanced PKCS12_create() 4651 instead of the low level API. 4652 [Steve Henson] 4653 4654 *) Extend ASN1 encoder to support indefinite length constructed 4655 encoding. This can output sequences tags and octet strings in 4656 this form. Modify pk7_asn1.c to support indefinite length 4657 encoding. This is experimental and needs additional code to 4658 be useful, such as an ASN1 bio and some enhanced streaming 4659 PKCS#7 code. 4660 4661 Extend template encode functionality so that tagging is passed 4662 down to the template encoder. 4663 [Steve Henson] 4664 4665 *) Let 'openssl req' fail if an argument to '-newkey' is not 4666 recognized instead of using RSA as a default. 4667 [Bodo Moeller] 4668 4669 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 4670 As these are not official, they are not included in "ALL"; 4671 the "ECCdraft" ciphersuite group alias can be used to select them. 4672 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 4673 4674 *) Add ECDH engine support. 4675 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 4676 4677 *) Add ECDH in new directory crypto/ecdh/. 4678 [Douglas Stebila (Sun Microsystems Laboratories)] 4679 4680 *) Let BN_rand_range() abort with an error after 100 iterations 4681 without success (which indicates a broken PRNG). 4682 [Bodo Moeller] 4683 4684 *) Change BN_mod_sqrt() so that it verifies that the input value 4685 is really the square of the return value. (Previously, 4686 BN_mod_sqrt would show GIGO behaviour.) 4687 [Bodo Moeller] 4688 4689 *) Add named elliptic curves over binary fields from X9.62, SECG, 4690 and WAP/WTLS; add OIDs that were still missing. 4691 4692 [Sheueling Chang Shantz and Douglas Stebila 4693 (Sun Microsystems Laboratories)] 4694 4695 *) Extend the EC library for elliptic curves over binary fields 4696 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 4697 New EC_METHOD: 4698 4699 EC_GF2m_simple_method 4700 4701 New API functions: 4702 4703 EC_GROUP_new_curve_GF2m 4704 EC_GROUP_set_curve_GF2m 4705 EC_GROUP_get_curve_GF2m 4706 EC_POINT_set_affine_coordinates_GF2m 4707 EC_POINT_get_affine_coordinates_GF2m 4708 EC_POINT_set_compressed_coordinates_GF2m 4709 4710 Point compression for binary fields is disabled by default for 4711 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 4712 enable it). 4713 4714 As binary polynomials are represented as BIGNUMs, various members 4715 of the EC_GROUP and EC_POINT data structures can be shared 4716 between the implementations for prime fields and binary fields; 4717 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 4718 are essentially identical to their ..._GFp counterparts. 4719 (For simplicity, the '..._GFp' prefix has been dropped from 4720 various internal method names.) 4721 4722 An internal 'field_div' method (similar to 'field_mul' and 4723 'field_sqr') has been added; this is used only for binary fields. 4724 4725 [Sheueling Chang Shantz and Douglas Stebila 4726 (Sun Microsystems Laboratories)] 4727 4728 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 4729 through methods ('mul', 'precompute_mult'). 4730 4731 The generic implementations (now internally called 'ec_wNAF_mul' 4732 and 'ec_wNAF_precomputed_mult') remain the default if these 4733 methods are undefined. 4734 4735 [Sheueling Chang Shantz and Douglas Stebila 4736 (Sun Microsystems Laboratories)] 4737 4738 *) New function EC_GROUP_get_degree, which is defined through 4739 EC_METHOD. For curves over prime fields, this returns the bit 4740 length of the modulus. 4741 4742 [Sheueling Chang Shantz and Douglas Stebila 4743 (Sun Microsystems Laboratories)] 4744 4745 *) New functions EC_GROUP_dup, EC_POINT_dup. 4746 (These simply call ..._new and ..._copy). 4747 4748 [Sheueling Chang Shantz and Douglas Stebila 4749 (Sun Microsystems Laboratories)] 4750 4751 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 4752 Polynomials are represented as BIGNUMs (where the sign bit is not 4753 used) in the following functions [macros]: 4754 4755 BN_GF2m_add 4756 BN_GF2m_sub [= BN_GF2m_add] 4757 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 4758 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 4759 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 4760 BN_GF2m_mod_inv 4761 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 4762 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 4763 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 4764 BN_GF2m_cmp [= BN_ucmp] 4765 4766 (Note that only the 'mod' functions are actually for fields GF(2^m). 4767 BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 4768 4769 For some functions, an the irreducible polynomial defining a 4770 field can be given as an 'unsigned int[]' with strictly 4771 decreasing elements giving the indices of those bits that are set; 4772 i.e., p[] represents the polynomial 4773 f(t) = t^p[0] + t^p[1] + ... + t^p[k] 4774 where 4775 p[0] > p[1] > ... > p[k] = 0. 4776 This applies to the following functions: 4777 4778 BN_GF2m_mod_arr 4779 BN_GF2m_mod_mul_arr 4780 BN_GF2m_mod_sqr_arr 4781 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 4782 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 4783 BN_GF2m_mod_exp_arr 4784 BN_GF2m_mod_sqrt_arr 4785 BN_GF2m_mod_solve_quad_arr 4786 BN_GF2m_poly2arr 4787 BN_GF2m_arr2poly 4788 4789 Conversion can be performed by the following functions: 4790 4791 BN_GF2m_poly2arr 4792 BN_GF2m_arr2poly 4793 4794 bntest.c has additional tests for binary polynomial arithmetic. 4795 4796 Two implementations for BN_GF2m_mod_div() are available. 4797 The default algorithm simply uses BN_GF2m_mod_inv() and 4798 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 4799 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 4800 copyright notice in crypto/bn/bn_gf2m.c before enabling it). 4801 4802 [Sheueling Chang Shantz and Douglas Stebila 4803 (Sun Microsystems Laboratories)] 4804 4805 *) Add new error code 'ERR_R_DISABLED' that can be used when some 4806 functionality is disabled at compile-time. 4807 [Douglas Stebila <douglas.stebila@sun.com>] 4808 4809 *) Change default behaviour of 'openssl asn1parse' so that more 4810 information is visible when viewing, e.g., a certificate: 4811 4812 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 4813 mode the content of non-printable OCTET STRINGs is output in a 4814 style similar to INTEGERs, but with '[HEX DUMP]' prepended to 4815 avoid the appearance of a printable string. 4816 [Nils Larsch <nla@trustcenter.de>] 4817 4818 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 4819 functions 4820 EC_GROUP_set_asn1_flag() 4821 EC_GROUP_get_asn1_flag() 4822 EC_GROUP_set_point_conversion_form() 4823 EC_GROUP_get_point_conversion_form() 4824 These control ASN1 encoding details: 4825 - Curves (i.e., groups) are encoded explicitly unless asn1_flag 4826 has been set to OPENSSL_EC_NAMED_CURVE. 4827 - Points are encoded in uncompressed form by default; options for 4828 asn1_for are as for point2oct, namely 4829 POINT_CONVERSION_COMPRESSED 4830 POINT_CONVERSION_UNCOMPRESSED 4831 POINT_CONVERSION_HYBRID 4832 4833 Also add 'seed' and 'seed_len' members to EC_GROUP with access 4834 functions 4835 EC_GROUP_set_seed() 4836 EC_GROUP_get0_seed() 4837 EC_GROUP_get_seed_len() 4838 This is used only for ASN1 purposes (so far). 4839 [Nils Larsch <nla@trustcenter.de>] 4840 4841 *) Add 'field_type' member to EC_METHOD, which holds the NID 4842 of the appropriate field type OID. The new function 4843 EC_METHOD_get_field_type() returns this value. 4844 [Nils Larsch <nla@trustcenter.de>] 4845 4846 *) Add functions 4847 EC_POINT_point2bn() 4848 EC_POINT_bn2point() 4849 EC_POINT_point2hex() 4850 EC_POINT_hex2point() 4851 providing useful interfaces to EC_POINT_point2oct() and 4852 EC_POINT_oct2point(). 4853 [Nils Larsch <nla@trustcenter.de>] 4854 4855 *) Change internals of the EC library so that the functions 4856 EC_GROUP_set_generator() 4857 EC_GROUP_get_generator() 4858 EC_GROUP_get_order() 4859 EC_GROUP_get_cofactor() 4860 are implemented directly in crypto/ec/ec_lib.c and not dispatched 4861 to methods, which would lead to unnecessary code duplication when 4862 adding different types of curves. 4863 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 4864 4865 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 4866 arithmetic, and such that modified wNAFs are generated 4867 (which avoid length expansion in many cases). 4868 [Bodo Moeller] 4869 4870 *) Add a function EC_GROUP_check_discriminant() (defined via 4871 EC_METHOD) that verifies that the curve discriminant is non-zero. 4872 4873 Add a function EC_GROUP_check() that makes some sanity tests 4874 on a EC_GROUP, its generator and order. This includes 4875 EC_GROUP_check_discriminant(). 4876 [Nils Larsch <nla@trustcenter.de>] 4877 4878 *) Add ECDSA in new directory crypto/ecdsa/. 4879 4880 Add applications 'openssl ecparam' and 'openssl ecdsa' 4881 (these are based on 'openssl dsaparam' and 'openssl dsa'). 4882 4883 ECDSA support is also included in various other files across the 4884 library. Most notably, 4885 - 'openssl req' now has a '-newkey ecdsa:file' option; 4886 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 4887 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 4888 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 4889 them suitable for ECDSA where domain parameters must be 4890 extracted before the specific public key; 4891 - ECDSA engine support has been added. 4892 [Nils Larsch <nla@trustcenter.de>] 4893 4894 *) Include some named elliptic curves, and add OIDs from X9.62, 4895 SECG, and WAP/WTLS. Each curve can be obtained from the new 4896 function 4897 EC_GROUP_new_by_curve_name(), 4898 and the list of available named curves can be obtained with 4899 EC_get_builtin_curves(). 4900 Also add a 'curve_name' member to EC_GROUP objects, which can be 4901 accessed via 4902 EC_GROUP_set_curve_name() 4903 EC_GROUP_get_curve_name() 4904 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 4905 4906 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 4907 was actually never needed) and in BN_mul(). The removal in BN_mul() 4908 required a small change in bn_mul_part_recursive() and the addition 4909 of the functions bn_cmp_part_words(), bn_sub_part_words() and 4910 bn_add_part_words(), which do the same thing as bn_cmp_words(), 4911 bn_sub_words() and bn_add_words() except they take arrays with 4912 differing sizes. 4913 [Richard Levitte] 4914 4915 Changes between 0.9.7l and 0.9.7m [23 Feb 2007] 4916 4917 *) Cleanse PEM buffers before freeing them since they may contain 4918 sensitive data. 4919 [Benjamin Bennett <ben@psc.edu>] 4920 4921 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 4922 a ciphersuite string such as "DEFAULT:RSA" cannot enable 4923 authentication-only ciphersuites. 4924 [Bodo Moeller] 4925 4926 *) Since AES128 and AES256 share a single mask bit in the logic of 4927 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 4928 kludge to work properly if AES128 is available and AES256 isn't. 4929 [Victor Duchovni] 4930 4931 *) Expand security boundary to match 1.1.1 module. 4932 [Steve Henson] 4933 4934 *) Remove redundant features: hash file source, editing of test vectors 4935 modify fipsld to use external fips_premain.c signature. 4936 [Steve Henson] 4937 4938 *) New perl script mkfipsscr.pl to create shell scripts or batch files to 4939 run algorithm test programs. 4940 [Steve Henson] 4941 4942 *) Make algorithm test programs more tolerant of whitespace. 4943 [Steve Henson] 4944 4945 *) Have SSL/TLS server implementation tolerate "mismatched" record 4946 protocol version while receiving ClientHello even if the 4947 ClientHello is fragmented. (The server can't insist on the 4948 particular protocol version it has chosen before the ServerHello 4949 message has informed the client about his choice.) 4950 [Bodo Moeller] 4951 4952 *) Load error codes if they are not already present instead of using a 4953 static variable. This allows them to be cleanly unloaded and reloaded. 4954 [Steve Henson] 4955 4956 Changes between 0.9.7k and 0.9.7l [28 Sep 2006] 4957 4958 *) Introduce limits to prevent malicious keys being able to 4959 cause a denial of service. (CVE-2006-2940) 4960 [Steve Henson, Bodo Moeller] 4961 4962 *) Fix ASN.1 parsing of certain invalid structures that can result 4963 in a denial of service. (CVE-2006-2937) [Steve Henson] 4964 4965 *) Fix buffer overflow in SSL_get_shared_ciphers() function. 4966 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 4967 4968 *) Fix SSL client code which could crash if connecting to a 4969 malicious SSLv2 server. (CVE-2006-4343) 4970 [Tavis Ormandy and Will Drewry, Google Security Team] 4971 4972 *) Change ciphersuite string processing so that an explicit 4973 ciphersuite selects this one ciphersuite (so that "AES256-SHA" 4974 will no longer include "AES128-SHA"), and any other similar 4975 ciphersuite (same bitmap) from *other* protocol versions (so that 4976 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 4977 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 4978 changes from 0.9.8b and 0.9.8d. 4979 [Bodo Moeller] 4980 4981 Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 4982 4983 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 4984 (CVE-2006-4339) [Ben Laurie and Google Security Team] 4985 4986 *) Change the Unix randomness entropy gathering to use poll() when 4987 possible instead of select(), since the latter has some 4988 undesirable limitations. 4989 [Darryl Miles via Richard Levitte and Bodo Moeller] 4990 4991 *) Disable rogue ciphersuites: 4992 4993 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 4994 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 4995 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 4996 4997 The latter two were purportedly from 4998 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 4999 appear there. 5000 5001 Also deactive the remaining ciphersuites from 5002 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 5003 unofficial, and the ID has long expired. 5004 [Bodo Moeller] 5005 5006 *) Fix RSA blinding Heisenbug (problems sometimes occured on 5007 dual-core machines) and other potential thread-safety issues. 5008 [Bodo Moeller] 5009 5010 Changes between 0.9.7i and 0.9.7j [04 May 2006] 5011 5012 *) Adapt fipsld and the build system to link against the validated FIPS 5013 module in FIPS mode. 5014 [Steve Henson] 5015 5016 *) Fixes for VC++ 2005 build under Windows. 5017 [Steve Henson] 5018 5019 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 5020 from a Windows bash shell such as MSYS. It is autodetected from the 5021 "config" script when run from a VC++ environment. Modify standard VC++ 5022 build to use fipscanister.o from the GNU make build. 5023 [Steve Henson] 5024 5025 Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 5026 5027 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 5028 The value now differs depending on if you build for FIPS or not. 5029 BEWARE! A program linked with a shared FIPSed libcrypto can't be 5030 safely run with a non-FIPSed libcrypto, as it may crash because of 5031 the difference induced by this change. 5032 [Andy Polyakov] 5033 5034 Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 5035 5036 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 5037 (part of SSL_OP_ALL). This option used to disable the 5038 countermeasure against man-in-the-middle protocol-version 5039 rollback in the SSL 2.0 server implementation, which is a bad 5040 idea. (CVE-2005-2969) 5041 5042 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 5043 for Information Security, National Institute of Advanced Industrial 5044 Science and Technology [AIST], Japan)] 5045 5046 *) Minimal support for X9.31 signatures and PSS padding modes. This is 5047 mainly for FIPS compliance and not fully integrated at this stage. 5048 [Steve Henson] 5049 5050 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 5051 the exponentiation using a fixed-length exponent. (Otherwise, 5052 the information leaked through timing could expose the secret key 5053 after many signatures; cf. Bleichenbacher's attack on DSA with 5054 biased k.) 5055 [Bodo Moeller] 5056 5057 *) Make a new fixed-window mod_exp implementation the default for 5058 RSA, DSA, and DH private-key operations so that the sequence of 5059 squares and multiplies and the memory access pattern are 5060 independent of the particular secret key. This will mitigate 5061 cache-timing and potential related attacks. 5062 5063 BN_mod_exp_mont_consttime() is the new exponentiation implementation, 5064 and this is automatically used by BN_mod_exp_mont() if the new flag 5065 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 5066 will use this BN flag for private exponents unless the flag 5067 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 5068 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 5069 5070 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 5071 5072 *) Change the client implementation for SSLv23_method() and 5073 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 5074 Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 5075 (Previously, the SSL 2.0 backwards compatible Client Hello 5076 message format would be used even with SSL_OP_NO_SSLv2.) 5077 [Bodo Moeller] 5078 5079 *) Add support for smime-type MIME parameter in S/MIME messages which some 5080 clients need. 5081 [Steve Henson] 5082 5083 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 5084 a threadsafe manner. Modify rsa code to use new function and add calls 5085 to dsa and dh code (which had race conditions before). 5086 [Steve Henson] 5087 5088 *) Include the fixed error library code in the C error file definitions 5089 instead of fixing them up at runtime. This keeps the error code 5090 structures constant. 5091 [Steve Henson] 5092 5093 Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 5094 5095 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 5096 OpenSSL 0.9.8.] 5097 5098 *) Fixes for newer kerberos headers. NB: the casts are needed because 5099 the 'length' field is signed on one version and unsigned on another 5100 with no (?) obvious way to tell the difference, without these VC++ 5101 complains. Also the "definition" of FAR (blank) is no longer included 5102 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 5103 some needed definitions. 5104 [Steve Henson] 5105 5106 *) Undo Cygwin change. 5107 [Ulf M��ller] 5108 5109 *) Added support for proxy certificates according to RFC 3820. 5110 Because they may be a security thread to unaware applications, 5111 they must be explicitely allowed in run-time. See 5112 docs/HOWTO/proxy_certificates.txt for further information. 5113 [Richard Levitte] 5114 5115 Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 5116 5117 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 5118 server and client random values. Previously 5119 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 5120 less random data when sizeof(time_t) > 4 (some 64 bit platforms). 5121 5122 This change has negligible security impact because: 5123 5124 1. Server and client random values still have 24 bytes of pseudo random 5125 data. 5126 5127 2. Server and client random values are sent in the clear in the initial 5128 handshake. 5129 5130 3. The master secret is derived using the premaster secret (48 bytes in 5131 size for static RSA ciphersuites) as well as client server and random 5132 values. 5133 5134 The OpenSSL team would like to thank the UK NISCC for bringing this issue 5135 to our attention. 5136 5137 [Stephen Henson, reported by UK NISCC] 5138 5139 *) Use Windows randomness collection on Cygwin. 5140 [Ulf M��ller] 5141 5142 *) Fix hang in EGD/PRNGD query when communication socket is closed 5143 prematurely by EGD/PRNGD. 5144 [Darren Tucker <dtucker@zip.com.au> via Lutz J��nicke, resolves #1014] 5145 5146 *) Prompt for pass phrases when appropriate for PKCS12 input format. 5147 [Steve Henson] 5148 5149 *) Back-port of selected performance improvements from development 5150 branch, as well as improved support for PowerPC platforms. 5151 [Andy Polyakov] 5152 5153 *) Add lots of checks for memory allocation failure, error codes to indicate 5154 failure and freeing up memory if a failure occurs. 5155 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 5156 5157 *) Add new -passin argument to dgst. 5158 [Steve Henson] 5159 5160 *) Perform some character comparisons of different types in X509_NAME_cmp: 5161 this is needed for some certificates that reencode DNs into UTF8Strings 5162 (in violation of RFC3280) and can't or wont issue name rollover 5163 certificates. 5164 [Steve Henson] 5165 5166 *) Make an explicit check during certificate validation to see that 5167 the CA setting in each certificate on the chain is correct. As a 5168 side effect always do the following basic checks on extensions, 5169 not just when there's an associated purpose to the check: 5170 5171 - if there is an unhandled critical extension (unless the user 5172 has chosen to ignore this fault) 5173 - if the path length has been exceeded (if one is set at all) 5174 - that certain extensions fit the associated purpose (if one has 5175 been given) 5176 [Richard Levitte] 5177 5178 Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 5179 5180 *) Avoid a race condition when CRLs are checked in a multi threaded 5181 environment. This would happen due to the reordering of the revoked 5182 entries during signature checking and serial number lookup. Now the 5183 encoding is cached and the serial number sort performed under a lock. 5184 Add new STACK function sk_is_sorted(). 5185 [Steve Henson] 5186 5187 *) Add Delta CRL to the extension code. 5188 [Steve Henson] 5189 5190 *) Various fixes to s3_pkt.c so alerts are sent properly. 5191 [David Holmes <d.holmes@f5.com>] 5192 5193 *) Reduce the chances of duplicate issuer name and serial numbers (in 5194 violation of RFC3280) using the OpenSSL certificate creation utilities. 5195 This is done by creating a random 64 bit value for the initial serial 5196 number when a serial number file is created or when a self signed 5197 certificate is created using 'openssl req -x509'. The initial serial 5198 number file is created using 'openssl x509 -next_serial' in CA.pl 5199 rather than being initialized to 1. 5200 [Steve Henson] 5201 5202 Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 5203 5204 *) Fix null-pointer assignment in do_change_cipher_spec() revealed 5205 by using the Codenomicon TLS Test Tool (CVE-2004-0079) 5206 [Joe Orton, Steve Henson] 5207 5208 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 5209 (CVE-2004-0112) 5210 [Joe Orton, Steve Henson] 5211 5212 *) Make it possible to have multiple active certificates with the same 5213 subject in the CA index file. This is done only if the keyword 5214 'unique_subject' is set to 'no' in the main CA section (default 5215 if 'CA_default') of the configuration file. The value is saved 5216 with the database itself in a separate index attribute file, 5217 named like the index file with '.attr' appended to the name. 5218 [Richard Levitte] 5219 5220 *) X509 verify fixes. Disable broken certificate workarounds when 5221 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 5222 keyUsage extension present. Don't accept CRLs with unhandled critical 5223 extensions: since verify currently doesn't process CRL extensions this 5224 rejects a CRL with *any* critical extensions. Add new verify error codes 5225 for these cases. 5226 [Steve Henson] 5227 5228 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 5229 A clarification of RFC2560 will require the use of OCTET STRINGs and 5230 some implementations cannot handle the current raw format. Since OpenSSL 5231 copies and compares OCSP nonces as opaque blobs without any attempt at 5232 parsing them this should not create any compatibility issues. 5233 [Steve Henson] 5234 5235 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 5236 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 5237 this HMAC (and other) operations are several times slower than OpenSSL 5238 < 0.9.7. 5239 [Steve Henson] 5240 5241 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 5242 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 5243 5244 *) Use the correct content when signing type "other". 5245 [Steve Henson] 5246 5247 Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 5248 5249 *) Fix various bugs revealed by running the NISCC test suite: 5250 5251 Stop out of bounds reads in the ASN1 code when presented with 5252 invalid tags (CVE-2003-0543 and CVE-2003-0544). 5253 5254 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 5255 5256 If verify callback ignores invalid public key errors don't try to check 5257 certificate signature with the NULL public key. 5258 5259 [Steve Henson] 5260 5261 *) New -ignore_err option in ocsp application to stop the server 5262 exiting on the first error in a request. 5263 [Steve Henson] 5264 5265 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 5266 if the server requested one: as stated in TLS 1.0 and SSL 3.0 5267 specifications. 5268 [Steve Henson] 5269 5270 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 5271 extra data after the compression methods not only for TLS 1.0 5272 but also for SSL 3.0 (as required by the specification). 5273 [Bodo Moeller; problem pointed out by Matthias Loepfe] 5274 5275 *) Change X509_certificate_type() to mark the key as exported/exportable 5276 when it's 512 *bits* long, not 512 bytes. 5277 [Richard Levitte] 5278 5279 *) Change AES_cbc_encrypt() so it outputs exact multiple of 5280 blocks during encryption. 5281 [Richard Levitte] 5282 5283 *) Various fixes to base64 BIO and non blocking I/O. On write 5284 flushes were not handled properly if the BIO retried. On read 5285 data was not being buffered properly and had various logic bugs. 5286 This also affects blocking I/O when the data being decoded is a 5287 certain size. 5288 [Steve Henson] 5289 5290 *) Various S/MIME bugfixes and compatibility changes: 5291 output correct application/pkcs7 MIME type if 5292 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 5293 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 5294 of files as .eml work). Correctly handle very long lines in MIME 5295 parser. 5296 [Steve Henson] 5297 5298 Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 5299 5300 *) Countermeasure against the Klima-Pokorny-Rosa extension of 5301 Bleichbacher's attack on PKCS #1 v1.5 padding: treat 5302 a protocol version number mismatch like a decryption error 5303 in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 5304 [Bodo Moeller] 5305 5306 *) Turn on RSA blinding by default in the default implementation 5307 to avoid a timing attack. Applications that don't want it can call 5308 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 5309 They would be ill-advised to do so in most cases. 5310 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 5311 5312 *) Change RSA blinding code so that it works when the PRNG is not 5313 seeded (in this case, the secret RSA exponent is abused as 5314 an unpredictable seed -- if it is not unpredictable, there 5315 is no point in blinding anyway). Make RSA blinding thread-safe 5316 by remembering the creator's thread ID in rsa->blinding and 5317 having all other threads use local one-time blinding factors 5318 (this requires more computation than sharing rsa->blinding, but 5319 avoids excessive locking; and if an RSA object is not shared 5320 between threads, blinding will still be very fast). 5321 [Bodo Moeller] 5322 5323 *) Fixed a typo bug that would cause ENGINE_set_default() to set an 5324 ENGINE as defaults for all supported algorithms irrespective of 5325 the 'flags' parameter. 'flags' is now honoured, so applications 5326 should make sure they are passing it correctly. 5327 [Geoff Thorpe] 5328 5329 *) Target "mingw" now allows native Windows code to be generated in 5330 the Cygwin environment as well as with the MinGW compiler. 5331 [Ulf Moeller] 5332 5333 Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 5334 5335 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 5336 via timing by performing a MAC computation even if incorrrect 5337 block cipher padding has been found. This is a countermeasure 5338 against active attacks where the attacker has to distinguish 5339 between bad padding and a MAC verification error. (CVE-2003-0078) 5340 5341 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 5342 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 5343 Martin Vuagnoux (EPFL, Ilion)] 5344 5345 *) Make the no-err option work as intended. The intention with no-err 5346 is not to have the whole error stack handling routines removed from 5347 libcrypto, it's only intended to remove all the function name and 5348 reason texts, thereby removing some of the footprint that may not 5349 be interesting if those errors aren't displayed anyway. 5350 5351 NOTE: it's still possible for any application or module to have it's 5352 own set of error texts inserted. The routines are there, just not 5353 used by default when no-err is given. 5354 [Richard Levitte] 5355 5356 *) Add support for FreeBSD on IA64. 5357 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 5358 5359 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 5360 Kerberos function mit_des_cbc_cksum(). Before this change, 5361 the value returned by DES_cbc_cksum() was like the one from 5362 mit_des_cbc_cksum(), except the bytes were swapped. 5363 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 5364 5365 *) Allow an application to disable the automatic SSL chain building. 5366 Before this a rather primitive chain build was always performed in 5367 ssl3_output_cert_chain(): an application had no way to send the 5368 correct chain if the automatic operation produced an incorrect result. 5369 5370 Now the chain builder is disabled if either: 5371 5372 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 5373 5374 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 5375 5376 The reasoning behind this is that an application would not want the 5377 auto chain building to take place if extra chain certificates are 5378 present and it might also want a means of sending no additional 5379 certificates (for example the chain has two certificates and the 5380 root is omitted). 5381 [Steve Henson] 5382 5383 *) Add the possibility to build without the ENGINE framework. 5384 [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 5385 5386 *) Under Win32 gmtime() can return NULL: check return value in 5387 OPENSSL_gmtime(). Add error code for case where gmtime() fails. 5388 [Steve Henson] 5389 5390 *) DSA routines: under certain error conditions uninitialized BN objects 5391 could be freed. Solution: make sure initialization is performed early 5392 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 5393 Nils Larsch <nla@trustcenter.de> via PR#459) 5394 [Lutz Jaenicke] 5395 5396 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 5397 checked on reconnect on the client side, therefore session resumption 5398 could still fail with a "ssl session id is different" error. This 5399 behaviour is masked when SSL_OP_ALL is used due to 5400 SSL_OP_MICROSOFT_SESS_ID_BUG being set. 5401 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 5402 followup to PR #377. 5403 [Lutz Jaenicke] 5404 5405 *) IA-32 assembler support enhancements: unified ELF targets, support 5406 for SCO/Caldera platforms, fix for Cygwin shared build. 5407 [Andy Polyakov] 5408 5409 *) Add support for FreeBSD on sparc64. As a consequence, support for 5410 FreeBSD on non-x86 processors is separate from x86 processors on 5411 the config script, much like the NetBSD support. 5412 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 5413 5414 Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 5415 5416 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 5417 OpenSSL 0.9.7.] 5418 5419 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 5420 code (06) was taken as the first octet of the session ID and the last 5421 octet was ignored consequently. As a result SSLv2 client side session 5422 caching could not have worked due to the session ID mismatch between 5423 client and server. 5424 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 5425 PR #377. 5426 [Lutz Jaenicke] 5427 5428 *) Change the declaration of needed Kerberos libraries to use EX_LIBS 5429 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 5430 removed entirely. 5431 [Richard Levitte] 5432 5433 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 5434 seems that in spite of existing for more than a year, many application 5435 author have done nothing to provide the necessary callbacks, which 5436 means that this particular engine will not work properly anywhere. 5437 This is a very unfortunate situation which forces us, in the name 5438 of usability, to give the hw_ncipher.c a static lock, which is part 5439 of libcrypto. 5440 NOTE: This is for the 0.9.7 series ONLY. This hack will never 5441 appear in 0.9.8 or later. We EXPECT application authors to have 5442 dealt properly with this when 0.9.8 is released (unless we actually 5443 make such changes in the libcrypto locking code that changes will 5444 have to be made anyway). 5445 [Richard Levitte] 5446 5447 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 5448 octets have been read, EOF or an error occurs. Without this change 5449 some truncated ASN1 structures will not produce an error. 5450 [Steve Henson] 5451 5452 *) Disable Heimdal support, since it hasn't been fully implemented. 5453 Still give the possibility to force the use of Heimdal, but with 5454 warnings and a request that patches get sent to openssl-dev. 5455 [Richard Levitte] 5456 5457 *) Add the VC-CE target, introduce the WINCE sysname, and add 5458 INSTALL.WCE and appropriate conditionals to make it build. 5459 [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 5460 5461 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 5462 cygssl-x.y.z.dll, where x, y and z are the major, minor and 5463 edit numbers of the version. 5464 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 5465 5466 *) Introduce safe string copy and catenation functions 5467 (BUF_strlcpy() and BUF_strlcat()). 5468 [Ben Laurie (CHATS) and Richard Levitte] 5469 5470 *) Avoid using fixed-size buffers for one-line DNs. 5471 [Ben Laurie (CHATS)] 5472 5473 *) Add BUF_MEM_grow_clean() to avoid information leakage when 5474 resizing buffers containing secrets, and use where appropriate. 5475 [Ben Laurie (CHATS)] 5476 5477 *) Avoid using fixed size buffers for configuration file location. 5478 [Ben Laurie (CHATS)] 5479 5480 *) Avoid filename truncation for various CA files. 5481 [Ben Laurie (CHATS)] 5482 5483 *) Use sizeof in preference to magic numbers. 5484 [Ben Laurie (CHATS)] 5485 5486 *) Avoid filename truncation in cert requests. 5487 [Ben Laurie (CHATS)] 5488 5489 *) Add assertions to check for (supposedly impossible) buffer 5490 overflows. 5491 [Ben Laurie (CHATS)] 5492 5493 *) Don't cache truncated DNS entries in the local cache (this could 5494 potentially lead to a spoofing attack). 5495 [Ben Laurie (CHATS)] 5496 5497 *) Fix various buffers to be large enough for hex/decimal 5498 representations in a platform independent manner. 5499 [Ben Laurie (CHATS)] 5500 5501 *) Add CRYPTO_realloc_clean() to avoid information leakage when 5502 resizing buffers containing secrets, and use where appropriate. 5503 [Ben Laurie (CHATS)] 5504 5505 *) Add BIO_indent() to avoid much slightly worrying code to do 5506 indents. 5507 [Ben Laurie (CHATS)] 5508 5509 *) Convert sprintf()/BIO_puts() to BIO_printf(). 5510 [Ben Laurie (CHATS)] 5511 5512 *) buffer_gets() could terminate with the buffer only half 5513 full. Fixed. 5514 [Ben Laurie (CHATS)] 5515 5516 *) Add assertions to prevent user-supplied crypto functions from 5517 overflowing internal buffers by having large block sizes, etc. 5518 [Ben Laurie (CHATS)] 5519 5520 *) New OPENSSL_assert() macro (similar to assert(), but enabled 5521 unconditionally). 5522 [Ben Laurie (CHATS)] 5523 5524 *) Eliminate unused copy of key in RC4. 5525 [Ben Laurie (CHATS)] 5526 5527 *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 5528 [Ben Laurie (CHATS)] 5529 5530 *) Fix off-by-one error in EGD path. 5531 [Ben Laurie (CHATS)] 5532 5533 *) If RANDFILE path is too long, ignore instead of truncating. 5534 [Ben Laurie (CHATS)] 5535 5536 *) Eliminate unused and incorrectly sized X.509 structure 5537 CBCParameter. 5538 [Ben Laurie (CHATS)] 5539 5540 *) Eliminate unused and dangerous function knumber(). 5541 [Ben Laurie (CHATS)] 5542 5543 *) Eliminate unused and dangerous structure, KSSL_ERR. 5544 [Ben Laurie (CHATS)] 5545 5546 *) Protect against overlong session ID context length in an encoded 5547 session object. Since these are local, this does not appear to be 5548 exploitable. 5549 [Ben Laurie (CHATS)] 5550 5551 *) Change from security patch (see 0.9.6e below) that did not affect 5552 the 0.9.6 release series: 5553 5554 Remote buffer overflow in SSL3 protocol - an attacker could 5555 supply an oversized master key in Kerberos-enabled versions. 5556 (CVE-2002-0657) 5557 [Ben Laurie (CHATS)] 5558 5559 *) Change the SSL kerb5 codes to match RFC 2712. 5560 [Richard Levitte] 5561 5562 *) Make -nameopt work fully for req and add -reqopt switch. 5563 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 5564 5565 *) The "block size" for block ciphers in CFB and OFB mode should be 1. 5566 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 5567 5568 *) Make sure tests can be performed even if the corresponding algorithms 5569 have been removed entirely. This was also the last step to make 5570 OpenSSL compilable with DJGPP under all reasonable conditions. 5571 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 5572 5573 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 5574 to allow version independent disabling of normally unselected ciphers, 5575 which may be activated as a side-effect of selecting a single cipher. 5576 5577 (E.g., cipher list string "RSA" enables ciphersuites that are left 5578 out of "ALL" because they do not provide symmetric encryption. 5579 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 5580 [Lutz Jaenicke, Bodo Moeller] 5581 5582 *) Add appropriate support for separate platform-dependent build 5583 directories. The recommended way to make a platform-dependent 5584 build directory is the following (tested on Linux), maybe with 5585 some local tweaks: 5586 5587 # Place yourself outside of the OpenSSL source tree. In 5588 # this example, the environment variable OPENSSL_SOURCE 5589 # is assumed to contain the absolute OpenSSL source directory. 5590 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 5591 cd objtree/"`uname -s`-`uname -r`-`uname -m`" 5592 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 5593 mkdir -p `dirname $F` 5594 ln -s $OPENSSL_SOURCE/$F $F 5595 done 5596 5597 To be absolutely sure not to disturb the source tree, a "make clean" 5598 is a good thing. If it isn't successfull, don't worry about it, 5599 it probably means the source directory is very clean. 5600 [Richard Levitte] 5601 5602 *) Make sure any ENGINE control commands make local copies of string 5603 pointers passed to them whenever necessary. Otherwise it is possible 5604 the caller may have overwritten (or deallocated) the original string 5605 data when a later ENGINE operation tries to use the stored values. 5606 [G��tz Babin-Ebell <babinebell@trustcenter.de>] 5607 5608 *) Improve diagnostics in file reading and command-line digests. 5609 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 5610 5611 *) Add AES modes CFB and OFB to the object database. Correct an 5612 error in AES-CFB decryption. 5613 [Richard Levitte] 5614 5615 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 5616 allows existing EVP_CIPHER_CTX structures to be reused after 5617 calling EVP_*Final(). This behaviour is used by encryption 5618 BIOs and some applications. This has the side effect that 5619 applications must explicitly clean up cipher contexts with 5620 EVP_CIPHER_CTX_cleanup() or they will leak memory. 5621 [Steve Henson] 5622 5623 *) Check the values of dna and dnb in bn_mul_recursive before calling 5624 bn_mul_comba (a non zero value means the a or b arrays do not contain 5625 n2 elements) and fallback to bn_mul_normal if either is not zero. 5626 [Steve Henson] 5627 5628 *) Fix escaping of non-ASCII characters when using the -subj option 5629 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 5630 [Lutz Jaenicke] 5631 5632 *) Make object definitions compliant to LDAP (RFC2256): SN is the short 5633 form for "surname", serialNumber has no short form. 5634 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 5635 therefore remove "mail" short name for "internet 7". 5636 The OID for unique identifiers in X509 certificates is 5637 x500UniqueIdentifier, not uniqueIdentifier. 5638 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 5639 [Lutz Jaenicke] 5640 5641 *) Add an "init" command to the ENGINE config module and auto initialize 5642 ENGINEs. Without any "init" command the ENGINE will be initialized 5643 after all ctrl commands have been executed on it. If init=1 the 5644 ENGINE is initailized at that point (ctrls before that point are run 5645 on the uninitialized ENGINE and after on the initialized one). If 5646 init=0 then the ENGINE will not be iniatialized at all. 5647 [Steve Henson] 5648 5649 *) Fix the 'app_verify_callback' interface so that the user-defined 5650 argument is actually passed to the callback: In the 5651 SSL_CTX_set_cert_verify_callback() prototype, the callback 5652 declaration has been changed from 5653 int (*cb)() 5654 into 5655 int (*cb)(X509_STORE_CTX *,void *); 5656 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 5657 i=s->ctx->app_verify_callback(&ctx) 5658 has been changed into 5659 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 5660 5661 To update applications using SSL_CTX_set_cert_verify_callback(), 5662 a dummy argument can be added to their callback functions. 5663 [D. K. Smetters <smetters@parc.xerox.com>] 5664 5665 *) Added the '4758cca' ENGINE to support IBM 4758 cards. 5666 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 5667 5668 *) Add and OPENSSL_LOAD_CONF define which will cause 5669 OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 5670 This allows older applications to transparently support certain 5671 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 5672 Two new functions OPENSSL_add_all_algorithms_noconf() which will never 5673 load the config file and OPENSSL_add_all_algorithms_conf() which will 5674 always load it have also been added. 5675 [Steve Henson] 5676 5677 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 5678 Adjust NIDs and EVP layer. 5679 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 5680 5681 *) Config modules support in openssl utility. 5682 5683 Most commands now load modules from the config file, 5684 though in a few (such as version) this isn't done 5685 because it couldn't be used for anything. 5686 5687 In the case of ca and req the config file used is 5688 the same as the utility itself: that is the -config 5689 command line option can be used to specify an 5690 alternative file. 5691 [Steve Henson] 5692 5693 *) Move default behaviour from OPENSSL_config(). If appname is NULL 5694 use "openssl_conf" if filename is NULL use default openssl config file. 5695 [Steve Henson] 5696 5697 *) Add an argument to OPENSSL_config() to allow the use of an alternative 5698 config section name. Add a new flag to tolerate a missing config file 5699 and move code to CONF_modules_load_file(). 5700 [Steve Henson] 5701 5702 *) Support for crypto accelerator cards from Accelerated Encryption 5703 Processing, www.aep.ie. (Use engine 'aep') 5704 The support was copied from 0.9.6c [engine] and adapted/corrected 5705 to work with the new engine framework. 5706 [AEP Inc. and Richard Levitte] 5707 5708 *) Support for SureWare crypto accelerator cards from Baltimore 5709 Technologies. (Use engine 'sureware') 5710 The support was copied from 0.9.6c [engine] and adapted 5711 to work with the new engine framework. 5712 [Richard Levitte] 5713 5714 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 5715 make the newer ENGINE framework commands for the CHIL engine work. 5716 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 5717 5718 *) Make it possible to produce shared libraries on ReliantUNIX. 5719 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 5720 5721 *) Add the configuration target debug-linux-ppro. 5722 Make 'openssl rsa' use the general key loading routines 5723 implemented in apps.c, and make those routines able to 5724 handle the key format FORMAT_NETSCAPE and the variant 5725 FORMAT_IISSGC. 5726 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 5727 5728 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 5729 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 5730 5731 *) Add -keyform to rsautl, and document -engine. 5732 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 5733 5734 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 5735 BIO_R_NO_SUCH_FILE error code rather than the generic 5736 ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 5737 [Ben Laurie] 5738 5739 *) Add new functions 5740 ERR_peek_last_error 5741 ERR_peek_last_error_line 5742 ERR_peek_last_error_line_data. 5743 These are similar to 5744 ERR_peek_error 5745 ERR_peek_error_line 5746 ERR_peek_error_line_data, 5747 but report on the latest error recorded rather than the first one 5748 still in the error queue. 5749 [Ben Laurie, Bodo Moeller] 5750 5751 *) default_algorithms option in ENGINE config module. This allows things 5752 like: 5753 default_algorithms = ALL 5754 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 5755 [Steve Henson] 5756 5757 *) Prelminary ENGINE config module. 5758 [Steve Henson] 5759 5760 *) New experimental application configuration code. 5761 [Steve Henson] 5762 5763 *) Change the AES code to follow the same name structure as all other 5764 symmetric ciphers, and behave the same way. Move everything to 5765 the directory crypto/aes, thereby obsoleting crypto/rijndael. 5766 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 5767 5768 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 5769 [Ben Laurie and Theo de Raadt] 5770 5771 *) Add option to output public keys in req command. 5772 [Massimiliano Pala madwolf@openca.org] 5773 5774 *) Use wNAFs in EC_POINTs_mul() for improved efficiency 5775 (up to about 10% better than before for P-192 and P-224). 5776 [Bodo Moeller] 5777 5778 *) New functions/macros 5779 5780 SSL_CTX_set_msg_callback(ctx, cb) 5781 SSL_CTX_set_msg_callback_arg(ctx, arg) 5782 SSL_set_msg_callback(ssl, cb) 5783 SSL_set_msg_callback_arg(ssl, arg) 5784 5785 to request calling a callback function 5786 5787 void cb(int write_p, int version, int content_type, 5788 const void *buf, size_t len, SSL *ssl, void *arg) 5789 5790 whenever a protocol message has been completely received 5791 (write_p == 0) or sent (write_p == 1). Here 'version' is the 5792 protocol version according to which the SSL library interprets 5793 the current protocol message (SSL2_VERSION, SSL3_VERSION, or 5794 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 5795 the content type as defined in the SSL 3.0/TLS 1.0 protocol 5796 specification (change_cipher_spec(20), alert(21), handshake(22)). 5797 'buf' and 'len' point to the actual message, 'ssl' to the 5798 SSL object, and 'arg' is the application-defined value set by 5799 SSL[_CTX]_set_msg_callback_arg(). 5800 5801 'openssl s_client' and 'openssl s_server' have new '-msg' options 5802 to enable a callback that displays all protocol messages. 5803 [Bodo Moeller] 5804 5805 *) Change the shared library support so shared libraries are built as 5806 soon as the corresponding static library is finished, and thereby get 5807 openssl and the test programs linked against the shared library. 5808 This still only happens when the keyword "shard" has been given to 5809 the configuration scripts. 5810 5811 NOTE: shared library support is still an experimental thing, and 5812 backward binary compatibility is still not guaranteed. 5813 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 5814 5815 *) Add support for Subject Information Access extension. 5816 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 5817 5818 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 5819 additional bytes when new memory had to be allocated, not just 5820 when reusing an existing buffer. 5821 [Bodo Moeller] 5822 5823 *) New command line and configuration option 'utf8' for the req command. 5824 This allows field values to be specified as UTF8 strings. 5825 [Steve Henson] 5826 5827 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 5828 runs for the former and machine-readable output for the latter. 5829 [Ben Laurie] 5830 5831 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 5832 of the e-mail address in the DN (i.e., it will go into a certificate 5833 extension only). The new configuration file option 'email_in_dn = no' 5834 has the same effect. 5835 [Massimiliano Pala madwolf@openca.org] 5836 5837 *) Change all functions with names starting with des_ to be starting 5838 with DES_ instead. Add wrappers that are compatible with libdes, 5839 but are named _ossl_old_des_*. Finally, add macros that map the 5840 des_* symbols to the corresponding _ossl_old_des_* if libdes 5841 compatibility is desired. If OpenSSL 0.9.6c compatibility is 5842 desired, the des_* symbols will be mapped to DES_*, with one 5843 exception. 5844 5845 Since we provide two compatibility mappings, the user needs to 5846 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 5847 compatibility is desired. The default (i.e., when that macro 5848 isn't defined) is OpenSSL 0.9.6c compatibility. 5849 5850 There are also macros that enable and disable the support of old 5851 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 5852 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 5853 are defined, the default will apply: to support the old des routines. 5854 5855 In either case, one must include openssl/des.h to get the correct 5856 definitions. Do not try to just include openssl/des_old.h, that 5857 won't work. 5858 5859 NOTE: This is a major break of an old API into a new one. Software 5860 authors are encouraged to switch to the DES_ style functions. Some 5861 time in the future, des_old.h and the libdes compatibility functions 5862 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 5863 default), and then completely removed. 5864 [Richard Levitte] 5865 5866 *) Test for certificates which contain unsupported critical extensions. 5867 If such a certificate is found during a verify operation it is 5868 rejected by default: this behaviour can be overridden by either 5869 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 5870 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 5871 X509_supported_extension() has also been added which returns 1 if a 5872 particular extension is supported. 5873 [Steve Henson] 5874 5875 *) Modify the behaviour of EVP cipher functions in similar way to digests 5876 to retain compatibility with existing code. 5877 [Steve Henson] 5878 5879 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 5880 compatibility with existing code. In particular the 'ctx' parameter does 5881 not have to be to be initialized before the call to EVP_DigestInit() and 5882 it is tidied up after a call to EVP_DigestFinal(). New function 5883 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 5884 EVP_MD_CTX_copy() changed to not require the destination to be 5885 initialized valid and new function EVP_MD_CTX_copy_ex() added which 5886 requires the destination to be valid. 5887 5888 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 5889 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 5890 [Steve Henson] 5891 5892 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 5893 so that complete 'Handshake' protocol structures are kept in memory 5894 instead of overwriting 'msg_type' and 'length' with 'body' data. 5895 [Bodo Moeller] 5896 5897 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 5898 [Massimo Santin via Richard Levitte] 5899 5900 *) Major restructuring to the underlying ENGINE code. This includes 5901 reduction of linker bloat, separation of pure "ENGINE" manipulation 5902 (initialisation, etc) from functionality dealing with implementations 5903 of specific crypto iterfaces. This change also introduces integrated 5904 support for symmetric ciphers and digest implementations - so ENGINEs 5905 can now accelerate these by providing EVP_CIPHER and EVP_MD 5906 implementations of their own. This is detailed in crypto/engine/README 5907 as it couldn't be adequately described here. However, there are a few 5908 API changes worth noting - some RSA, DSA, DH, and RAND functions that 5909 were changed in the original introduction of ENGINE code have now 5910 reverted back - the hooking from this code to ENGINE is now a good 5911 deal more passive and at run-time, operations deal directly with 5912 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 5913 dereferencing through an ENGINE pointer any more. Also, the ENGINE 5914 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 5915 they were not being used by the framework as there is no concept of a 5916 BIGNUM_METHOD and they could not be generalised to the new 5917 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 5918 ENGINE_cpy() has been removed as it cannot be consistently defined in 5919 the new code. 5920 [Geoff Thorpe] 5921 5922 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 5923 [Steve Henson] 5924 5925 *) Change mkdef.pl to sort symbols that get the same entry number, 5926 and make sure the automatically generated functions ERR_load_* 5927 become part of libeay.num as well. 5928 [Richard Levitte] 5929 5930 *) New function SSL_renegotiate_pending(). This returns true once 5931 renegotiation has been requested (either SSL_renegotiate() call 5932 or HelloRequest/ClientHello receveived from the peer) and becomes 5933 false once a handshake has been completed. 5934 (For servers, SSL_renegotiate() followed by SSL_do_handshake() 5935 sends a HelloRequest, but does not ensure that a handshake takes 5936 place. SSL_renegotiate_pending() is useful for checking if the 5937 client has followed the request.) 5938 [Bodo Moeller] 5939 5940 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 5941 By default, clients may request session resumption even during 5942 renegotiation (if session ID contexts permit); with this option, 5943 session resumption is possible only in the first handshake. 5944 5945 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 5946 more bits available for options that should not be part of 5947 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 5948 [Bodo Moeller] 5949 5950 *) Add some demos for certificate and certificate request creation. 5951 [Steve Henson] 5952 5953 *) Make maximum certificate chain size accepted from the peer application 5954 settable (SSL*_get/set_max_cert_list()), as proposed by 5955 "Douglas E. Engert" <deengert@anl.gov>. 5956 [Lutz Jaenicke] 5957 5958 *) Add support for shared libraries for Unixware-7 5959 (Boyd Lynn Gerber <gerberb@zenez.com>). 5960 [Lutz Jaenicke] 5961 5962 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 5963 be done prior to destruction. Use this to unload error strings from 5964 ENGINEs that load their own error strings. NB: This adds two new API 5965 functions to "get" and "set" this destroy handler in an ENGINE. 5966 [Geoff Thorpe] 5967 5968 *) Alter all existing ENGINE implementations (except "openssl" and 5969 "openbsd") to dynamically instantiate their own error strings. This 5970 makes them more flexible to be built both as statically-linked ENGINEs 5971 and self-contained shared-libraries loadable via the "dynamic" ENGINE. 5972 Also, add stub code to each that makes building them as self-contained 5973 shared-libraries easier (see README.ENGINE). 5974 [Geoff Thorpe] 5975 5976 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 5977 implementations into applications that are completely implemented in 5978 self-contained shared-libraries. The "dynamic" ENGINE exposes control 5979 commands that can be used to configure what shared-library to load and 5980 to control aspects of the way it is handled. Also, made an update to 5981 the README.ENGINE file that brings its information up-to-date and 5982 provides some information and instructions on the "dynamic" ENGINE 5983 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 5984 [Geoff Thorpe] 5985 5986 *) Make it possible to unload ranges of ERR strings with a new 5987 "ERR_unload_strings" function. 5988 [Geoff Thorpe] 5989 5990 *) Add a copy() function to EVP_MD. 5991 [Ben Laurie] 5992 5993 *) Make EVP_MD routines take a context pointer instead of just the 5994 md_data void pointer. 5995 [Ben Laurie] 5996 5997 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 5998 that the digest can only process a single chunk of data 5999 (typically because it is provided by a piece of 6000 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 6001 is only going to provide a single chunk of data, and hence the 6002 framework needn't accumulate the data for oneshot drivers. 6003 [Ben Laurie] 6004 6005 *) As with "ERR", make it possible to replace the underlying "ex_data" 6006 functions. This change also alters the storage and management of global 6007 ex_data state - it's now all inside ex_data.c and all "class" code (eg. 6008 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 6009 index counters. The API functions that use this state have been changed 6010 to take a "class_index" rather than pointers to the class's local STACK 6011 and counter, and there is now an API function to dynamically create new 6012 classes. This centralisation allows us to (a) plug a lot of the 6013 thread-safety problems that existed, and (b) makes it possible to clean 6014 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 6015 such data would previously have always leaked in application code and 6016 workarounds were in place to make the memory debugging turn a blind eye 6017 to it. Application code that doesn't use this new function will still 6018 leak as before, but their memory debugging output will announce it now 6019 rather than letting it slide. 6020 6021 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 6022 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 6023 has a return value to indicate success or failure. 6024 [Geoff Thorpe] 6025 6026 *) Make it possible to replace the underlying "ERR" functions such that the 6027 global state (2 LHASH tables and 2 locks) is only used by the "default" 6028 implementation. This change also adds two functions to "get" and "set" 6029 the implementation prior to it being automatically set the first time 6030 any other ERR function takes place. Ie. an application can call "get", 6031 pass the return value to a module it has just loaded, and that module 6032 can call its own "set" function using that value. This means the 6033 module's "ERR" operations will use (and modify) the error state in the 6034 application and not in its own statically linked copy of OpenSSL code. 6035 [Geoff Thorpe] 6036 6037 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 6038 reference counts. This performs normal REF_PRINT/REF_CHECK macros on 6039 the operation, and provides a more encapsulated way for external code 6040 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 6041 to use these functions rather than manually incrementing the counts. 6042 6043 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 6044 [Geoff Thorpe] 6045 6046 *) Add EVP test program. 6047 [Ben Laurie] 6048 6049 *) Add symmetric cipher support to ENGINE. Expect the API to change! 6050 [Ben Laurie] 6051 6052 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 6053 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 6054 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 6055 These allow a CRL to be built without having to access X509_CRL fields 6056 directly. Modify 'ca' application to use new functions. 6057 [Steve Henson] 6058 6059 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 6060 bug workarounds. Rollback attack detection is a security feature. 6061 The problem will only arise on OpenSSL servers when TLSv1 is not 6062 available (sslv3_server_method() or SSL_OP_NO_TLSv1). 6063 Software authors not wanting to support TLSv1 will have special reasons 6064 for their choice and can explicitly enable this option. 6065 [Bodo Moeller, Lutz Jaenicke] 6066 6067 *) Rationalise EVP so it can be extended: don't include a union of 6068 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 6069 (similar to those existing for EVP_CIPHER_CTX). 6070 Usage example: 6071 6072 EVP_MD_CTX md; 6073 6074 EVP_MD_CTX_init(&md); /* new function call */ 6075 EVP_DigestInit(&md, EVP_sha1()); 6076 EVP_DigestUpdate(&md, in, len); 6077 EVP_DigestFinal(&md, out, NULL); 6078 EVP_MD_CTX_cleanup(&md); /* new function call */ 6079 6080 [Ben Laurie] 6081 6082 *) Make DES key schedule conform to the usual scheme, as well as 6083 correcting its structure. This means that calls to DES functions 6084 now have to pass a pointer to a des_key_schedule instead of a 6085 plain des_key_schedule (which was actually always a pointer 6086 anyway): E.g., 6087 6088 des_key_schedule ks; 6089 6090 des_set_key_checked(..., &ks); 6091 des_ncbc_encrypt(..., &ks, ...); 6092 6093 (Note that a later change renames 'des_...' into 'DES_...'.) 6094 [Ben Laurie] 6095 6096 *) Initial reduction of linker bloat: the use of some functions, such as 6097 PEM causes large amounts of unused functions to be linked in due to 6098 poor organisation. For example pem_all.c contains every PEM function 6099 which has a knock on effect of linking in large amounts of (unused) 6100 ASN1 code. Grouping together similar functions and splitting unrelated 6101 functions prevents this. 6102 [Steve Henson] 6103 6104 *) Cleanup of EVP macros. 6105 [Ben Laurie] 6106 6107 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 6108 correct _ecb suffix. 6109 [Ben Laurie] 6110 6111 *) Add initial OCSP responder support to ocsp application. The 6112 revocation information is handled using the text based index 6113 use by the ca application. The responder can either handle 6114 requests generated internally, supplied in files (for example 6115 via a CGI script) or using an internal minimal server. 6116 [Steve Henson] 6117 6118 *) Add configuration choices to get zlib compression for TLS. 6119 [Richard Levitte] 6120 6121 *) Changes to Kerberos SSL for RFC 2712 compliance: 6122 1. Implemented real KerberosWrapper, instead of just using 6123 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 6124 2. Implemented optional authenticator field of KerberosWrapper. 6125 6126 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 6127 and authenticator structs; see crypto/krb5/. 6128 6129 Generalized Kerberos calls to support multiple Kerberos libraries. 6130 [Vern Staats <staatsvr@asc.hpc.mil>, 6131 Jeffrey Altman <jaltman@columbia.edu> 6132 via Richard Levitte] 6133 6134 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 6135 already does with RSA. testdsa.h now has 'priv_key/pub_key' 6136 values for each of the key sizes rather than having just 6137 parameters (and 'speed' generating keys each time). 6138 [Geoff Thorpe] 6139 6140 *) Speed up EVP routines. 6141 Before: 6142encrypt 6143type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 6144des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 6145des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 6146des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 6147decrypt 6148des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 6149des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 6150des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 6151 After: 6152encrypt 6153des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 6154decrypt 6155des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 6156 [Ben Laurie] 6157 6158 *) Added the OS2-EMX target. 6159 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 6160 6161 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 6162 to support NCONF routines in extension code. New function CONF_set_nconf() 6163 to allow functions which take an NCONF to also handle the old LHASH 6164 structure: this means that the old CONF compatible routines can be 6165 retained (in particular wrt extensions) without having to duplicate the 6166 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 6167 [Steve Henson] 6168 6169 *) Enhance the general user interface with mechanisms for inner control 6170 and with possibilities to have yes/no kind of prompts. 6171 [Richard Levitte] 6172 6173 *) Change all calls to low level digest routines in the library and 6174 applications to use EVP. Add missing calls to HMAC_cleanup() and 6175 don't assume HMAC_CTX can be copied using memcpy(). 6176 [Verdon Walker <VWalker@novell.com>, Steve Henson] 6177 6178 *) Add the possibility to control engines through control names but with 6179 arbitrary arguments instead of just a string. 6180 Change the key loaders to take a UI_METHOD instead of a callback 6181 function pointer. NOTE: this breaks binary compatibility with earlier 6182 versions of OpenSSL [engine]. 6183 Adapt the nCipher code for these new conditions and add a card insertion 6184 callback. 6185 [Richard Levitte] 6186 6187 *) Enhance the general user interface with mechanisms to better support 6188 dialog box interfaces, application-defined prompts, the possibility 6189 to use defaults (for example default passwords from somewhere else) 6190 and interrupts/cancellations. 6191 [Richard Levitte] 6192 6193 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 6194 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 6195 [Steve Henson] 6196 6197 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 6198 tidy up some unnecessarily weird code in 'sk_new()'). 6199 [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 6200 6201 *) Change the key loading routines for ENGINEs to use the same kind 6202 callback (pem_password_cb) as all other routines that need this 6203 kind of callback. 6204 [Richard Levitte] 6205 6206 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 6207 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 6208 than this minimum value is recommended. 6209 [Lutz Jaenicke] 6210 6211 *) New random seeder for OpenVMS, using the system process statistics 6212 that are easily reachable. 6213 [Richard Levitte] 6214 6215 *) Windows apparently can't transparently handle global 6216 variables defined in DLLs. Initialisations such as: 6217 6218 const ASN1_ITEM *it = &ASN1_INTEGER_it; 6219 6220 wont compile. This is used by the any applications that need to 6221 declare their own ASN1 modules. This was fixed by adding the option 6222 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 6223 needed for static libraries under Win32. 6224 [Steve Henson] 6225 6226 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 6227 setting of purpose and trust fields. New X509_STORE trust and 6228 purpose functions and tidy up setting in other SSL functions. 6229 [Steve Henson] 6230 6231 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 6232 structure. These are inherited by X509_STORE_CTX when it is 6233 initialised. This allows various defaults to be set in the 6234 X509_STORE structure (such as flags for CRL checking and custom 6235 purpose or trust settings) for functions which only use X509_STORE_CTX 6236 internally such as S/MIME. 6237 6238 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 6239 trust settings if they are not set in X509_STORE. This allows X509_STORE 6240 purposes and trust (in S/MIME for example) to override any set by default. 6241 6242 Add command line options for CRL checking to smime, s_client and s_server 6243 applications. 6244 [Steve Henson] 6245 6246 *) Initial CRL based revocation checking. If the CRL checking flag(s) 6247 are set then the CRL is looked up in the X509_STORE structure and 6248 its validity and signature checked, then if the certificate is found 6249 in the CRL the verify fails with a revoked error. 6250 6251 Various new CRL related callbacks added to X509_STORE_CTX structure. 6252 6253 Command line options added to 'verify' application to support this. 6254 6255 This needs some additional work, such as being able to handle multiple 6256 CRLs with different times, extension based lookup (rather than just 6257 by subject name) and ultimately more complete V2 CRL extension 6258 handling. 6259 [Steve Henson] 6260 6261 *) Add a general user interface API (crypto/ui/). This is designed 6262 to replace things like des_read_password and friends (backward 6263 compatibility functions using this new API are provided). 6264 The purpose is to remove prompting functions from the DES code 6265 section as well as provide for prompting through dialog boxes in 6266 a window system and the like. 6267 [Richard Levitte] 6268 6269 *) Add "ex_data" support to ENGINE so implementations can add state at a 6270 per-structure level rather than having to store it globally. 6271 [Geoff] 6272 6273 *) Make it possible for ENGINE structures to be copied when retrieved by 6274 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 6275 This causes the "original" ENGINE structure to act like a template, 6276 analogous to the RSA vs. RSA_METHOD type of separation. Because of this 6277 operational state can be localised to each ENGINE structure, despite the 6278 fact they all share the same "methods". New ENGINE structures returned in 6279 this case have no functional references and the return value is the single 6280 structural reference. This matches the single structural reference returned 6281 by ENGINE_by_id() normally, when it is incremented on the pre-existing 6282 ENGINE structure. 6283 [Geoff] 6284 6285 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 6286 needs to match any other type at all we need to manually clear the 6287 tag cache. 6288 [Steve Henson] 6289 6290 *) Changes to the "openssl engine" utility to include; 6291 - verbosity levels ('-v', '-vv', and '-vvv') that provide information 6292 about an ENGINE's available control commands. 6293 - executing control commands from command line arguments using the 6294 '-pre' and '-post' switches. '-post' is only used if '-t' is 6295 specified and the ENGINE is successfully initialised. The syntax for 6296 the individual commands are colon-separated, for example; 6297 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 6298 [Geoff] 6299 6300 *) New dynamic control command support for ENGINEs. ENGINEs can now 6301 declare their own commands (numbers), names (strings), descriptions, 6302 and input types for run-time discovery by calling applications. A 6303 subset of these commands are implicitly classed as "executable" 6304 depending on their input type, and only these can be invoked through 6305 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 6306 can be based on user input, config files, etc). The distinction is 6307 that "executable" commands cannot return anything other than a boolean 6308 result and can only support numeric or string input, whereas some 6309 discoverable commands may only be for direct use through 6310 ENGINE_ctrl(), eg. supporting the exchange of binary data, function 6311 pointers, or other custom uses. The "executable" commands are to 6312 support parameterisations of ENGINE behaviour that can be 6313 unambiguously defined by ENGINEs and used consistently across any 6314 OpenSSL-based application. Commands have been added to all the 6315 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 6316 control over shared-library paths without source code alterations. 6317 [Geoff] 6318 6319 *) Changed all ENGINE implementations to dynamically allocate their 6320 ENGINEs rather than declaring them statically. Apart from this being 6321 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 6322 this also allows the implementations to compile without using the 6323 internal engine_int.h header. 6324 [Geoff] 6325 6326 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 6327 'const' value. Any code that should be able to modify a RAND_METHOD 6328 should already have non-const pointers to it (ie. they should only 6329 modify their own ones). 6330 [Geoff] 6331 6332 *) Made a variety of little tweaks to the ENGINE code. 6333 - "atalla" and "ubsec" string definitions were moved from header files 6334 to C code. "nuron" string definitions were placed in variables 6335 rather than hard-coded - allowing parameterisation of these values 6336 later on via ctrl() commands. 6337 - Removed unused "#if 0"'d code. 6338 - Fixed engine list iteration code so it uses ENGINE_free() to release 6339 structural references. 6340 - Constified the RAND_METHOD element of ENGINE structures. 6341 - Constified various get/set functions as appropriate and added 6342 missing functions (including a catch-all ENGINE_cpy that duplicates 6343 all ENGINE values onto a new ENGINE except reference counts/state). 6344 - Removed NULL parameter checks in get/set functions. Setting a method 6345 or function to NULL is a way of cancelling out a previously set 6346 value. Passing a NULL ENGINE parameter is just plain stupid anyway 6347 and doesn't justify the extra error symbols and code. 6348 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 6349 flags from engine_int.h to engine.h. 6350 - Changed prototypes for ENGINE handler functions (init(), finish(), 6351 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 6352 [Geoff] 6353 6354 *) Implement binary inversion algorithm for BN_mod_inverse in addition 6355 to the algorithm using long division. The binary algorithm can be 6356 used only if the modulus is odd. On 32-bit systems, it is faster 6357 only for relatively small moduli (roughly 20-30% for 128-bit moduli, 6358 roughly 5-15% for 256-bit moduli), so we use it only for moduli 6359 up to 450 bits. In 64-bit environments, the binary algorithm 6360 appears to be advantageous for much longer moduli; here we use it 6361 for moduli up to 2048 bits. 6362 [Bodo Moeller] 6363 6364 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 6365 could not support the combine flag in choice fields. 6366 [Steve Henson] 6367 6368 *) Add a 'copy_extensions' option to the 'ca' utility. This copies 6369 extensions from a certificate request to the certificate. 6370 [Steve Henson] 6371 6372 *) Allow multiple 'certopt' and 'nameopt' options to be separated 6373 by commas. Add 'namopt' and 'certopt' options to the 'ca' config 6374 file: this allows the display of the certificate about to be 6375 signed to be customised, to allow certain fields to be included 6376 or excluded and extension details. The old system didn't display 6377 multicharacter strings properly, omitted fields not in the policy 6378 and couldn't display additional details such as extensions. 6379 [Steve Henson] 6380 6381 *) Function EC_POINTs_mul for multiple scalar multiplication 6382 of an arbitrary number of elliptic curve points 6383 \sum scalars[i]*points[i], 6384 optionally including the generator defined for the EC_GROUP: 6385 scalar*generator + \sum scalars[i]*points[i]. 6386 6387 EC_POINT_mul is a simple wrapper function for the typical case 6388 that the point list has just one item (besides the optional 6389 generator). 6390 [Bodo Moeller] 6391 6392 *) First EC_METHODs for curves over GF(p): 6393 6394 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 6395 operations and provides various method functions that can also 6396 operate with faster implementations of modular arithmetic. 6397 6398 EC_GFp_mont_method() reuses most functions that are part of 6399 EC_GFp_simple_method, but uses Montgomery arithmetic. 6400 6401 [Bodo Moeller; point addition and point doubling 6402 implementation directly derived from source code provided by 6403 Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 6404 6405 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 6406 crypto/ec/ec_lib.c): 6407 6408 Curves are EC_GROUP objects (with an optional group generator) 6409 based on EC_METHODs that are built into the library. 6410 6411 Points are EC_POINT objects based on EC_GROUP objects. 6412 6413 Most of the framework would be able to handle curves over arbitrary 6414 finite fields, but as there are no obvious types for fields other 6415 than GF(p), some functions are limited to that for now. 6416 [Bodo Moeller] 6417 6418 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 6419 that the file contains a complete HTTP response. 6420 [Richard Levitte] 6421 6422 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 6423 change the def and num file printf format specifier from "%-40sXXX" 6424 to "%-39s XXX". The latter will always guarantee a space after the 6425 field while the former will cause them to run together if the field 6426 is 40 of more characters long. 6427 [Steve Henson] 6428 6429 *) Constify the cipher and digest 'method' functions and structures 6430 and modify related functions to take constant EVP_MD and EVP_CIPHER 6431 pointers. 6432 [Steve Henson] 6433 6434 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 6435 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 6436 [Bodo Moeller] 6437 6438 *) Modify EVP_Digest*() routines so they now return values. Although the 6439 internal software routines can never fail additional hardware versions 6440 might. 6441 [Steve Henson] 6442 6443 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 6444 6445 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 6446 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 6447 6448 ASN1 error codes 6449 ERR_R_NESTED_ASN1_ERROR 6450 ... 6451 ERR_R_MISSING_ASN1_EOS 6452 were 4 .. 9, conflicting with 6453 ERR_LIB_RSA (= ERR_R_RSA_LIB) 6454 ... 6455 ERR_LIB_PEM (= ERR_R_PEM_LIB). 6456 They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 6457 6458 Add new error code 'ERR_R_INTERNAL_ERROR'. 6459 [Bodo Moeller] 6460 6461 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 6462 suffices. 6463 [Bodo Moeller] 6464 6465 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 6466 sets the subject name for a new request or supersedes the 6467 subject name in a given request. Formats that can be parsed are 6468 'CN=Some Name, OU=myOU, C=IT' 6469 and 6470 'CN=Some Name/OU=myOU/C=IT'. 6471 6472 Add options '-batch' and '-verbose' to 'openssl req'. 6473 [Massimiliano Pala <madwolf@hackmasters.net>] 6474 6475 *) Introduce the possibility to access global variables through 6476 functions on platform were that's the best way to handle exporting 6477 global variables in shared libraries. To enable this functionality, 6478 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 6479 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 6480 is normally done by Configure or something similar). 6481 6482 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 6483 in the source file (foo.c) like this: 6484 6485 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 6486 OPENSSL_IMPLEMENT_GLOBAL(double,bar); 6487 6488 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 6489 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 6490 6491 OPENSSL_DECLARE_GLOBAL(int,foo); 6492 #define foo OPENSSL_GLOBAL_REF(foo) 6493 OPENSSL_DECLARE_GLOBAL(double,bar); 6494 #define bar OPENSSL_GLOBAL_REF(bar) 6495 6496 The #defines are very important, and therefore so is including the 6497 header file everywhere where the defined globals are used. 6498 6499 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 6500 of ASN.1 items, but that structure is a bit different. 6501 6502 The largest change is in util/mkdef.pl which has been enhanced with 6503 better and easier to understand logic to choose which symbols should 6504 go into the Windows .def files as well as a number of fixes and code 6505 cleanup (among others, algorithm keywords are now sorted 6506 lexicographically to avoid constant rewrites). 6507 [Richard Levitte] 6508 6509 *) In BN_div() keep a copy of the sign of 'num' before writing the 6510 result to 'rm' because if rm==num the value will be overwritten 6511 and produce the wrong result if 'num' is negative: this caused 6512 problems with BN_mod() and BN_nnmod(). 6513 [Steve Henson] 6514 6515 *) Function OCSP_request_verify(). This checks the signature on an 6516 OCSP request and verifies the signer certificate. The signer 6517 certificate is just checked for a generic purpose and OCSP request 6518 trust settings. 6519 [Steve Henson] 6520 6521 *) Add OCSP_check_validity() function to check the validity of OCSP 6522 responses. OCSP responses are prepared in real time and may only 6523 be a few seconds old. Simply checking that the current time lies 6524 between thisUpdate and nextUpdate max reject otherwise valid responses 6525 caused by either OCSP responder or client clock inaccuracy. Instead 6526 we allow thisUpdate and nextUpdate to fall within a certain period of 6527 the current time. The age of the response can also optionally be 6528 checked. Two new options -validity_period and -status_age added to 6529 ocsp utility. 6530 [Steve Henson] 6531 6532 *) If signature or public key algorithm is unrecognized print out its 6533 OID rather that just UNKNOWN. 6534 [Steve Henson] 6535 6536 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 6537 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 6538 ID to be generated from the issuer certificate alone which can then be 6539 passed to OCSP_id_issuer_cmp(). 6540 [Steve Henson] 6541 6542 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 6543 ASN1 modules to export functions returning ASN1_ITEM pointers 6544 instead of the ASN1_ITEM structures themselves. This adds several 6545 new macros which allow the underlying ASN1 function/structure to 6546 be accessed transparently. As a result code should not use ASN1_ITEM 6547 references directly (such as &X509_it) but instead use the relevant 6548 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 6549 use of the new ASN1 code on platforms where exporting structures 6550 is problematical (for example in shared libraries) but exporting 6551 functions returning pointers to structures is not. 6552 [Steve Henson] 6553 6554 *) Add support for overriding the generation of SSL/TLS session IDs. 6555 These callbacks can be registered either in an SSL_CTX or per SSL. 6556 The purpose of this is to allow applications to control, if they wish, 6557 the arbitrary values chosen for use as session IDs, particularly as it 6558 can be useful for session caching in multiple-server environments. A 6559 command-line switch for testing this (and any client code that wishes 6560 to use such a feature) has been added to "s_server". 6561 [Geoff Thorpe, Lutz Jaenicke] 6562 6563 *) Modify mkdef.pl to recognise and parse preprocessor conditionals 6564 of the form '#if defined(...) || defined(...) || ...' and 6565 '#if !defined(...) && !defined(...) && ...'. This also avoids 6566 the growing number of special cases it was previously handling. 6567 [Richard Levitte] 6568 6569 *) Make all configuration macros available for application by making 6570 sure they are available in opensslconf.h, by giving them names starting 6571 with "OPENSSL_" to avoid conflicts with other packages and by making 6572 sure e_os2.h will cover all platform-specific cases together with 6573 opensslconf.h. 6574 Additionally, it is now possible to define configuration/platform- 6575 specific names (called "system identities"). In the C code, these 6576 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 6577 macro with the name beginning with "OPENSSL_SYS_", which is determined 6578 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 6579 what is available. 6580 [Richard Levitte] 6581 6582 *) New option -set_serial to 'req' and 'x509' this allows the serial 6583 number to use to be specified on the command line. Previously self 6584 signed certificates were hard coded with serial number 0 and the 6585 CA options of 'x509' had to use a serial number in a file which was 6586 auto incremented. 6587 [Steve Henson] 6588 6589 *) New options to 'ca' utility to support V2 CRL entry extensions. 6590 Currently CRL reason, invalidity date and hold instruction are 6591 supported. Add new CRL extensions to V3 code and some new objects. 6592 [Steve Henson] 6593 6594 *) New function EVP_CIPHER_CTX_set_padding() this is used to 6595 disable standard block padding (aka PKCS#5 padding) in the EVP 6596 API, which was previously mandatory. This means that the data is 6597 not padded in any way and so the total length much be a multiple 6598 of the block size, otherwise an error occurs. 6599 [Steve Henson] 6600 6601 *) Initial (incomplete) OCSP SSL support. 6602 [Steve Henson] 6603 6604 *) New function OCSP_parse_url(). This splits up a URL into its host, 6605 port and path components: primarily to parse OCSP URLs. New -url 6606 option to ocsp utility. 6607 [Steve Henson] 6608 6609 *) New nonce behavior. The return value of OCSP_check_nonce() now 6610 reflects the various checks performed. Applications can decide 6611 whether to tolerate certain situations such as an absent nonce 6612 in a response when one was present in a request: the ocsp application 6613 just prints out a warning. New function OCSP_add1_basic_nonce() 6614 this is to allow responders to include a nonce in a response even if 6615 the request is nonce-less. 6616 [Steve Henson] 6617 6618 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 6619 skipped when using openssl x509 multiple times on a single input file, 6620 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 6621 [Bodo Moeller] 6622 6623 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 6624 set string type: to handle setting ASN1_TIME structures. Fix ca 6625 utility to correctly initialize revocation date of CRLs. 6626 [Steve Henson] 6627 6628 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 6629 the clients preferred ciphersuites and rather use its own preferences. 6630 Should help to work around M$ SGC (Server Gated Cryptography) bug in 6631 Internet Explorer by ensuring unchanged hash method during stepup. 6632 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 6633 [Lutz Jaenicke] 6634 6635 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 6636 to aes and add a new 'exist' option to print out symbols that don't 6637 appear to exist. 6638 [Steve Henson] 6639 6640 *) Additional options to ocsp utility to allow flags to be set and 6641 additional certificates supplied. 6642 [Steve Henson] 6643 6644 *) Add the option -VAfile to 'openssl ocsp', so the user can give the 6645 OCSP client a number of certificate to only verify the response 6646 signature against. 6647 [Richard Levitte] 6648 6649 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 6650 handle the new API. Currently only ECB, CBC modes supported. Add new 6651 AES OIDs. 6652 6653 Add TLS AES ciphersuites as described in RFC3268, "Advanced 6654 Encryption Standard (AES) Ciphersuites for Transport Layer 6655 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 6656 not enabled by default and were not part of the "ALL" ciphersuite 6657 alias because they were not yet official; they could be 6658 explicitly requested by specifying the "AESdraft" ciphersuite 6659 group alias. In the final release of OpenSSL 0.9.7, the group 6660 alias is called "AES" and is part of "ALL".) 6661 [Ben Laurie, Steve Henson, Bodo Moeller] 6662 6663 *) New function OCSP_copy_nonce() to copy nonce value (if present) from 6664 request to response. 6665 [Steve Henson] 6666 6667 *) Functions for OCSP responders. OCSP_request_onereq_count(), 6668 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 6669 extract information from a certificate request. OCSP_response_create() 6670 creates a response and optionally adds a basic response structure. 6671 OCSP_basic_add1_status() adds a complete single response to a basic 6672 response and returns the OCSP_SINGLERESP structure just added (to allow 6673 extensions to be included for example). OCSP_basic_add1_cert() adds a 6674 certificate to a basic response and OCSP_basic_sign() signs a basic 6675 response with various flags. New helper functions ASN1_TIME_check() 6676 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 6677 (converts ASN1_TIME to GeneralizedTime). 6678 [Steve Henson] 6679 6680 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 6681 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 6682 structure from a certificate. X509_pubkey_digest() digests the public_key 6683 contents: this is used in various key identifiers. 6684 [Steve Henson] 6685 6686 *) Make sk_sort() tolerate a NULL argument. 6687 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 6688 6689 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 6690 passed by the function are trusted implicitly. If any of them signed the 6691 response then it is assumed to be valid and is not verified. 6692 [Steve Henson] 6693 6694 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 6695 to data. This was previously part of the PKCS7 ASN1 code. This 6696 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 6697 [Steve Henson, reported by Kenneth R. Robinette 6698 <support@securenetterm.com>] 6699 6700 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 6701 routines: without these tracing memory leaks is very painful. 6702 Fix leaks in PKCS12 and PKCS7 routines. 6703 [Steve Henson] 6704 6705 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 6706 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 6707 effectively meant GeneralizedTime would never be used. Now it 6708 is initialised to -1 but X509_time_adj() now has to check the value 6709 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 6710 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 6711 [Steve Henson, reported by Kenneth R. Robinette 6712 <support@securenetterm.com>] 6713 6714 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 6715 result in a zero length in the ASN1_INTEGER structure which was 6716 not consistent with the structure when d2i_ASN1_INTEGER() was used 6717 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 6718 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 6719 where it did not print out a minus for negative ASN1_INTEGER. 6720 [Steve Henson] 6721 6722 *) Add summary printout to ocsp utility. The various functions which 6723 convert status values to strings have been renamed to: 6724 OCSP_response_status_str(), OCSP_cert_status_str() and 6725 OCSP_crl_reason_str() and are no longer static. New options 6726 to verify nonce values and to disable verification. OCSP response 6727 printout format cleaned up. 6728 [Steve Henson] 6729 6730 *) Add additional OCSP certificate checks. These are those specified 6731 in RFC2560. This consists of two separate checks: the CA of the 6732 certificate being checked must either be the OCSP signer certificate 6733 or the issuer of the OCSP signer certificate. In the latter case the 6734 OCSP signer certificate must contain the OCSP signing extended key 6735 usage. This check is performed by attempting to match the OCSP 6736 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 6737 in the OCSP_CERTID structures of the response. 6738 [Steve Henson] 6739 6740 *) Initial OCSP certificate verification added to OCSP_basic_verify() 6741 and related routines. This uses the standard OpenSSL certificate 6742 verify routines to perform initial checks (just CA validity) and 6743 to obtain the certificate chain. Then additional checks will be 6744 performed on the chain. Currently the root CA is checked to see 6745 if it is explicitly trusted for OCSP signing. This is used to set 6746 a root CA as a global signing root: that is any certificate that 6747 chains to that CA is an acceptable OCSP signing certificate. 6748 [Steve Henson] 6749 6750 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 6751 extensions from a separate configuration file. 6752 As when reading extensions from the main configuration file, 6753 the '-extensions ...' option may be used for specifying the 6754 section to use. 6755 [Massimiliano Pala <madwolf@comune.modena.it>] 6756 6757 *) New OCSP utility. Allows OCSP requests to be generated or 6758 read. The request can be sent to a responder and the output 6759 parsed, outputed or printed in text form. Not complete yet: 6760 still needs to check the OCSP response validity. 6761 [Steve Henson] 6762 6763 *) New subcommands for 'openssl ca': 6764 'openssl ca -status <serial>' prints the status of the cert with 6765 the given serial number (according to the index file). 6766 'openssl ca -updatedb' updates the expiry status of certificates 6767 in the index file. 6768 [Massimiliano Pala <madwolf@comune.modena.it>] 6769 6770 *) New '-newreq-nodes' command option to CA.pl. This is like 6771 '-newreq', but calls 'openssl req' with the '-nodes' option 6772 so that the resulting key is not encrypted. 6773 [Damien Miller <djm@mindrot.org>] 6774 6775 *) New configuration for the GNU Hurd. 6776 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 6777 6778 *) Initial code to implement OCSP basic response verify. This 6779 is currently incomplete. Currently just finds the signer's 6780 certificate and verifies the signature on the response. 6781 [Steve Henson] 6782 6783 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 6784 value of OPENSSLDIR. This is available via the new '-d' option 6785 to 'openssl version', and is also included in 'openssl version -a'. 6786 [Bodo Moeller] 6787 6788 *) Allowing defining memory allocation callbacks that will be given 6789 file name and line number information in additional arguments 6790 (a const char* and an int). The basic functionality remains, as 6791 well as the original possibility to just replace malloc(), 6792 realloc() and free() by functions that do not know about these 6793 additional arguments. To register and find out the current 6794 settings for extended allocation functions, the following 6795 functions are provided: 6796 6797 CRYPTO_set_mem_ex_functions 6798 CRYPTO_set_locked_mem_ex_functions 6799 CRYPTO_get_mem_ex_functions 6800 CRYPTO_get_locked_mem_ex_functions 6801 6802 These work the same way as CRYPTO_set_mem_functions and friends. 6803 CRYPTO_get_[locked_]mem_functions now writes 0 where such an 6804 extended allocation function is enabled. 6805 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 6806 a conventional allocation function is enabled. 6807 [Richard Levitte, Bodo Moeller] 6808 6809 *) Finish off removing the remaining LHASH function pointer casts. 6810 There should no longer be any prototype-casting required when using 6811 the LHASH abstraction, and any casts that remain are "bugs". See 6812 the callback types and macros at the head of lhash.h for details 6813 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 6814 [Geoff Thorpe] 6815 6816 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 6817 If /dev/[u]random devices are not available or do not return enough 6818 entropy, EGD style sockets (served by EGD or PRNGD) will automatically 6819 be queried. 6820 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 6821 /etc/entropy will be queried once each in this sequence, quering stops 6822 when enough entropy was collected without querying more sockets. 6823 [Lutz Jaenicke] 6824 6825 *) Change the Unix RAND_poll() variant to be able to poll several 6826 random devices, as specified by DEVRANDOM, until a sufficient amount 6827 of data has been collected. We spend at most 10 ms on each file 6828 (select timeout) and read in non-blocking mode. DEVRANDOM now 6829 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 6830 (previously it was just the string "/dev/urandom"), so on typical 6831 platforms the 10 ms delay will never occur. 6832 Also separate out the Unix variant to its own file, rand_unix.c. 6833 For VMS, there's a currently-empty rand_vms.c. 6834 [Richard Levitte] 6835 6836 *) Move OCSP client related routines to ocsp_cl.c. These 6837 provide utility functions which an application needing 6838 to issue a request to an OCSP responder and analyse the 6839 response will typically need: as opposed to those which an 6840 OCSP responder itself would need which will be added later. 6841 6842 OCSP_request_sign() signs an OCSP request with an API similar 6843 to PKCS7_sign(). OCSP_response_status() returns status of OCSP 6844 response. OCSP_response_get1_basic() extracts basic response 6845 from response. OCSP_resp_find_status(): finds and extracts status 6846 information from an OCSP_CERTID structure (which will be created 6847 when the request structure is built). These are built from lower 6848 level functions which work on OCSP_SINGLERESP structures but 6849 wont normally be used unless the application wishes to examine 6850 extensions in the OCSP response for example. 6851 6852 Replace nonce routines with a pair of functions. 6853 OCSP_request_add1_nonce() adds a nonce value and optionally 6854 generates a random value. OCSP_check_nonce() checks the 6855 validity of the nonce in an OCSP response. 6856 [Steve Henson] 6857 6858 *) Change function OCSP_request_add() to OCSP_request_add0_id(). 6859 This doesn't copy the supplied OCSP_CERTID and avoids the 6860 need to free up the newly created id. Change return type 6861 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 6862 This can then be used to add extensions to the request. 6863 Deleted OCSP_request_new(), since most of its functionality 6864 is now in OCSP_REQUEST_new() (and the case insensitive name 6865 clash) apart from the ability to set the request name which 6866 will be added elsewhere. 6867 [Steve Henson] 6868 6869 *) Update OCSP API. Remove obsolete extensions argument from 6870 various functions. Extensions are now handled using the new 6871 OCSP extension code. New simple OCSP HTTP function which 6872 can be used to send requests and parse the response. 6873 [Steve Henson] 6874 6875 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 6876 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 6877 uses the special reorder version of SET OF to sort the attributes 6878 and reorder them to match the encoded order. This resolves a long 6879 standing problem: a verify on a PKCS7 structure just after signing 6880 it used to fail because the attribute order did not match the 6881 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 6882 it uses the received order. This is necessary to tolerate some broken 6883 software that does not order SET OF. This is handled by encoding 6884 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 6885 to produce the required SET OF. 6886 [Steve Henson] 6887 6888 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 6889 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 6890 files to get correct declarations of the ASN.1 item variables. 6891 [Richard Levitte] 6892 6893 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 6894 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 6895 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 6896 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 6897 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 6898 ASN1_ITEM and no wrapper functions. 6899 [Steve Henson] 6900 6901 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 6902 replace the old function pointer based I/O routines. Change most of 6903 the *_d2i_bio() and *_d2i_fp() functions to use these. 6904 [Steve Henson] 6905 6906 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 6907 lines, recognice more "algorithms" that can be deselected, and make 6908 it complain about algorithm deselection that isn't recognised. 6909 [Richard Levitte] 6910 6911 *) New ASN1 functions to handle dup, sign, verify, digest, pack and 6912 unpack operations in terms of ASN1_ITEM. Modify existing wrappers 6913 to use new functions. Add NO_ASN1_OLD which can be set to remove 6914 some old style ASN1 functions: this can be used to determine if old 6915 code will still work when these eventually go away. 6916 [Steve Henson] 6917 6918 *) New extension functions for OCSP structures, these follow the 6919 same conventions as certificates and CRLs. 6920 [Steve Henson] 6921 6922 *) New function X509V3_add1_i2d(). This automatically encodes and 6923 adds an extension. Its behaviour can be customised with various 6924 flags to append, replace or delete. Various wrappers added for 6925 certifcates and CRLs. 6926 [Steve Henson] 6927 6928 *) Fix to avoid calling the underlying ASN1 print routine when 6929 an extension cannot be parsed. Correct a typo in the 6930 OCSP_SERVICELOC extension. Tidy up print OCSP format. 6931 [Steve Henson] 6932 6933 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 6934 entries for variables. 6935 [Steve Henson] 6936 6937 *) Add functionality to apps/openssl.c for detecting locking 6938 problems: As the program is single-threaded, all we have 6939 to do is register a locking callback using an array for 6940 storing which locks are currently held by the program. 6941 [Bodo Moeller] 6942 6943 *) Use a lock around the call to CRYPTO_get_ex_new_index() in 6944 SSL_get_ex_data_X509_STORE_idx(), which is used in 6945 ssl_verify_cert_chain() and thus can be called at any time 6946 during TLS/SSL handshakes so that thread-safety is essential. 6947 Unfortunately, the ex_data design is not at all suited 6948 for multi-threaded use, so it probably should be abolished. 6949 [Bodo Moeller] 6950 6951 *) Added Broadcom "ubsec" ENGINE to OpenSSL. 6952 [Broadcom, tweaked and integrated by Geoff Thorpe] 6953 6954 *) Move common extension printing code to new function 6955 X509V3_print_extensions(). Reorganise OCSP print routines and 6956 implement some needed OCSP ASN1 functions. Add OCSP extensions. 6957 [Steve Henson] 6958 6959 *) New function X509_signature_print() to remove duplication in some 6960 print routines. 6961 [Steve Henson] 6962 6963 *) Add a special meaning when SET OF and SEQUENCE OF flags are both 6964 set (this was treated exactly the same as SET OF previously). This 6965 is used to reorder the STACK representing the structure to match the 6966 encoding. This will be used to get round a problem where a PKCS7 6967 structure which was signed could not be verified because the STACK 6968 order did not reflect the encoded order. 6969 [Steve Henson] 6970 6971 *) Reimplement the OCSP ASN1 module using the new code. 6972 [Steve Henson] 6973 6974 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 6975 for its ASN1 operations. The old style function pointers still exist 6976 for now but they will eventually go away. 6977 [Steve Henson] 6978 6979 *) Merge in replacement ASN1 code from the ASN1 branch. This almost 6980 completely replaces the old ASN1 functionality with a table driven 6981 encoder and decoder which interprets an ASN1_ITEM structure describing 6982 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 6983 largely maintained. Almost all of the old asn1_mac.h macro based ASN1 6984 has also been converted to the new form. 6985 [Steve Henson] 6986 6987 *) Change BN_mod_exp_recp so that negative moduli are tolerated 6988 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 6989 so that BN_mod_exp_mont and BN_mod_exp_mont_word work 6990 for negative moduli. 6991 [Bodo Moeller] 6992 6993 *) Fix BN_uadd and BN_usub: Always return non-negative results instead 6994 of not touching the result's sign bit. 6995 [Bodo Moeller] 6996 6997 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 6998 set. 6999 [Bodo Moeller] 7000 7001 *) Changed the LHASH code to use prototypes for callbacks, and created 7002 macros to declare and implement thin (optionally static) functions 7003 that provide type-safety and avoid function pointer casting for the 7004 type-specific callbacks. 7005 [Geoff Thorpe] 7006 7007 *) Added Kerberos Cipher Suites to be used with TLS, as written in 7008 RFC 2712. 7009 [Veers Staats <staatsvr@asc.hpc.mil>, 7010 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 7011 7012 *) Reformat the FAQ so the different questions and answers can be divided 7013 in sections depending on the subject. 7014 [Richard Levitte] 7015 7016 *) Have the zlib compression code load ZLIB.DLL dynamically under 7017 Windows. 7018 [Richard Levitte] 7019 7020 *) New function BN_mod_sqrt for computing square roots modulo a prime 7021 (using the probabilistic Tonelli-Shanks algorithm unless 7022 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 7023 be handled deterministically). 7024 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 7025 7026 *) Make BN_mod_inverse faster by explicitly handling small quotients 7027 in the Euclid loop. (Speed gain about 20% for small moduli [256 or 7028 512 bits], about 30% for larger ones [1024 or 2048 bits].) 7029 [Bodo Moeller] 7030 7031 *) New function BN_kronecker. 7032 [Bodo Moeller] 7033 7034 *) Fix BN_gcd so that it works on negative inputs; the result is 7035 positive unless both parameters are zero. 7036 Previously something reasonably close to an infinite loop was 7037 possible because numbers could be growing instead of shrinking 7038 in the implementation of Euclid's algorithm. 7039 [Bodo Moeller] 7040 7041 *) Fix BN_is_word() and BN_is_one() macros to take into account the 7042 sign of the number in question. 7043 7044 Fix BN_is_word(a,w) to work correctly for w == 0. 7045 7046 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 7047 because its test if the absolute value of 'a' equals 'w'. 7048 Note that BN_abs_is_word does *not* handle w == 0 reliably; 7049 it exists mostly for use in the implementations of BN_is_zero(), 7050 BN_is_one(), and BN_is_word(). 7051 [Bodo Moeller] 7052 7053 *) New function BN_swap. 7054 [Bodo Moeller] 7055 7056 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 7057 the exponentiation functions are more likely to produce reasonable 7058 results on negative inputs. 7059 [Bodo Moeller] 7060 7061 *) Change BN_mod_mul so that the result is always non-negative. 7062 Previously, it could be negative if one of the factors was negative; 7063 I don't think anyone really wanted that behaviour. 7064 [Bodo Moeller] 7065 7066 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 7067 (except for exponentiation, which stays in crypto/bn/bn_exp.c, 7068 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 7069 and add new functions: 7070 7071 BN_nnmod 7072 BN_mod_sqr 7073 BN_mod_add 7074 BN_mod_add_quick 7075 BN_mod_sub 7076 BN_mod_sub_quick 7077 BN_mod_lshift1 7078 BN_mod_lshift1_quick 7079 BN_mod_lshift 7080 BN_mod_lshift_quick 7081 7082 These functions always generate non-negative results. 7083 7084 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 7085 such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 7086 7087 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 7088 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 7089 be reduced modulo m. 7090 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 7091 7092#if 0 7093 The following entry accidentily appeared in the CHANGES file 7094 distributed with OpenSSL 0.9.7. The modifications described in 7095 it do *not* apply to OpenSSL 0.9.7. 7096 7097 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 7098 was actually never needed) and in BN_mul(). The removal in BN_mul() 7099 required a small change in bn_mul_part_recursive() and the addition 7100 of the functions bn_cmp_part_words(), bn_sub_part_words() and 7101 bn_add_part_words(), which do the same thing as bn_cmp_words(), 7102 bn_sub_words() and bn_add_words() except they take arrays with 7103 differing sizes. 7104 [Richard Levitte] 7105#endif 7106 7107 *) In 'openssl passwd', verify passwords read from the terminal 7108 unless the '-salt' option is used (which usually means that 7109 verification would just waste user's time since the resulting 7110 hash is going to be compared with some given password hash) 7111 or the new '-noverify' option is used. 7112 7113 This is an incompatible change, but it does not affect 7114 non-interactive use of 'openssl passwd' (passwords on the command 7115 line, '-stdin' option, '-in ...' option) and thus should not 7116 cause any problems. 7117 [Bodo Moeller] 7118 7119 *) Remove all references to RSAref, since there's no more need for it. 7120 [Richard Levitte] 7121 7122 *) Make DSO load along a path given through an environment variable 7123 (SHLIB_PATH) with shl_load(). 7124 [Richard Levitte] 7125 7126 *) Constify the ENGINE code as a result of BIGNUM constification. 7127 Also constify the RSA code and most things related to it. In a 7128 few places, most notable in the depth of the ASN.1 code, ugly 7129 casts back to non-const were required (to be solved at a later 7130 time) 7131 [Richard Levitte] 7132 7133 *) Make it so the openssl application has all engines loaded by default. 7134 [Richard Levitte] 7135 7136 *) Constify the BIGNUM routines a little more. 7137 [Richard Levitte] 7138 7139 *) Add the following functions: 7140 7141 ENGINE_load_cswift() 7142 ENGINE_load_chil() 7143 ENGINE_load_atalla() 7144 ENGINE_load_nuron() 7145 ENGINE_load_builtin_engines() 7146 7147 That way, an application can itself choose if external engines that 7148 are built-in in OpenSSL shall ever be used or not. The benefit is 7149 that applications won't have to be linked with libdl or other dso 7150 libraries unless it's really needed. 7151 7152 Changed 'openssl engine' to load all engines on demand. 7153 Changed the engine header files to avoid the duplication of some 7154 declarations (they differed!). 7155 [Richard Levitte] 7156 7157 *) 'openssl engine' can now list capabilities. 7158 [Richard Levitte] 7159 7160 *) Better error reporting in 'openssl engine'. 7161 [Richard Levitte] 7162 7163 *) Never call load_dh_param(NULL) in s_server. 7164 [Bodo Moeller] 7165 7166 *) Add engine application. It can currently list engines by name and 7167 identity, and test if they are actually available. 7168 [Richard Levitte] 7169 7170 *) Improve RPM specification file by forcing symbolic linking and making 7171 sure the installed documentation is also owned by root.root. 7172 [Damien Miller <djm@mindrot.org>] 7173 7174 *) Give the OpenSSL applications more possibilities to make use of 7175 keys (public as well as private) handled by engines. 7176 [Richard Levitte] 7177 7178 *) Add OCSP code that comes from CertCo. 7179 [Richard Levitte] 7180 7181 *) Add VMS support for the Rijndael code. 7182 [Richard Levitte] 7183 7184 *) Added untested support for Nuron crypto accelerator. 7185 [Ben Laurie] 7186 7187 *) Add support for external cryptographic devices. This code was 7188 previously distributed separately as the "engine" branch. 7189 [Geoff Thorpe, Richard Levitte] 7190 7191 *) Rework the filename-translation in the DSO code. It is now possible to 7192 have far greater control over how a "name" is turned into a filename 7193 depending on the operating environment and any oddities about the 7194 different shared library filenames on each system. 7195 [Geoff Thorpe] 7196 7197 *) Support threads on FreeBSD-elf in Configure. 7198 [Richard Levitte] 7199 7200 *) Fix for SHA1 assembly problem with MASM: it produces 7201 warnings about corrupt line number information when assembling 7202 with debugging information. This is caused by the overlapping 7203 of two sections. 7204 [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 7205 7206 *) NCONF changes. 7207 NCONF_get_number() has no error checking at all. As a replacement, 7208 NCONF_get_number_e() is defined (_e for "error checking") and is 7209 promoted strongly. The old NCONF_get_number is kept around for 7210 binary backward compatibility. 7211 Make it possible for methods to load from something other than a BIO, 7212 by providing a function pointer that is given a name instead of a BIO. 7213 For example, this could be used to load configuration data from an 7214 LDAP server. 7215 [Richard Levitte] 7216 7217 *) Fix for non blocking accept BIOs. Added new I/O special reason 7218 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 7219 with non blocking I/O was not possible because no retry code was 7220 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 7221 this case. 7222 [Steve Henson] 7223 7224 *) Added the beginnings of Rijndael support. 7225 [Ben Laurie] 7226 7227 *) Fix for bug in DirectoryString mask setting. Add support for 7228 X509_NAME_print_ex() in 'req' and X509_print_ex() function 7229 to allow certificate printing to more controllable, additional 7230 'certopt' option to 'x509' to allow new printing options to be 7231 set. 7232 [Steve Henson] 7233 7234 *) Clean old EAY MD5 hack from e_os.h. 7235 [Richard Levitte] 7236 7237 Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 7238 7239 *) Fix null-pointer assignment in do_change_cipher_spec() revealed 7240 by using the Codenomicon TLS Test Tool (CVE-2004-0079) 7241 [Joe Orton, Steve Henson] 7242 7243 Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 7244 7245 *) Fix additional bug revealed by the NISCC test suite: 7246 7247 Stop bug triggering large recursion when presented with 7248 certain ASN.1 tags (CVE-2003-0851) 7249 [Steve Henson] 7250 7251 Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 7252 7253 *) Fix various bugs revealed by running the NISCC test suite: 7254 7255 Stop out of bounds reads in the ASN1 code when presented with 7256 invalid tags (CVE-2003-0543 and CVE-2003-0544). 7257 7258 If verify callback ignores invalid public key errors don't try to check 7259 certificate signature with the NULL public key. 7260 7261 [Steve Henson] 7262 7263 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 7264 if the server requested one: as stated in TLS 1.0 and SSL 3.0 7265 specifications. 7266 [Steve Henson] 7267 7268 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 7269 extra data after the compression methods not only for TLS 1.0 7270 but also for SSL 3.0 (as required by the specification). 7271 [Bodo Moeller; problem pointed out by Matthias Loepfe] 7272 7273 *) Change X509_certificate_type() to mark the key as exported/exportable 7274 when it's 512 *bits* long, not 512 bytes. 7275 [Richard Levitte] 7276 7277 Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 7278 7279 *) Countermeasure against the Klima-Pokorny-Rosa extension of 7280 Bleichbacher's attack on PKCS #1 v1.5 padding: treat 7281 a protocol version number mismatch like a decryption error 7282 in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 7283 [Bodo Moeller] 7284 7285 *) Turn on RSA blinding by default in the default implementation 7286 to avoid a timing attack. Applications that don't want it can call 7287 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 7288 They would be ill-advised to do so in most cases. 7289 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 7290 7291 *) Change RSA blinding code so that it works when the PRNG is not 7292 seeded (in this case, the secret RSA exponent is abused as 7293 an unpredictable seed -- if it is not unpredictable, there 7294 is no point in blinding anyway). Make RSA blinding thread-safe 7295 by remembering the creator's thread ID in rsa->blinding and 7296 having all other threads use local one-time blinding factors 7297 (this requires more computation than sharing rsa->blinding, but 7298 avoids excessive locking; and if an RSA object is not shared 7299 between threads, blinding will still be very fast). 7300 [Bodo Moeller] 7301 7302 Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 7303 7304 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 7305 via timing by performing a MAC computation even if incorrrect 7306 block cipher padding has been found. This is a countermeasure 7307 against active attacks where the attacker has to distinguish 7308 between bad padding and a MAC verification error. (CVE-2003-0078) 7309 7310 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 7311 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 7312 Martin Vuagnoux (EPFL, Ilion)] 7313 7314 Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 7315 7316 *) New function OPENSSL_cleanse(), which is used to cleanse a section of 7317 memory from it's contents. This is done with a counter that will 7318 place alternating values in each byte. This can be used to solve 7319 two issues: 1) the removal of calls to memset() by highly optimizing 7320 compilers, and 2) cleansing with other values than 0, since those can 7321 be read through on certain media, for example a swap space on disk. 7322 [Geoff Thorpe] 7323 7324 *) Bugfix: client side session caching did not work with external caching, 7325 because the session->cipher setting was not restored when reloading 7326 from the external cache. This problem was masked, when 7327 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 7328 (Found by Steve Haslam <steve@araqnid.ddts.net>.) 7329 [Lutz Jaenicke] 7330 7331 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 7332 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 7333 [Zeev Lieber <zeev-l@yahoo.com>] 7334 7335 *) Undo an undocumented change introduced in 0.9.6e which caused 7336 repeated calls to OpenSSL_add_all_ciphers() and 7337 OpenSSL_add_all_digests() to be ignored, even after calling 7338 EVP_cleanup(). 7339 [Richard Levitte] 7340 7341 *) Change the default configuration reader to deal with last line not 7342 being properly terminated. 7343 [Richard Levitte] 7344 7345 *) Change X509_NAME_cmp() so it applies the special rules on handling 7346 DN values that are of type PrintableString, as well as RDNs of type 7347 emailAddress where the value has the type ia5String. 7348 [stefank@valicert.com via Richard Levitte] 7349 7350 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 7351 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 7352 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 7353 the bitwise-OR of the two for use by the majority of applications 7354 wanting this behaviour, and update the docs. The documented 7355 behaviour and actual behaviour were inconsistent and had been 7356 changing anyway, so this is more a bug-fix than a behavioural 7357 change. 7358 [Geoff Thorpe, diagnosed by Nadav Har'El] 7359 7360 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 7361 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 7362 [Bodo Moeller] 7363 7364 *) Fix initialization code race conditions in 7365 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 7366 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 7367 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 7368 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 7369 ssl2_get_cipher_by_char(), 7370 ssl3_get_cipher_by_char(). 7371 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 7372 7373 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 7374 the cached sessions are flushed, as the remove_cb() might use ex_data 7375 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 7376 (see [openssl.org #212]). 7377 [Geoff Thorpe, Lutz Jaenicke] 7378 7379 *) Fix typo in OBJ_txt2obj which incorrectly passed the content 7380 length, instead of the encoding length to d2i_ASN1_OBJECT. 7381 [Steve Henson] 7382 7383 Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 7384 7385 *) [In 0.9.6g-engine release:] 7386 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 7387 [Lynn Gazis <lgazis@rainbow.com>] 7388 7389 Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 7390 7391 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 7392 and get fix the header length calculation. 7393 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 7394 Alon Kantor <alonk@checkpoint.com> (and others), 7395 Steve Henson] 7396 7397 *) Use proper error handling instead of 'assertions' in buffer 7398 overflow checks added in 0.9.6e. This prevents DoS (the 7399 assertions could call abort()). 7400 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 7401 7402 Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 7403 7404 *) Add various sanity checks to asn1_get_length() to reject 7405 the ASN1 length bytes if they exceed sizeof(long), will appear 7406 negative or the content length exceeds the length of the 7407 supplied buffer. 7408 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 7409 7410 *) Fix cipher selection routines: ciphers without encryption had no flags 7411 for the cipher strength set and where therefore not handled correctly 7412 by the selection routines (PR #130). 7413 [Lutz Jaenicke] 7414 7415 *) Fix EVP_dsa_sha macro. 7416 [Nils Larsch] 7417 7418 *) New option 7419 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 7420 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 7421 that was added in OpenSSL 0.9.6d. 7422 7423 As the countermeasure turned out to be incompatible with some 7424 broken SSL implementations, the new option is part of SSL_OP_ALL. 7425 SSL_OP_ALL is usually employed when compatibility with weird SSL 7426 implementations is desired (e.g. '-bugs' option to 's_client' and 7427 's_server'), so the new option is automatically set in many 7428 applications. 7429 [Bodo Moeller] 7430 7431 *) Changes in security patch: 7432 7433 Changes marked "(CHATS)" were sponsored by the Defense Advanced 7434 Research Projects Agency (DARPA) and Air Force Research Laboratory, 7435 Air Force Materiel Command, USAF, under agreement number 7436 F30602-01-2-0537. 7437 7438 *) Add various sanity checks to asn1_get_length() to reject 7439 the ASN1 length bytes if they exceed sizeof(long), will appear 7440 negative or the content length exceeds the length of the 7441 supplied buffer. (CVE-2002-0659) 7442 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 7443 7444 *) Assertions for various potential buffer overflows, not known to 7445 happen in practice. 7446 [Ben Laurie (CHATS)] 7447 7448 *) Various temporary buffers to hold ASCII versions of integers were 7449 too small for 64 bit platforms. (CVE-2002-0655) 7450 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 7451 7452 *) Remote buffer overflow in SSL3 protocol - an attacker could 7453 supply an oversized session ID to a client. (CVE-2002-0656) 7454 [Ben Laurie (CHATS)] 7455 7456 *) Remote buffer overflow in SSL2 protocol - an attacker could 7457 supply an oversized client master key. (CVE-2002-0656) 7458 [Ben Laurie (CHATS)] 7459 7460 Changes between 0.9.6c and 0.9.6d [9 May 2002] 7461 7462 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 7463 encoded as NULL) with id-dsa-with-sha1. 7464 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 7465 7466 *) Check various X509_...() return values in apps/req.c. 7467 [Nils Larsch <nla@trustcenter.de>] 7468 7469 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 7470 an end-of-file condition would erronously be flagged, when the CRLF 7471 was just at the end of a processed block. The bug was discovered when 7472 processing data through a buffering memory BIO handing the data to a 7473 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 7474 <ptsekov@syntrex.com> and Nedelcho Stanev. 7475 [Lutz Jaenicke] 7476 7477 *) Implement a countermeasure against a vulnerability recently found 7478 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 7479 before application data chunks to avoid the use of known IVs 7480 with data potentially chosen by the attacker. 7481 [Bodo Moeller] 7482 7483 *) Fix length checks in ssl3_get_client_hello(). 7484 [Bodo Moeller] 7485 7486 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 7487 to prevent ssl3_read_internal() from incorrectly assuming that 7488 ssl3_read_bytes() found application data while handshake 7489 processing was enabled when in fact s->s3->in_read_app_data was 7490 merely automatically cleared during the initial handshake. 7491 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 7492 7493 *) Fix object definitions for Private and Enterprise: they were not 7494 recognized in their shortname (=lowercase) representation. Extend 7495 obj_dat.pl to issue an error when using undefined keywords instead 7496 of silently ignoring the problem (Svenning Sorensen 7497 <sss@sss.dnsalias.net>). 7498 [Lutz Jaenicke] 7499 7500 *) Fix DH_generate_parameters() so that it works for 'non-standard' 7501 generators, i.e. generators other than 2 and 5. (Previously, the 7502 code did not properly initialise the 'add' and 'rem' values to 7503 BN_generate_prime().) 7504 7505 In the new general case, we do not insist that 'generator' is 7506 actually a primitive root: This requirement is rather pointless; 7507 a generator of the order-q subgroup is just as good, if not 7508 better. 7509 [Bodo Moeller] 7510 7511 *) Map new X509 verification errors to alerts. Discovered and submitted by 7512 Tom Wu <tom@arcot.com>. 7513 [Lutz Jaenicke] 7514 7515 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 7516 returning non-zero before the data has been completely received 7517 when using non-blocking I/O. 7518 [Bodo Moeller; problem pointed out by John Hughes] 7519 7520 *) Some of the ciphers missed the strength entry (SSL_LOW etc). 7521 [Ben Laurie, Lutz Jaenicke] 7522 7523 *) Fix bug in SSL_clear(): bad sessions were not removed (found by 7524 Yoram Zahavi <YoramZ@gilian.com>). 7525 [Lutz Jaenicke] 7526 7527 *) Add information about CygWin 1.3 and on, and preserve proper 7528 configuration for the versions before that. 7529 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 7530 7531 *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 7532 check whether we deal with a copy of a session and do not delete from 7533 the cache in this case. Problem reported by "Izhar Shoshani Levi" 7534 <izhar@checkpoint.com>. 7535 [Lutz Jaenicke] 7536 7537 *) Do not store session data into the internal session cache, if it 7538 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 7539 flag is set). Proposed by Aslam <aslam@funk.com>. 7540 [Lutz Jaenicke] 7541 7542 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 7543 value is 0. 7544 [Richard Levitte] 7545 7546 *) [In 0.9.6d-engine release:] 7547 Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 7548 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 7549 7550 *) Add the configuration target linux-s390x. 7551 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 7552 7553 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 7554 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 7555 variable as an indication that a ClientHello message has been 7556 received. As the flag value will be lost between multiple 7557 invocations of ssl3_accept when using non-blocking I/O, the 7558 function may not be aware that a handshake has actually taken 7559 place, thus preventing a new session from being added to the 7560 session cache. 7561 7562 To avoid this problem, we now set s->new_session to 2 instead of 7563 using a local variable. 7564 [Lutz Jaenicke, Bodo Moeller] 7565 7566 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 7567 if the SSL_R_LENGTH_MISMATCH error is detected. 7568 [Geoff Thorpe, Bodo Moeller] 7569 7570 *) New 'shared_ldflag' column in Configure platform table. 7571 [Richard Levitte] 7572 7573 *) Fix EVP_CIPHER_mode macro. 7574 ["Dan S. Camper" <dan@bti.net>] 7575 7576 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 7577 type, we must throw them away by setting rr->length to 0. 7578 [D P Chang <dpc@qualys.com>] 7579 7580 Changes between 0.9.6b and 0.9.6c [21 dec 2001] 7581 7582 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 7583 <Dominikus.Scherkl@biodata.com>. (The previous implementation 7584 worked incorrectly for those cases where range = 10..._2 and 7585 3*range is two bits longer than range.) 7586 [Bodo Moeller] 7587 7588 *) Only add signing time to PKCS7 structures if it is not already 7589 present. 7590 [Steve Henson] 7591 7592 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 7593 OBJ_ld_ce should be OBJ_id_ce. 7594 Also some ip-pda OIDs in crypto/objects/objects.txt were 7595 incorrect (cf. RFC 3039). 7596 [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 7597 7598 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 7599 returns early because it has nothing to do. 7600 [Andy Schneider <andy.schneider@bjss.co.uk>] 7601 7602 *) [In 0.9.6c-engine release:] 7603 Fix mutex callback return values in crypto/engine/hw_ncipher.c. 7604 [Andy Schneider <andy.schneider@bjss.co.uk>] 7605 7606 *) [In 0.9.6c-engine release:] 7607 Add support for Cryptographic Appliance's keyserver technology. 7608 (Use engine 'keyclient') 7609 [Cryptographic Appliances and Geoff Thorpe] 7610 7611 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 7612 is called via tools/c89.sh because arguments have to be 7613 rearranged (all '-L' options must appear before the first object 7614 modules). 7615 [Richard Shapiro <rshapiro@abinitio.com>] 7616 7617 *) [In 0.9.6c-engine release:] 7618 Add support for Broadcom crypto accelerator cards, backported 7619 from 0.9.7. 7620 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 7621 7622 *) [In 0.9.6c-engine release:] 7623 Add support for SureWare crypto accelerator cards from 7624 Baltimore Technologies. (Use engine 'sureware') 7625 [Baltimore Technologies and Mark Cox] 7626 7627 *) [In 0.9.6c-engine release:] 7628 Add support for crypto accelerator cards from Accelerated 7629 Encryption Processing, www.aep.ie. (Use engine 'aep') 7630 [AEP Inc. and Mark Cox] 7631 7632 *) Add a configuration entry for gcc on UnixWare. 7633 [Gary Benson <gbenson@redhat.com>] 7634 7635 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 7636 messages are stored in a single piece (fixed-length part and 7637 variable-length part combined) and fix various bugs found on the way. 7638 [Bodo Moeller] 7639 7640 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 7641 instead. BIO_gethostbyname() does not know what timeouts are 7642 appropriate, so entries would stay in cache even when they have 7643 become invalid. 7644 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 7645 7646 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 7647 faced with a pathologically small ClientHello fragment that does 7648 not contain client_version: Instead of aborting with an error, 7649 simply choose the highest available protocol version (i.e., 7650 TLS 1.0 unless it is disabled). In practice, ClientHello 7651 messages are never sent like this, but this change gives us 7652 strictly correct behaviour at least for TLS. 7653 [Bodo Moeller] 7654 7655 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 7656 never resets s->method to s->ctx->method when called from within 7657 one of the SSL handshake functions. 7658 [Bodo Moeller; problem pointed out by Niko Baric] 7659 7660 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 7661 (sent using the client's version number) if client_version is 7662 smaller than the protocol version in use. Also change 7663 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 7664 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 7665 the client will at least see that alert. 7666 [Bodo Moeller] 7667 7668 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 7669 correctly. 7670 [Bodo Moeller] 7671 7672 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 7673 client receives HelloRequest while in a handshake. 7674 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 7675 7676 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 7677 should end in 'break', not 'goto end' which circuments various 7678 cleanups done in state SSL_ST_OK. But session related stuff 7679 must be disabled for SSL_ST_OK in the case that we just sent a 7680 HelloRequest. 7681 7682 Also avoid some overhead by not calling ssl_init_wbio_buffer() 7683 before just sending a HelloRequest. 7684 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 7685 7686 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 7687 reveal whether illegal block cipher padding was found or a MAC 7688 verification error occured. (Neither SSLerr() codes nor alerts 7689 are directly visible to potential attackers, but the information 7690 may leak via logfiles.) 7691 7692 Similar changes are not required for the SSL 2.0 implementation 7693 because the number of padding bytes is sent in clear for SSL 2.0, 7694 and the extra bytes are just ignored. However ssl/s2_pkt.c 7695 failed to verify that the purported number of padding bytes is in 7696 the legal range. 7697 [Bodo Moeller] 7698 7699 *) Add OpenUNIX-8 support including shared libraries 7700 (Boyd Lynn Gerber <gerberb@zenez.com>). 7701 [Lutz Jaenicke] 7702 7703 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 7704 'wristwatch attack' using huge encoding parameters (cf. 7705 James H. Manger's CRYPTO 2001 paper). Note that the 7706 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 7707 encoding parameters and hence was not vulnerable. 7708 [Bodo Moeller] 7709 7710 *) BN_sqr() bug fix. 7711 [Ulf M��ller, reported by Jim Ellis <jim.ellis@cavium.com>] 7712 7713 *) Rabin-Miller test analyses assume uniformly distributed witnesses, 7714 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 7715 followed by modular reduction. 7716 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 7717 7718 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 7719 equivalent based on BN_pseudo_rand() instead of BN_rand(). 7720 [Bodo Moeller] 7721 7722 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 7723 This function was broken, as the check for a new client hello message 7724 to handle SGC did not allow these large messages. 7725 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 7726 [Lutz Jaenicke] 7727 7728 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 7729 [Lutz Jaenicke] 7730 7731 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 7732 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 7733 [Lutz Jaenicke] 7734 7735 *) Rework the configuration and shared library support for Tru64 Unix. 7736 The configuration part makes use of modern compiler features and 7737 still retains old compiler behavior for those that run older versions 7738 of the OS. The shared library support part includes a variant that 7739 uses the RPATH feature, and is available through the special 7740 configuration target "alpha-cc-rpath", which will never be selected 7741 automatically. 7742 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 7743 7744 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 7745 with the same message size as in ssl3_get_certificate_request(). 7746 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 7747 messages might inadvertently be reject as too long. 7748 [Petr Lampa <lampa@fee.vutbr.cz>] 7749 7750 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 7751 [Andy Polyakov] 7752 7753 *) Modified SSL library such that the verify_callback that has been set 7754 specificly for an SSL object with SSL_set_verify() is actually being 7755 used. Before the change, a verify_callback set with this function was 7756 ignored and the verify_callback() set in the SSL_CTX at the time of 7757 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 7758 to allow the necessary settings. 7759 [Lutz Jaenicke] 7760 7761 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 7762 explicitly to NULL, as at least on Solaris 8 this seems not always to be 7763 done automatically (in contradiction to the requirements of the C 7764 standard). This made problems when used from OpenSSH. 7765 [Lutz Jaenicke] 7766 7767 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 7768 dh->length and always used 7769 7770 BN_rand_range(priv_key, dh->p). 7771 7772 BN_rand_range() is not necessary for Diffie-Hellman, and this 7773 specific range makes Diffie-Hellman unnecessarily inefficient if 7774 dh->length (recommended exponent length) is much smaller than the 7775 length of dh->p. We could use BN_rand_range() if the order of 7776 the subgroup was stored in the DH structure, but we only have 7777 dh->length. 7778 7779 So switch back to 7780 7781 BN_rand(priv_key, l, ...) 7782 7783 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 7784 otherwise. 7785 [Bodo Moeller] 7786 7787 *) In 7788 7789 RSA_eay_public_encrypt 7790 RSA_eay_private_decrypt 7791 RSA_eay_private_encrypt (signing) 7792 RSA_eay_public_decrypt (signature verification) 7793 7794 (default implementations for RSA_public_encrypt, 7795 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 7796 always reject numbers >= n. 7797 [Bodo Moeller] 7798 7799 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 7800 to synchronize access to 'locking_thread'. This is necessary on 7801 systems where access to 'locking_thread' (an 'unsigned long' 7802 variable) is not atomic. 7803 [Bodo Moeller] 7804 7805 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 7806 *before* setting the 'crypto_lock_rand' flag. The previous code had 7807 a race condition if 0 is a valid thread ID. 7808 [Travis Vitek <vitek@roguewave.com>] 7809 7810 *) Add support for shared libraries under Irix. 7811 [Albert Chin-A-Young <china@thewrittenword.com>] 7812 7813 *) Add configuration option to build on Linux on both big-endian and 7814 little-endian MIPS. 7815 [Ralf Baechle <ralf@uni-koblenz.de>] 7816 7817 *) Add the possibility to create shared libraries on HP-UX. 7818 [Richard Levitte] 7819 7820 Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 7821 7822 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 7823 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 7824 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 7825 PRNG state recovery was possible based on the output of 7826 one PRNG request appropriately sized to gain knowledge on 7827 'md' followed by enough consecutive 1-byte PRNG requests 7828 to traverse all of 'state'. 7829 7830 1. When updating 'md_local' (the current thread's copy of 'md') 7831 during PRNG output generation, hash all of the previous 7832 'md_local' value, not just the half used for PRNG output. 7833 7834 2. Make the number of bytes from 'state' included into the hash 7835 independent from the number of PRNG bytes requested. 7836 7837 The first measure alone would be sufficient to avoid 7838 Markku-Juhani's attack. (Actually it had never occurred 7839 to me that the half of 'md_local' used for chaining was the 7840 half from which PRNG output bytes were taken -- I had always 7841 assumed that the secret half would be used.) The second 7842 measure makes sure that additional data from 'state' is never 7843 mixed into 'md_local' in small portions; this heuristically 7844 further strengthens the PRNG. 7845 [Bodo Moeller] 7846 7847 *) Fix crypto/bn/asm/mips3.s. 7848 [Andy Polyakov] 7849 7850 *) When only the key is given to "enc", the IV is undefined. Print out 7851 an error message in this case. 7852 [Lutz Jaenicke] 7853 7854 *) Handle special case when X509_NAME is empty in X509 printing routines. 7855 [Steve Henson] 7856 7857 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 7858 positive and less than q. 7859 [Bodo Moeller] 7860 7861 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 7862 used: it isn't thread safe and the add_lock_callback should handle 7863 that itself. 7864 [Paul Rose <Paul.Rose@bridge.com>] 7865 7866 *) Verify that incoming data obeys the block size in 7867 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 7868 [Bodo Moeller] 7869 7870 *) Fix OAEP check. 7871 [Ulf M��ller, Bodo M��ller] 7872 7873 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 7874 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 7875 when fixing the server behaviour for backwards-compatible 'client 7876 hello' messages. (Note that the attack is impractical against 7877 SSL 3.0 and TLS 1.0 anyway because length and version checking 7878 means that the probability of guessing a valid ciphertext is 7879 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 7880 paper.) 7881 7882 Before 0.9.5, the countermeasure (hide the error by generating a 7883 random 'decryption result') did not work properly because 7884 ERR_clear_error() was missing, meaning that SSL_get_error() would 7885 detect the supposedly ignored error. 7886 7887 Both problems are now fixed. 7888 [Bodo Moeller] 7889 7890 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 7891 (previously it was 1024). 7892 [Bodo Moeller] 7893 7894 *) Fix for compatibility mode trust settings: ignore trust settings 7895 unless some valid trust or reject settings are present. 7896 [Steve Henson] 7897 7898 *) Fix for blowfish EVP: its a variable length cipher. 7899 [Steve Henson] 7900 7901 *) Fix various bugs related to DSA S/MIME verification. Handle missing 7902 parameters in DSA public key structures and return an error in the 7903 DSA routines if parameters are absent. 7904 [Steve Henson] 7905 7906 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 7907 in the current directory if neither $RANDFILE nor $HOME was set. 7908 RAND_file_name() in 0.9.6a returned NULL in this case. This has 7909 caused some confusion to Windows users who haven't defined $HOME. 7910 Thus RAND_file_name() is changed again: e_os.h can define a 7911 DEFAULT_HOME, which will be used if $HOME is not set. 7912 For Windows, we use "C:"; on other platforms, we still require 7913 environment variables. 7914 7915 *) Move 'if (!initialized) RAND_poll()' into regions protected by 7916 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 7917 having multiple threads call RAND_poll() concurrently. 7918 [Bodo Moeller] 7919 7920 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 7921 combination of a flag and a thread ID variable. 7922 Otherwise while one thread is in ssleay_rand_bytes (which sets the 7923 flag), *other* threads can enter ssleay_add_bytes without obeying 7924 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 7925 that they do not hold after the first thread unsets add_do_not_lock). 7926 [Bodo Moeller] 7927 7928 *) Change bctest again: '-x' expressions are not available in all 7929 versions of 'test'. 7930 [Bodo Moeller] 7931 7932 Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 7933 7934 *) Fix a couple of memory leaks in PKCS7_dataDecode() 7935 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 7936 7937 *) Change Configure and Makefiles to provide EXE_EXT, which will contain 7938 the default extension for executables, if any. Also, make the perl 7939 scripts that use symlink() to test if it really exists and use "cp" 7940 if it doesn't. All this made OpenSSL compilable and installable in 7941 CygWin. 7942 [Richard Levitte] 7943 7944 *) Fix for asn1_GetSequence() for indefinite length constructed data. 7945 If SEQUENCE is length is indefinite just set c->slen to the total 7946 amount of data available. 7947 [Steve Henson, reported by shige@FreeBSD.org] 7948 [This change does not apply to 0.9.7.] 7949 7950 *) Change bctest to avoid here-documents inside command substitution 7951 (workaround for FreeBSD /bin/sh bug). 7952 For compatibility with Ultrix, avoid shell functions (introduced 7953 in the bctest version that searches along $PATH). 7954 [Bodo Moeller] 7955 7956 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 7957 with des_encrypt() defined on some operating systems, like Solaris 7958 and UnixWare. 7959 [Richard Levitte] 7960 7961 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 7962 On the Importance of Eliminating Errors in Cryptographic 7963 Computations, J. Cryptology 14 (2001) 2, 101-119, 7964 http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 7965 [Ulf Moeller] 7966 7967 *) MIPS assembler BIGNUM division bug fix. 7968 [Andy Polyakov] 7969 7970 *) Disabled incorrect Alpha assembler code. 7971 [Richard Levitte] 7972 7973 *) Fix PKCS#7 decode routines so they correctly update the length 7974 after reading an EOC for the EXPLICIT tag. 7975 [Steve Henson] 7976 [This change does not apply to 0.9.7.] 7977 7978 *) Fix bug in PKCS#12 key generation routines. This was triggered 7979 if a 3DES key was generated with a 0 initial byte. Include 7980 PKCS12_BROKEN_KEYGEN compilation option to retain the old 7981 (but broken) behaviour. 7982 [Steve Henson] 7983 7984 *) Enhance bctest to search for a working bc along $PATH and print 7985 it when found. 7986 [Tim Rice <tim@multitalents.net> via Richard Levitte] 7987 7988 *) Fix memory leaks in err.c: free err_data string if necessary; 7989 don't write to the wrong index in ERR_set_error_data. 7990 [Bodo Moeller] 7991 7992 *) Implement ssl23_peek (analogous to ssl23_read), which previously 7993 did not exist. 7994 [Bodo Moeller] 7995 7996 *) Replace rdtsc with _emit statements for VC++ version 5. 7997 [Jeremy Cooper <jeremy@baymoo.org>] 7998 7999 *) Make it possible to reuse SSLv2 sessions. 8000 [Richard Levitte] 8001 8002 *) In copy_email() check for >= 0 as a return value for 8003 X509_NAME_get_index_by_NID() since 0 is a valid index. 8004 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 8005 8006 *) Avoid coredump with unsupported or invalid public keys by checking if 8007 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 8008 PKCS7_verify() fails with non detached data. 8009 [Steve Henson] 8010 8011 *) Don't use getenv in library functions when run as setuid/setgid. 8012 New function OPENSSL_issetugid(). 8013 [Ulf Moeller] 8014 8015 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 8016 due to incorrect handling of multi-threading: 8017 8018 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 8019 8020 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 8021 8022 3. Count how many times MemCheck_off() has been called so that 8023 nested use can be treated correctly. This also avoids 8024 inband-signalling in the previous code (which relied on the 8025 assumption that thread ID 0 is impossible). 8026 [Bodo Moeller] 8027 8028 *) Add "-rand" option also to s_client and s_server. 8029 [Lutz Jaenicke] 8030 8031 *) Fix CPU detection on Irix 6.x. 8032 [Kurt Hockenbury <khockenb@stevens-tech.edu> and 8033 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 8034 8035 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 8036 was empty. 8037 [Steve Henson] 8038 [This change does not apply to 0.9.7.] 8039 8040 *) Use the cached encoding of an X509_NAME structure rather than 8041 copying it. This is apparently the reason for the libsafe "errors" 8042 but the code is actually correct. 8043 [Steve Henson] 8044 8045 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 8046 Bleichenbacher's DSA attack. 8047 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 8048 to be set and top=0 forces the highest bit to be set; top=-1 is new 8049 and leaves the highest bit random. 8050 [Ulf Moeller, Bodo Moeller] 8051 8052 *) In the NCONF_...-based implementations for CONF_... queries 8053 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 8054 a temporary CONF structure with the data component set to NULL 8055 (which gives segmentation faults in lh_retrieve). 8056 Instead, use NULL for the CONF pointer in CONF_get_string and 8057 CONF_get_number (which may use environment variables) and directly 8058 return NULL from CONF_get_section. 8059 [Bodo Moeller] 8060 8061 *) Fix potential buffer overrun for EBCDIC. 8062 [Ulf Moeller] 8063 8064 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 8065 keyUsage if basicConstraints absent for a CA. 8066 [Steve Henson] 8067 8068 *) Make SMIME_write_PKCS7() write mail header values with a format that 8069 is more generally accepted (no spaces before the semicolon), since 8070 some programs can't parse those values properly otherwise. Also make 8071 sure BIO's that break lines after each write do not create invalid 8072 headers. 8073 [Richard Levitte] 8074 8075 *) Make the CRL encoding routines work with empty SEQUENCE OF. The 8076 macros previously used would not encode an empty SEQUENCE OF 8077 and break the signature. 8078 [Steve Henson] 8079 [This change does not apply to 0.9.7.] 8080 8081 *) Zero the premaster secret after deriving the master secret in 8082 DH ciphersuites. 8083 [Steve Henson] 8084 8085 *) Add some EVP_add_digest_alias registrations (as found in 8086 OpenSSL_add_all_digests()) to SSL_library_init() 8087 aka OpenSSL_add_ssl_algorithms(). This provides improved 8088 compatibility with peers using X.509 certificates 8089 with unconventional AlgorithmIdentifier OIDs. 8090 [Bodo Moeller] 8091 8092 *) Fix for Irix with NO_ASM. 8093 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 8094 8095 *) ./config script fixes. 8096 [Ulf Moeller, Richard Levitte] 8097 8098 *) Fix 'openssl passwd -1'. 8099 [Bodo Moeller] 8100 8101 *) Change PKCS12_key_gen_asc() so it can cope with non null 8102 terminated strings whose length is passed in the passlen 8103 parameter, for example from PEM callbacks. This was done 8104 by adding an extra length parameter to asc2uni(). 8105 [Steve Henson, reported by <oddissey@samsung.co.kr>] 8106 8107 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 8108 call failed, free the DSA structure. 8109 [Bodo Moeller] 8110 8111 *) Fix to uni2asc() to cope with zero length Unicode strings. 8112 These are present in some PKCS#12 files. 8113 [Steve Henson] 8114 8115 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 8116 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 8117 when writing a 32767 byte record. 8118 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 8119 8120 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 8121 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 8122 8123 (RSA objects have a reference count access to which is protected 8124 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 8125 so they are meant to be shared between threads.) 8126 [Bodo Moeller, Geoff Thorpe; original patch submitted by 8127 "Reddie, Steven" <Steven.Reddie@ca.com>] 8128 8129 *) Fix a deadlock in CRYPTO_mem_leaks(). 8130 [Bodo Moeller] 8131 8132 *) Use better test patterns in bntest. 8133 [Ulf M��ller] 8134 8135 *) rand_win.c fix for Borland C. 8136 [Ulf M��ller] 8137 8138 *) BN_rshift bugfix for n == 0. 8139 [Bodo Moeller] 8140 8141 *) Add a 'bctest' script that checks for some known 'bc' bugs 8142 so that 'make test' does not abort just because 'bc' is broken. 8143 [Bodo Moeller] 8144 8145 *) Store verify_result within SSL_SESSION also for client side to 8146 avoid potential security hole. (Re-used sessions on the client side 8147 always resulted in verify_result==X509_V_OK, not using the original 8148 result of the server certificate verification.) 8149 [Lutz Jaenicke] 8150 8151 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 8152 SSL3_RT_APPLICATION_DATA, return 0. 8153 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 8154 [Bodo Moeller] 8155 8156 *) Fix SSL_peek: 8157 Both ssl2_peek and ssl3_peek, which were totally broken in earlier 8158 releases, have been re-implemented by renaming the previous 8159 implementations of ssl2_read and ssl3_read to ssl2_read_internal 8160 and ssl3_read_internal, respectively, and adding 'peek' parameters 8161 to them. The new ssl[23]_{read,peek} functions are calls to 8162 ssl[23]_read_internal with the 'peek' flag set appropriately. 8163 A 'peek' parameter has also been added to ssl3_read_bytes, which 8164 does the actual work for ssl3_read_internal. 8165 [Bodo Moeller] 8166 8167 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 8168 the method-specific "init()" handler. Also clean up ex_data after 8169 calling the method-specific "finish()" handler. Previously, this was 8170 happening the other way round. 8171 [Geoff Thorpe] 8172 8173 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 8174 The previous value, 12, was not always sufficient for BN_mod_exp(). 8175 [Bodo Moeller] 8176 8177 *) Make sure that shared libraries get the internal name engine with 8178 the full version number and not just 0. This should mark the 8179 shared libraries as not backward compatible. Of course, this should 8180 be changed again when we can guarantee backward binary compatibility. 8181 [Richard Levitte] 8182 8183 *) Fix typo in get_cert_by_subject() in by_dir.c 8184 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 8185 8186 *) Rework the system to generate shared libraries: 8187 8188 - Make note of the expected extension for the shared libraries and 8189 if there is a need for symbolic links from for example libcrypto.so.0 8190 to libcrypto.so.0.9.7. There is extended info in Configure for 8191 that. 8192 8193 - Make as few rebuilds of the shared libraries as possible. 8194 8195 - Still avoid linking the OpenSSL programs with the shared libraries. 8196 8197 - When installing, install the shared libraries separately from the 8198 static ones. 8199 [Richard Levitte] 8200 8201 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 8202 8203 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 8204 and not in SSL_clear because the latter is also used by the 8205 accept/connect functions; previously, the settings made by 8206 SSL_set_read_ahead would be lost during the handshake. 8207 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 8208 8209 *) Correct util/mkdef.pl to be selective about disabled algorithms. 8210 Previously, it would create entries for disableed algorithms no 8211 matter what. 8212 [Richard Levitte] 8213 8214 *) Added several new manual pages for SSL_* function. 8215 [Lutz Jaenicke] 8216 8217 Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 8218 8219 *) In ssl23_get_client_hello, generate an error message when faced 8220 with an initial SSL 3.0/TLS record that is too small to contain the 8221 first two bytes of the ClientHello message, i.e. client_version. 8222 (Note that this is a pathologic case that probably has never happened 8223 in real life.) The previous approach was to use the version number 8224 from the record header as a substitute; but our protocol choice 8225 should not depend on that one because it is not authenticated 8226 by the Finished messages. 8227 [Bodo Moeller] 8228 8229 *) More robust randomness gathering functions for Windows. 8230 [Jeffrey Altman <jaltman@columbia.edu>] 8231 8232 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 8233 not set then we don't setup the error code for issuer check errors 8234 to avoid possibly overwriting other errors which the callback does 8235 handle. If an application does set the flag then we assume it knows 8236 what it is doing and can handle the new informational codes 8237 appropriately. 8238 [Steve Henson] 8239 8240 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 8241 a general "ANY" type, as such it should be able to decode anything 8242 including tagged types. However it didn't check the class so it would 8243 wrongly interpret tagged types in the same way as their universal 8244 counterpart and unknown types were just rejected. Changed so that the 8245 tagged and unknown types are handled in the same way as a SEQUENCE: 8246 that is the encoding is stored intact. There is also a new type 8247 "V_ASN1_OTHER" which is used when the class is not universal, in this 8248 case we have no idea what the actual type is so we just lump them all 8249 together. 8250 [Steve Henson] 8251 8252 *) On VMS, stdout may very well lead to a file that is written to 8253 in a record-oriented fashion. That means that every write() will 8254 write a separate record, which will be read separately by the 8255 programs trying to read from it. This can be very confusing. 8256 8257 The solution is to put a BIO filter in the way that will buffer 8258 text until a linefeed is reached, and then write everything a 8259 line at a time, so every record written will be an actual line, 8260 not chunks of lines and not (usually doesn't happen, but I've 8261 seen it once) several lines in one record. BIO_f_linebuffer() is 8262 the answer. 8263 8264 Currently, it's a VMS-only method, because that's where it has 8265 been tested well enough. 8266 [Richard Levitte] 8267 8268 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 8269 it can return incorrect results. 8270 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 8271 but it was in 0.9.6-beta[12].) 8272 [Bodo Moeller] 8273 8274 *) Disable the check for content being present when verifying detached 8275 signatures in pk7_smime.c. Some versions of Netscape (wrongly) 8276 include zero length content when signing messages. 8277 [Steve Henson] 8278 8279 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 8280 BIO_ctrl (for BIO pairs). 8281 [Bodo M��ller] 8282 8283 *) Add DSO method for VMS. 8284 [Richard Levitte] 8285 8286 *) Bug fix: Montgomery multiplication could produce results with the 8287 wrong sign. 8288 [Ulf M��ller] 8289 8290 *) Add RPM specification openssl.spec and modify it to build three 8291 packages. The default package contains applications, application 8292 documentation and run-time libraries. The devel package contains 8293 include files, static libraries and function documentation. The 8294 doc package contains the contents of the doc directory. The original 8295 openssl.spec was provided by Damien Miller <djm@mindrot.org>. 8296 [Richard Levitte] 8297 8298 *) Add a large number of documentation files for many SSL routines. 8299 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 8300 8301 *) Add a configuration entry for Sony News 4. 8302 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 8303 8304 *) Don't set the two most significant bits to one when generating a 8305 random number < q in the DSA library. 8306 [Ulf M��ller] 8307 8308 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 8309 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 8310 the underlying transport is blocking) if a handshake took place. 8311 (The default behaviour is needed by applications such as s_client 8312 and s_server that use select() to determine when to use SSL_read; 8313 but for applications that know in advance when to expect data, it 8314 just makes things more complicated.) 8315 [Bodo Moeller] 8316 8317 *) Add RAND_egd_bytes(), which gives control over the number of bytes read 8318 from EGD. 8319 [Ben Laurie] 8320 8321 *) Add a few more EBCDIC conditionals that make `req' and `x509' 8322 work better on such systems. 8323 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 8324 8325 *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 8326 Update PKCS12_parse() so it copies the friendlyName and the 8327 keyid to the certificates aux info. 8328 [Steve Henson] 8329 8330 *) Fix bug in PKCS7_verify() which caused an infinite loop 8331 if there was more than one signature. 8332 [Sven Uszpelkat <su@celocom.de>] 8333 8334 *) Major change in util/mkdef.pl to include extra information 8335 about each symbol, as well as presentig variables as well 8336 as functions. This change means that there's n more need 8337 to rebuild the .num files when some algorithms are excluded. 8338 [Richard Levitte] 8339 8340 *) Allow the verify time to be set by an application, 8341 rather than always using the current time. 8342 [Steve Henson] 8343 8344 *) Phase 2 verify code reorganisation. The certificate 8345 verify code now looks up an issuer certificate by a 8346 number of criteria: subject name, authority key id 8347 and key usage. It also verifies self signed certificates 8348 by the same criteria. The main comparison function is 8349 X509_check_issued() which performs these checks. 8350 8351 Lot of changes were necessary in order to support this 8352 without completely rewriting the lookup code. 8353 8354 Authority and subject key identifier are now cached. 8355 8356 The LHASH 'certs' is X509_STORE has now been replaced 8357 by a STACK_OF(X509_OBJECT). This is mainly because an 8358 LHASH can't store or retrieve multiple objects with 8359 the same hash value. 8360 8361 As a result various functions (which were all internal 8362 use only) have changed to handle the new X509_STORE 8363 structure. This will break anything that messed round 8364 with X509_STORE internally. 8365 8366 The functions X509_STORE_add_cert() now checks for an 8367 exact match, rather than just subject name. 8368 8369 The X509_STORE API doesn't directly support the retrieval 8370 of multiple certificates matching a given criteria, however 8371 this can be worked round by performing a lookup first 8372 (which will fill the cache with candidate certificates) 8373 and then examining the cache for matches. This is probably 8374 the best we can do without throwing out X509_LOOKUP 8375 entirely (maybe later...). 8376 8377 The X509_VERIFY_CTX structure has been enhanced considerably. 8378 8379 All certificate lookup operations now go via a get_issuer() 8380 callback. Although this currently uses an X509_STORE it 8381 can be replaced by custom lookups. This is a simple way 8382 to bypass the X509_STORE hackery necessary to make this 8383 work and makes it possible to use more efficient techniques 8384 in future. A very simple version which uses a simple 8385 STACK for its trusted certificate store is also provided 8386 using X509_STORE_CTX_trusted_stack(). 8387 8388 The verify_cb() and verify() callbacks now have equivalents 8389 in the X509_STORE_CTX structure. 8390 8391 X509_STORE_CTX also has a 'flags' field which can be used 8392 to customise the verify behaviour. 8393 [Steve Henson] 8394 8395 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 8396 excludes S/MIME capabilities. 8397 [Steve Henson] 8398 8399 *) When a certificate request is read in keep a copy of the 8400 original encoding of the signed data and use it when outputing 8401 again. Signatures then use the original encoding rather than 8402 a decoded, encoded version which may cause problems if the 8403 request is improperly encoded. 8404 [Steve Henson] 8405 8406 *) For consistency with other BIO_puts implementations, call 8407 buffer_write(b, ...) directly in buffer_puts instead of calling 8408 BIO_write(b, ...). 8409 8410 In BIO_puts, increment b->num_write as in BIO_write. 8411 [Peter.Sylvester@EdelWeb.fr] 8412 8413 *) Fix BN_mul_word for the case where the word is 0. (We have to use 8414 BN_zero, we may not return a BIGNUM with an array consisting of 8415 words set to zero.) 8416 [Bodo Moeller] 8417 8418 *) Avoid calling abort() from within the library when problems are 8419 detected, except if preprocessor symbols have been defined 8420 (such as REF_CHECK, BN_DEBUG etc.). 8421 [Bodo Moeller] 8422 8423 *) New openssl application 'rsautl'. This utility can be 8424 used for low level RSA operations. DER public key 8425 BIO/fp routines also added. 8426 [Steve Henson] 8427 8428 *) New Configure entry and patches for compiling on QNX 4. 8429 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 8430 8431 *) A demo state-machine implementation was sponsored by 8432 Nuron (http://www.nuron.com/) and is now available in 8433 demos/state_machine. 8434 [Ben Laurie] 8435 8436 *) New options added to the 'dgst' utility for signature 8437 generation and verification. 8438 [Steve Henson] 8439 8440 *) Unrecognized PKCS#7 content types are now handled via a 8441 catch all ASN1_TYPE structure. This allows unsupported 8442 types to be stored as a "blob" and an application can 8443 encode and decode it manually. 8444 [Steve Henson] 8445 8446 *) Fix various signed/unsigned issues to make a_strex.c 8447 compile under VC++. 8448 [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 8449 8450 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 8451 length if passed a buffer. ASN1_INTEGER_to_BN failed 8452 if passed a NULL BN and its argument was negative. 8453 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 8454 8455 *) Modification to PKCS#7 encoding routines to output definite 8456 length encoding. Since currently the whole structures are in 8457 memory there's not real point in using indefinite length 8458 constructed encoding. However if OpenSSL is compiled with 8459 the flag PKCS7_INDEFINITE_ENCODING the old form is used. 8460 [Steve Henson] 8461 8462 *) Added BIO_vprintf() and BIO_vsnprintf(). 8463 [Richard Levitte] 8464 8465 *) Added more prefixes to parse for in the the strings written 8466 through a logging bio, to cover all the levels that are available 8467 through syslog. The prefixes are now: 8468 8469 PANIC, EMERG, EMR => LOG_EMERG 8470 ALERT, ALR => LOG_ALERT 8471 CRIT, CRI => LOG_CRIT 8472 ERROR, ERR => LOG_ERR 8473 WARNING, WARN, WAR => LOG_WARNING 8474 NOTICE, NOTE, NOT => LOG_NOTICE 8475 INFO, INF => LOG_INFO 8476 DEBUG, DBG => LOG_DEBUG 8477 8478 and as before, if none of those prefixes are present at the 8479 beginning of the string, LOG_ERR is chosen. 8480 8481 On Win32, the LOG_* levels are mapped according to this: 8482 8483 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 8484 LOG_WARNING => EVENTLOG_WARNING_TYPE 8485 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 8486 8487 [Richard Levitte] 8488 8489 *) Made it possible to reconfigure with just the configuration 8490 argument "reconf" or "reconfigure". The command line arguments 8491 are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 8492 and are retrieved from there when reconfiguring. 8493 [Richard Levitte] 8494 8495 *) MD4 implemented. 8496 [Assar Westerlund <assar@sics.se>, Richard Levitte] 8497 8498 *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 8499 [Richard Levitte] 8500 8501 *) The obj_dat.pl script was messing up the sorting of object 8502 names. The reason was that it compared the quoted version 8503 of strings as a result "OCSP" > "OCSP Signing" because 8504 " > SPACE. Changed script to store unquoted versions of 8505 names and add quotes on output. It was also omitting some 8506 names from the lookup table if they were given a default 8507 value (that is if SN is missing it is given the same 8508 value as LN and vice versa), these are now added on the 8509 grounds that if an object has a name we should be able to 8510 look it up. Finally added warning output when duplicate 8511 short or long names are found. 8512 [Steve Henson] 8513 8514 *) Changes needed for Tandem NSK. 8515 [Scott Uroff <scott@xypro.com>] 8516 8517 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 8518 RSA_padding_check_SSLv23(), special padding was never detected 8519 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 8520 version rollback attacks was not effective. 8521 8522 In s23_clnt.c, don't use special rollback-attack detection padding 8523 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 8524 client; similarly, in s23_srvr.c, don't do the rollback check if 8525 SSL 2.0 is the only protocol enabled in the server. 8526 [Bodo Moeller] 8527 8528 *) Make it possible to get hexdumps of unprintable data with 'openssl 8529 asn1parse'. By implication, the functions ASN1_parse_dump() and 8530 BIO_dump_indent() are added. 8531 [Richard Levitte] 8532 8533 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 8534 these print out strings and name structures based on various 8535 flags including RFC2253 support and proper handling of 8536 multibyte characters. Added options to the 'x509' utility 8537 to allow the various flags to be set. 8538 [Steve Henson] 8539 8540 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 8541 Also change the functions X509_cmp_current_time() and 8542 X509_gmtime_adj() work with an ASN1_TIME structure, 8543 this will enable certificates using GeneralizedTime in validity 8544 dates to be checked. 8545 [Steve Henson] 8546 8547 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 8548 negative public key encodings) on by default, 8549 NO_NEG_PUBKEY_BUG can be set to disable it. 8550 [Steve Henson] 8551 8552 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 8553 content octets. An i2c_ASN1_OBJECT is unnecessary because 8554 the encoding can be trivially obtained from the structure. 8555 [Steve Henson] 8556 8557 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 8558 not read locks (CRYPTO_r_[un]lock). 8559 [Bodo Moeller] 8560 8561 *) A first attempt at creating official support for shared 8562 libraries through configuration. I've kept it so the 8563 default is static libraries only, and the OpenSSL programs 8564 are always statically linked for now, but there are 8565 preparations for dynamic linking in place. 8566 This has been tested on Linux and Tru64. 8567 [Richard Levitte] 8568 8569 *) Randomness polling function for Win9x, as described in: 8570 Peter Gutmann, Software Generation of Practically Strong 8571 Random Numbers. 8572 [Ulf M��ller] 8573 8574 *) Fix so PRNG is seeded in req if using an already existing 8575 DSA key. 8576 [Steve Henson] 8577 8578 *) New options to smime application. -inform and -outform 8579 allow alternative formats for the S/MIME message including 8580 PEM and DER. The -content option allows the content to be 8581 specified separately. This should allow things like Netscape 8582 form signing output easier to verify. 8583 [Steve Henson] 8584 8585 *) Fix the ASN1 encoding of tags using the 'long form'. 8586 [Steve Henson] 8587 8588 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 8589 STRING types. These convert content octets to and from the 8590 underlying type. The actual tag and length octets are 8591 already assumed to have been read in and checked. These 8592 are needed because all other string types have virtually 8593 identical handling apart from the tag. By having versions 8594 of the ASN1 functions that just operate on content octets 8595 IMPLICIT tagging can be handled properly. It also allows 8596 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 8597 and ASN1_INTEGER are identical apart from the tag. 8598 [Steve Henson] 8599 8600 *) Change the handling of OID objects as follows: 8601 8602 - New object identifiers are inserted in objects.txt, following 8603 the syntax given in objects.README. 8604 - objects.pl is used to process obj_mac.num and create a new 8605 obj_mac.h. 8606 - obj_dat.pl is used to create a new obj_dat.h, using the data in 8607 obj_mac.h. 8608 8609 This is currently kind of a hack, and the perl code in objects.pl 8610 isn't very elegant, but it works as I intended. The simplest way 8611 to check that it worked correctly is to look in obj_dat.h and 8612 check the array nid_objs and make sure the objects haven't moved 8613 around (this is important!). Additions are OK, as well as 8614 consistent name changes. 8615 [Richard Levitte] 8616 8617 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 8618 [Bodo Moeller] 8619 8620 *) Addition of the command line parameter '-rand file' to 'openssl req'. 8621 The given file adds to whatever has already been seeded into the 8622 random pool through the RANDFILE configuration file option or 8623 environment variable, or the default random state file. 8624 [Richard Levitte] 8625 8626 *) mkstack.pl now sorts each macro group into lexical order. 8627 Previously the output order depended on the order the files 8628 appeared in the directory, resulting in needless rewriting 8629 of safestack.h . 8630 [Steve Henson] 8631 8632 *) Patches to make OpenSSL compile under Win32 again. Mostly 8633 work arounds for the VC++ problem that it treats func() as 8634 func(void). Also stripped out the parts of mkdef.pl that 8635 added extra typesafe functions: these no longer exist. 8636 [Steve Henson] 8637 8638 *) Reorganisation of the stack code. The macros are now all 8639 collected in safestack.h . Each macro is defined in terms of 8640 a "stack macro" of the form SKM_<name>(type, a, b). The 8641 DEBUG_SAFESTACK is now handled in terms of function casts, 8642 this has the advantage of retaining type safety without the 8643 use of additional functions. If DEBUG_SAFESTACK is not defined 8644 then the non typesafe macros are used instead. Also modified the 8645 mkstack.pl script to handle the new form. Needs testing to see 8646 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 8647 the default if no major problems. Similar behaviour for ASN1_SET_OF 8648 and PKCS12_STACK_OF. 8649 [Steve Henson] 8650 8651 *) When some versions of IIS use the 'NET' form of private key the 8652 key derivation algorithm is different. Normally MD5(password) is 8653 used as a 128 bit RC4 key. In the modified case 8654 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 8655 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 8656 as the old Netscape_RSA functions except they have an additional 8657 'sgckey' parameter which uses the modified algorithm. Also added 8658 an -sgckey command line option to the rsa utility. Thanks to 8659 Adrian Peck <bertie@ncipher.com> for posting details of the modified 8660 algorithm to openssl-dev. 8661 [Steve Henson] 8662 8663 *) The evp_local.h macros were using 'c.##kname' which resulted in 8664 invalid expansion on some systems (SCO 5.0.5 for example). 8665 Corrected to 'c.kname'. 8666 [Phillip Porch <root@theporch.com>] 8667 8668 *) New X509_get1_email() and X509_REQ_get1_email() functions that return 8669 a STACK of email addresses from a certificate or request, these look 8670 in the subject name and the subject alternative name extensions and 8671 omit any duplicate addresses. 8672 [Steve Henson] 8673 8674 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 8675 This makes DSA verification about 2 % faster. 8676 [Bodo Moeller] 8677 8678 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 8679 (meaning that now 2^5 values will be precomputed, which is only 4 KB 8680 plus overhead for 1024 bit moduli). 8681 This makes exponentiations about 0.5 % faster for 1024 bit 8682 exponents (as measured by "openssl speed rsa2048"). 8683 [Bodo Moeller] 8684 8685 *) Rename memory handling macros to avoid conflicts with other 8686 software: 8687 Malloc => OPENSSL_malloc 8688 Malloc_locked => OPENSSL_malloc_locked 8689 Realloc => OPENSSL_realloc 8690 Free => OPENSSL_free 8691 [Richard Levitte] 8692 8693 *) New function BN_mod_exp_mont_word for small bases (roughly 15% 8694 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 8695 [Bodo Moeller] 8696 8697 *) CygWin32 support. 8698 [John Jarvie <jjarvie@newsguy.com>] 8699 8700 *) The type-safe stack code has been rejigged. It is now only compiled 8701 in when OpenSSL is configured with the DEBUG_SAFESTACK option and 8702 by default all type-specific stack functions are "#define"d back to 8703 standard stack functions. This results in more streamlined output 8704 but retains the type-safety checking possibilities of the original 8705 approach. 8706 [Geoff Thorpe] 8707 8708 *) The STACK code has been cleaned up, and certain type declarations 8709 that didn't make a lot of sense have been brought in line. This has 8710 also involved a cleanup of sorts in safestack.h to more correctly 8711 map type-safe stack functions onto their plain stack counterparts. 8712 This work has also resulted in a variety of "const"ifications of 8713 lots of the code, especially "_cmp" operations which should normally 8714 be prototyped with "const" parameters anyway. 8715 [Geoff Thorpe] 8716 8717 *) When generating bytes for the first time in md_rand.c, 'stir the pool' 8718 by seeding with STATE_SIZE dummy bytes (with zero entropy count). 8719 (The PRNG state consists of two parts, the large pool 'state' and 'md', 8720 where all of 'md' is used each time the PRNG is used, but 'state' 8721 is used only indexed by a cyclic counter. As entropy may not be 8722 well distributed from the beginning, 'md' is important as a 8723 chaining variable. However, the output function chains only half 8724 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 8725 all of 'md', and seeding with STATE_SIZE dummy bytes will result 8726 in all of 'state' being rewritten, with the new values depending 8727 on virtually all of 'md'. This overcomes the 80 bit limitation.) 8728 [Bodo Moeller] 8729 8730 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 8731 the handshake is continued after ssl_verify_cert_chain(); 8732 otherwise, if SSL_VERIFY_NONE is set, remaining error codes 8733 can lead to 'unexplainable' connection aborts later. 8734 [Bodo Moeller; problem tracked down by Lutz Jaenicke] 8735 8736 *) Major EVP API cipher revision. 8737 Add hooks for extra EVP features. This allows various cipher 8738 parameters to be set in the EVP interface. Support added for variable 8739 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 8740 setting of RC2 and RC5 parameters. 8741 8742 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 8743 ciphers. 8744 8745 Remove lots of duplicated code from the EVP library. For example *every* 8746 cipher init() function handles the 'iv' in the same way according to the 8747 cipher mode. They also all do nothing if the 'key' parameter is NULL and 8748 for CFB and OFB modes they zero ctx->num. 8749 8750 New functionality allows removal of S/MIME code RC2 hack. 8751 8752 Most of the routines have the same form and so can be declared in terms 8753 of macros. 8754 8755 By shifting this to the top level EVP_CipherInit() it can be removed from 8756 all individual ciphers. If the cipher wants to handle IVs or keys 8757 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 8758 flags. 8759 8760 Change lots of functions like EVP_EncryptUpdate() to now return a 8761 value: although software versions of the algorithms cannot fail 8762 any installed hardware versions can. 8763 [Steve Henson] 8764 8765 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 8766 this option is set, tolerate broken clients that send the negotiated 8767 protocol version number instead of the requested protocol version 8768 number. 8769 [Bodo Moeller] 8770 8771 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 8772 i.e. non-zero for export ciphersuites, zero otherwise. 8773 Previous versions had this flag inverted, inconsistent with 8774 rsa_tmp_cb (..._TMP_RSA_CB). 8775 [Bodo Moeller; problem reported by Amit Chopra] 8776 8777 *) Add missing DSA library text string. Work around for some IIS 8778 key files with invalid SEQUENCE encoding. 8779 [Steve Henson] 8780 8781 *) Add a document (doc/standards.txt) that list all kinds of standards 8782 and so on that are implemented in OpenSSL. 8783 [Richard Levitte] 8784 8785 *) Enhance c_rehash script. Old version would mishandle certificates 8786 with the same subject name hash and wouldn't handle CRLs at all. 8787 Added -fingerprint option to crl utility, to support new c_rehash 8788 features. 8789 [Steve Henson] 8790 8791 *) Eliminate non-ANSI declarations in crypto.h and stack.h. 8792 [Ulf M��ller] 8793 8794 *) Fix for SSL server purpose checking. Server checking was 8795 rejecting certificates which had extended key usage present 8796 but no ssl client purpose. 8797 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 8798 8799 *) Make PKCS#12 code work with no password. The PKCS#12 spec 8800 is a little unclear about how a blank password is handled. 8801 Since the password in encoded as a BMPString with terminating 8802 double NULL a zero length password would end up as just the 8803 double NULL. However no password at all is different and is 8804 handled differently in the PKCS#12 key generation code. NS 8805 treats a blank password as zero length. MSIE treats it as no 8806 password on export: but it will try both on import. We now do 8807 the same: PKCS12_parse() tries zero length and no password if 8808 the password is set to "" or NULL (NULL is now a valid password: 8809 it wasn't before) as does the pkcs12 application. 8810 [Steve Henson] 8811 8812 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 8813 perror when PEM_read_bio_X509_REQ fails, the error message must 8814 be obtained from the error queue. 8815 [Bodo Moeller] 8816 8817 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 8818 it in ERR_remove_state if appropriate, and change ERR_get_state 8819 accordingly to avoid race conditions (this is necessary because 8820 thread_hash is no longer constant once set). 8821 [Bodo Moeller] 8822 8823 *) Bugfix for linux-elf makefile.one. 8824 [Ulf M��ller] 8825 8826 *) RSA_get_default_method() will now cause a default 8827 RSA_METHOD to be chosen if one doesn't exist already. 8828 Previously this was only set during a call to RSA_new() 8829 or RSA_new_method(NULL) meaning it was possible for 8830 RSA_get_default_method() to return NULL. 8831 [Geoff Thorpe] 8832 8833 *) Added native name translation to the existing DSO code 8834 that will convert (if the flag to do so is set) filenames 8835 that are sufficiently small and have no path information 8836 into a canonical native form. Eg. "blah" converted to 8837 "libblah.so" or "blah.dll" etc. 8838 [Geoff Thorpe] 8839 8840 *) New function ERR_error_string_n(e, buf, len) which is like 8841 ERR_error_string(e, buf), but writes at most 'len' bytes 8842 including the 0 terminator. For ERR_error_string_n, 'buf' 8843 may not be NULL. 8844 [Damien Miller <djm@mindrot.org>, Bodo Moeller] 8845 8846 *) CONF library reworked to become more general. A new CONF 8847 configuration file reader "class" is implemented as well as a 8848 new functions (NCONF_*, for "New CONF") to handle it. The now 8849 old CONF_* functions are still there, but are reimplemented to 8850 work in terms of the new functions. Also, a set of functions 8851 to handle the internal storage of the configuration data is 8852 provided to make it easier to write new configuration file 8853 reader "classes" (I can definitely see something reading a 8854 configuration file in XML format, for example), called _CONF_*, 8855 or "the configuration storage API"... 8856 8857 The new configuration file reading functions are: 8858 8859 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 8860 NCONF_get_section, NCONF_get_string, NCONF_get_numbre 8861 8862 NCONF_default, NCONF_WIN32 8863 8864 NCONF_dump_fp, NCONF_dump_bio 8865 8866 NCONF_default and NCONF_WIN32 are method (or "class") choosers, 8867 NCONF_new creates a new CONF object. This works in the same way 8868 as other interfaces in OpenSSL, like the BIO interface. 8869 NCONF_dump_* dump the internal storage of the configuration file, 8870 which is useful for debugging. All other functions take the same 8871 arguments as the old CONF_* functions wth the exception of the 8872 first that must be a `CONF *' instead of a `LHASH *'. 8873 8874 To make it easer to use the new classes with the old CONF_* functions, 8875 the function CONF_set_default_method is provided. 8876 [Richard Levitte] 8877 8878 *) Add '-tls1' option to 'openssl ciphers', which was already 8879 mentioned in the documentation but had not been implemented. 8880 (This option is not yet really useful because even the additional 8881 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 8882 [Bodo Moeller] 8883 8884 *) Initial DSO code added into libcrypto for letting OpenSSL (and 8885 OpenSSL-based applications) load shared libraries and bind to 8886 them in a portable way. 8887 [Geoff Thorpe, with contributions from Richard Levitte] 8888 8889 Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 8890 8891 *) Make sure _lrotl and _lrotr are only used with MSVC. 8892 8893 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 8894 (the default implementation of RAND_status). 8895 8896 *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 8897 to '-clrext' (= clear extensions), as intended and documented. 8898 [Bodo Moeller; inconsistency pointed out by Michael Attili 8899 <attili@amaxo.com>] 8900 8901 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 8902 was larger than the MD block size. 8903 [Steve Henson, pointed out by Yost William <YostW@tce.com>] 8904 8905 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 8906 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 8907 using the passed key: if the passed key was a private key the result 8908 of X509_print(), for example, would be to print out all the private key 8909 components. 8910 [Steve Henson] 8911 8912 *) des_quad_cksum() byte order bug fix. 8913 [Ulf M��ller, using the problem description in krb4-0.9.7, where 8914 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 8915 8916 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 8917 discouraged. 8918 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 8919 8920 *) For easily testing in shell scripts whether some command 8921 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 8922 returns with exit code 0 iff no command of the given name is available. 8923 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 8924 the output goes to stdout and nothing is printed to stderr. 8925 Additional arguments are always ignored. 8926 8927 Since for each cipher there is a command of the same name, 8928 the 'no-cipher' compilation switches can be tested this way. 8929 8930 ('openssl no-XXX' is not able to detect pseudo-commands such 8931 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 8932 [Bodo Moeller] 8933 8934 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 8935 [Bodo Moeller] 8936 8937 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 8938 is set; it will be thrown away anyway because each handshake creates 8939 its own key. 8940 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 8941 to parameters -- in previous versions (since OpenSSL 0.9.3) the 8942 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 8943 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 8944 [Bodo Moeller] 8945 8946 *) New s_client option -ign_eof: EOF at stdin is ignored, and 8947 'Q' and 'R' lose their special meanings (quit/renegotiate). 8948 This is part of what -quiet does; unlike -quiet, -ign_eof 8949 does not suppress any output. 8950 [Richard Levitte] 8951 8952 *) Add compatibility options to the purpose and trust code. The 8953 purpose X509_PURPOSE_ANY is "any purpose" which automatically 8954 accepts a certificate or CA, this was the previous behaviour, 8955 with all the associated security issues. 8956 8957 X509_TRUST_COMPAT is the old trust behaviour: only and 8958 automatically trust self signed roots in certificate store. A 8959 new trust setting X509_TRUST_DEFAULT is used to specify that 8960 a purpose has no associated trust setting and it should instead 8961 use the value in the default purpose. 8962 [Steve Henson] 8963 8964 *) Fix the PKCS#8 DSA private key code so it decodes keys again 8965 and fix a memory leak. 8966 [Steve Henson] 8967 8968 *) In util/mkerr.pl (which implements 'make errors'), preserve 8969 reason strings from the previous version of the .c file, as 8970 the default to have only downcase letters (and digits) in 8971 automatically generated reasons codes is not always appropriate. 8972 [Bodo Moeller] 8973 8974 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 8975 using strerror. Previously, ERR_reason_error_string() returned 8976 library names as reason strings for SYSerr; but SYSerr is a special 8977 case where small numbers are errno values, not library numbers. 8978 [Bodo Moeller] 8979 8980 *) Add '-dsaparam' option to 'openssl dhparam' application. This 8981 converts DSA parameters into DH parameters. (When creating parameters, 8982 DSA_generate_parameters is used.) 8983 [Bodo Moeller] 8984 8985 *) Include 'length' (recommended exponent length) in C code generated 8986 by 'openssl dhparam -C'. 8987 [Bodo Moeller] 8988 8989 *) The second argument to set_label in perlasm was already being used 8990 so couldn't be used as a "file scope" flag. Moved to third argument 8991 which was free. 8992 [Steve Henson] 8993 8994 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 8995 instead of RAND_bytes for encryption IVs and salts. 8996 [Bodo Moeller] 8997 8998 *) Include RAND_status() into RAND_METHOD instead of implementing 8999 it only for md_rand.c Otherwise replacing the PRNG by calling 9000 RAND_set_rand_method would be impossible. 9001 [Bodo Moeller] 9002 9003 *) Don't let DSA_generate_key() enter an infinite loop if the random 9004 number generation fails. 9005 [Bodo Moeller] 9006 9007 *) New 'rand' application for creating pseudo-random output. 9008 [Bodo Moeller] 9009 9010 *) Added configuration support for Linux/IA64 9011 [Rolf Haberrecker <rolf@suse.de>] 9012 9013 *) Assembler module support for Mingw32. 9014 [Ulf M��ller] 9015 9016 *) Shared library support for HPUX (in shlib/). 9017 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 9018 9019 *) Shared library support for Solaris gcc. 9020 [Lutz Behnke <behnke@trustcenter.de>] 9021 9022 Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 9023 9024 *) PKCS7_encrypt() was adding text MIME headers twice because they 9025 were added manually and by SMIME_crlf_copy(). 9026 [Steve Henson] 9027 9028 *) In bntest.c don't call BN_rand with zero bits argument. 9029 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 9030 9031 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 9032 case was implemented. This caused BN_div_recp() to fail occasionally. 9033 [Ulf M��ller] 9034 9035 *) Add an optional second argument to the set_label() in the perl 9036 assembly language builder. If this argument exists and is set 9037 to 1 it signals that the assembler should use a symbol whose 9038 scope is the entire file, not just the current function. This 9039 is needed with MASM which uses the format label:: for this scope. 9040 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 9041 9042 *) Change the ASN1 types so they are typedefs by default. Before 9043 almost all types were #define'd to ASN1_STRING which was causing 9044 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 9045 for example. 9046 [Steve Henson] 9047 9048 *) Change names of new functions to the new get1/get0 naming 9049 convention: After 'get1', the caller owns a reference count 9050 and has to call ..._free; 'get0' returns a pointer to some 9051 data structure without incrementing reference counters. 9052 (Some of the existing 'get' functions increment a reference 9053 counter, some don't.) 9054 Similarly, 'set1' and 'add1' functions increase reference 9055 counters or duplicate objects. 9056 [Steve Henson] 9057 9058 *) Allow for the possibility of temp RSA key generation failure: 9059 the code used to assume it always worked and crashed on failure. 9060 [Steve Henson] 9061 9062 *) Fix potential buffer overrun problem in BIO_printf(). 9063 [Ulf M��ller, using public domain code by Patrick Powell; problem 9064 pointed out by David Sacerdote <das33@cornell.edu>] 9065 9066 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 9067 RAND_egd() and RAND_status(). In the command line application, 9068 the EGD socket can be specified like a seed file using RANDFILE 9069 or -rand. 9070 [Ulf M��ller] 9071 9072 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 9073 Some CAs (e.g. Verisign) distribute certificates in this form. 9074 [Steve Henson] 9075 9076 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 9077 list to exclude them. This means that no special compilation option 9078 is needed to use anonymous DH: it just needs to be included in the 9079 cipher list. 9080 [Steve Henson] 9081 9082 *) Change the EVP_MD_CTX_type macro so its meaning consistent with 9083 EVP_MD_type. The old functionality is available in a new macro called 9084 EVP_MD_md(). Change code that uses it and update docs. 9085 [Steve Henson] 9086 9087 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 9088 where the 'void *' argument is replaced by a function pointer argument. 9089 Previously 'void *' was abused to point to functions, which works on 9090 many platforms, but is not correct. As these functions are usually 9091 called by macros defined in OpenSSL header files, most source code 9092 should work without changes. 9093 [Richard Levitte] 9094 9095 *) <openssl/opensslconf.h> (which is created by Configure) now contains 9096 sections with information on -D... compiler switches used for 9097 compiling the library so that applications can see them. To enable 9098 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 9099 must be defined. E.g., 9100 #define OPENSSL_ALGORITHM_DEFINES 9101 #include <openssl/opensslconf.h> 9102 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 9103 [Richard Levitte, Ulf and Bodo M��ller] 9104 9105 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 9106 record layer. 9107 [Bodo Moeller] 9108 9109 *) Change the 'other' type in certificate aux info to a STACK_OF 9110 X509_ALGOR. Although not an AlgorithmIdentifier as such it has 9111 the required ASN1 format: arbitrary types determined by an OID. 9112 [Steve Henson] 9113 9114 *) Add some PEM_write_X509_REQ_NEW() functions and a command line 9115 argument to 'req'. This is not because the function is newer or 9116 better than others it just uses the work 'NEW' in the certificate 9117 request header lines. Some software needs this. 9118 [Steve Henson] 9119 9120 *) Reorganise password command line arguments: now passwords can be 9121 obtained from various sources. Delete the PEM_cb function and make 9122 it the default behaviour: i.e. if the callback is NULL and the 9123 usrdata argument is not NULL interpret it as a null terminated pass 9124 phrase. If usrdata and the callback are NULL then the pass phrase 9125 is prompted for as usual. 9126 [Steve Henson] 9127 9128 *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 9129 the support is automatically enabled. The resulting binaries will 9130 autodetect the card and use it if present. 9131 [Ben Laurie and Compaq Inc.] 9132 9133 *) Work around for Netscape hang bug. This sends certificate request 9134 and server done in one record. Since this is perfectly legal in the 9135 SSL/TLS protocol it isn't a "bug" option and is on by default. See 9136 the bugs/SSLv3 entry for more info. 9137 [Steve Henson] 9138 9139 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 9140 [Andy Polyakov] 9141 9142 *) Add -rand argument to smime and pkcs12 applications and read/write 9143 of seed file. 9144 [Steve Henson] 9145 9146 *) New 'passwd' tool for crypt(3) and apr1 password hashes. 9147 [Bodo Moeller] 9148 9149 *) Add command line password options to the remaining applications. 9150 [Steve Henson] 9151 9152 *) Bug fix for BN_div_recp() for numerators with an even number of 9153 bits. 9154 [Ulf M��ller] 9155 9156 *) More tests in bntest.c, and changed test_bn output. 9157 [Ulf M��ller] 9158 9159 *) ./config recognizes MacOS X now. 9160 [Andy Polyakov] 9161 9162 *) Bug fix for BN_div() when the first words of num and divsor are 9163 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 9164 [Ulf M��ller] 9165 9166 *) Add support for various broken PKCS#8 formats, and command line 9167 options to produce them. 9168 [Steve Henson] 9169 9170 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 9171 get temporary BIGNUMs from a BN_CTX. 9172 [Ulf M��ller] 9173 9174 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 9175 for p == 0. 9176 [Ulf M��ller] 9177 9178 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 9179 include a #define from the old name to the new. The original intent 9180 was that statically linked binaries could for example just call 9181 SSLeay_add_all_ciphers() to just add ciphers to the table and not 9182 link with digests. This never worked becayse SSLeay_add_all_digests() 9183 and SSLeay_add_all_ciphers() were in the same source file so calling 9184 one would link with the other. They are now in separate source files. 9185 [Steve Henson] 9186 9187 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 9188 [Steve Henson] 9189 9190 *) Use a less unusual form of the Miller-Rabin primality test (it used 9191 a binary algorithm for exponentiation integrated into the Miller-Rabin 9192 loop, our standard modexp algorithms are faster). 9193 [Bodo Moeller] 9194 9195 *) Support for the EBCDIC character set completed. 9196 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 9197 9198 *) Source code cleanups: use const where appropriate, eliminate casts, 9199 use void * instead of char * in lhash. 9200 [Ulf M��ller] 9201 9202 *) Bugfix: ssl3_send_server_key_exchange was not restartable 9203 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 9204 this the server could overwrite ephemeral keys that the client 9205 has already seen). 9206 [Bodo Moeller] 9207 9208 *) Turn DSA_is_prime into a macro that calls BN_is_prime, 9209 using 50 iterations of the Rabin-Miller test. 9210 9211 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 9212 iterations of the Rabin-Miller test as required by the appendix 9213 to FIPS PUB 186[-1]) instead of DSA_is_prime. 9214 As BN_is_prime_fasttest includes trial division, DSA parameter 9215 generation becomes much faster. 9216 9217 This implies a change for the callback functions in DSA_is_prime 9218 and DSA_generate_parameters: The callback function is called once 9219 for each positive witness in the Rabin-Miller test, not just 9220 occasionally in the inner loop; and the parameters to the 9221 callback function now provide an iteration count for the outer 9222 loop rather than for the current invocation of the inner loop. 9223 DSA_generate_parameters additionally can call the callback 9224 function with an 'iteration count' of -1, meaning that a 9225 candidate has passed the trial division test (when q is generated 9226 from an application-provided seed, trial division is skipped). 9227 [Bodo Moeller] 9228 9229 *) New function BN_is_prime_fasttest that optionally does trial 9230 division before starting the Rabin-Miller test and has 9231 an additional BN_CTX * argument (whereas BN_is_prime always 9232 has to allocate at least one BN_CTX). 9233 'callback(1, -1, cb_arg)' is called when a number has passed the 9234 trial division stage. 9235 [Bodo Moeller] 9236 9237 *) Fix for bug in CRL encoding. The validity dates weren't being handled 9238 as ASN1_TIME. 9239 [Steve Henson] 9240 9241 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 9242 [Steve Henson] 9243 9244 *) New function BN_pseudo_rand(). 9245 [Ulf M��ller] 9246 9247 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 9248 bignum version of BN_from_montgomery() with the working code from 9249 SSLeay 0.9.0 (the word based version is faster anyway), and clean up 9250 the comments. 9251 [Ulf M��ller] 9252 9253 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 9254 made it impossible to use the same SSL_SESSION data structure in 9255 SSL2 clients in multiple threads. 9256 [Bodo Moeller] 9257 9258 *) The return value of RAND_load_file() no longer counts bytes obtained 9259 by stat(). RAND_load_file(..., -1) is new and uses the complete file 9260 to seed the PRNG (previously an explicit byte count was required). 9261 [Ulf M��ller, Bodo M��ller] 9262 9263 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 9264 used (char *) instead of (void *) and had casts all over the place. 9265 [Steve Henson] 9266 9267 *) Make BN_generate_prime() return NULL on error if ret!=NULL. 9268 [Ulf M��ller] 9269 9270 *) Retain source code compatibility for BN_prime_checks macro: 9271 BN_is_prime(..., BN_prime_checks, ...) now uses 9272 BN_prime_checks_for_size to determine the appropriate number of 9273 Rabin-Miller iterations. 9274 [Ulf M��ller] 9275 9276 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 9277 DH_CHECK_P_NOT_SAFE_PRIME. 9278 (Check if this is true? OpenPGP calls them "strong".) 9279 [Ulf M��ller] 9280 9281 *) Merge the functionality of "dh" and "gendh" programs into a new program 9282 "dhparam". The old programs are retained for now but will handle DH keys 9283 (instead of parameters) in future. 9284 [Steve Henson] 9285 9286 *) Make the ciphers, s_server and s_client programs check the return values 9287 when a new cipher list is set. 9288 [Steve Henson] 9289 9290 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 9291 ciphers. Before when the 56bit ciphers were enabled the sorting was 9292 wrong. 9293 9294 The syntax for the cipher sorting has been extended to support sorting by 9295 cipher-strength (using the strength_bits hard coded in the tables). 9296 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 9297 9298 Fix a bug in the cipher-command parser: when supplying a cipher command 9299 string with an "undefined" symbol (neither command nor alphanumeric 9300 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 9301 an error is flagged. 9302 9303 Due to the strength-sorting extension, the code of the 9304 ssl_create_cipher_list() function was completely rearranged. I hope that 9305 the readability was also increased :-) 9306 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 9307 9308 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 9309 for the first serial number and places 2 in the serial number file. This 9310 avoids problems when the root CA is created with serial number zero and 9311 the first user certificate has the same issuer name and serial number 9312 as the root CA. 9313 [Steve Henson] 9314 9315 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 9316 the new code. Add documentation for this stuff. 9317 [Steve Henson] 9318 9319 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 9320 X509_*() to X509at_*() on the grounds that they don't handle X509 9321 structures and behave in an analagous way to the X509v3 functions: 9322 they shouldn't be called directly but wrapper functions should be used 9323 instead. 9324 9325 So we also now have some wrapper functions that call the X509at functions 9326 when passed certificate requests. (TO DO: similar things can be done with 9327 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 9328 things. Some of these need some d2i or i2d and print functionality 9329 because they handle more complex structures.) 9330 [Steve Henson] 9331 9332 *) Add missing #ifndefs that caused missing symbols when building libssl 9333 as a shared library without RSA. Use #ifndef NO_SSL2 instead of 9334 NO_RSA in ssl/s2*.c. 9335 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M��ller] 9336 9337 *) Precautions against using the PRNG uninitialized: RAND_bytes() now 9338 has a return value which indicates the quality of the random data 9339 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 9340 error queue. New function RAND_pseudo_bytes() generates output that is 9341 guaranteed to be unique but not unpredictable. RAND_add is like 9342 RAND_seed, but takes an extra argument for an entropy estimate 9343 (RAND_seed always assumes full entropy). 9344 [Ulf M��ller] 9345 9346 *) Do more iterations of Rabin-Miller probable prime test (specifically, 9347 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 9348 instead of only 2 for all lengths; see BN_prime_checks_for_size definition 9349 in crypto/bn/bn_prime.c for the complete table). This guarantees a 9350 false-positive rate of at most 2^-80 for random input. 9351 [Bodo Moeller] 9352 9353 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 9354 [Bodo Moeller] 9355 9356 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 9357 in the 0.9.5 release), this returns the chain 9358 from an X509_CTX structure with a dup of the stack and all 9359 the X509 reference counts upped: so the stack will exist 9360 after X509_CTX_cleanup() has been called. Modify pkcs12.c 9361 to use this. 9362 9363 Also make SSL_SESSION_print() print out the verify return 9364 code. 9365 [Steve Henson] 9366 9367 *) Add manpage for the pkcs12 command. Also change the default 9368 behaviour so MAC iteration counts are used unless the new 9369 -nomaciter option is used. This improves file security and 9370 only older versions of MSIE (4.0 for example) need it. 9371 [Steve Henson] 9372 9373 *) Honor the no-xxx Configure options when creating .DEF files. 9374 [Ulf M��ller] 9375 9376 *) Add PKCS#10 attributes to field table: challengePassword, 9377 unstructuredName and unstructuredAddress. These are taken from 9378 draft PKCS#9 v2.0 but are compatible with v1.2 provided no 9379 international characters are used. 9380 9381 More changes to X509_ATTRIBUTE code: allow the setting of types 9382 based on strings. Remove the 'loc' parameter when adding 9383 attributes because these will be a SET OF encoding which is sorted 9384 in ASN1 order. 9385 [Steve Henson] 9386 9387 *) Initial changes to the 'req' utility to allow request generation 9388 automation. This will allow an application to just generate a template 9389 file containing all the field values and have req construct the 9390 request. 9391 9392 Initial support for X509_ATTRIBUTE handling. Stacks of these are 9393 used all over the place including certificate requests and PKCS#7 9394 structures. They are currently handled manually where necessary with 9395 some primitive wrappers for PKCS#7. The new functions behave in a 9396 manner analogous to the X509 extension functions: they allow 9397 attributes to be looked up by NID and added. 9398 9399 Later something similar to the X509V3 code would be desirable to 9400 automatically handle the encoding, decoding and printing of the 9401 more complex types. The string types like challengePassword can 9402 be handled by the string table functions. 9403 9404 Also modified the multi byte string table handling. Now there is 9405 a 'global mask' which masks out certain types. The table itself 9406 can use the flag STABLE_NO_MASK to ignore the mask setting: this 9407 is useful when for example there is only one permissible type 9408 (as in countryName) and using the mask might result in no valid 9409 types at all. 9410 [Steve Henson] 9411 9412 *) Clean up 'Finished' handling, and add functions SSL_get_finished and 9413 SSL_get_peer_finished to allow applications to obtain the latest 9414 Finished messages sent to the peer or expected from the peer, 9415 respectively. (SSL_get_peer_finished is usually the Finished message 9416 actually received from the peer, otherwise the protocol will be aborted.) 9417 9418 As the Finished message are message digests of the complete handshake 9419 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 9420 be used for external authentication procedures when the authentication 9421 provided by SSL/TLS is not desired or is not enough. 9422 [Bodo Moeller] 9423 9424 *) Enhanced support for Alpha Linux is added. Now ./config checks if 9425 the host supports BWX extension and if Compaq C is present on the 9426 $PATH. Just exploiting of the BWX extension results in 20-30% 9427 performance kick for some algorithms, e.g. DES and RC4 to mention 9428 a couple. Compaq C in turn generates ~20% faster code for MD5 and 9429 SHA1. 9430 [Andy Polyakov] 9431 9432 *) Add support for MS "fast SGC". This is arguably a violation of the 9433 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 9434 weak crypto and after checking the certificate is SGC a second one 9435 with strong crypto. MS SGC stops the first handshake after receiving 9436 the server certificate message and sends a second client hello. Since 9437 a server will typically do all the time consuming operations before 9438 expecting any further messages from the client (server key exchange 9439 is the most expensive) there is little difference between the two. 9440 9441 To get OpenSSL to support MS SGC we have to permit a second client 9442 hello message after we have sent server done. In addition we have to 9443 reset the MAC if we do get this second client hello. 9444 [Steve Henson] 9445 9446 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 9447 if a DER encoded private key is RSA or DSA traditional format. Changed 9448 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 9449 format DER encoded private key. Newer code should use PKCS#8 format which 9450 has the key type encoded in the ASN1 structure. Added DER private key 9451 support to pkcs8 application. 9452 [Steve Henson] 9453 9454 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 9455 ciphersuites has been selected (as required by the SSL 3/TLS 1 9456 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 9457 is set, we interpret this as a request to violate the specification 9458 (the worst that can happen is a handshake failure, and 'correct' 9459 behaviour would result in a handshake failure anyway). 9460 [Bodo Moeller] 9461 9462 *) In SSL_CTX_add_session, take into account that there might be multiple 9463 SSL_SESSION structures with the same session ID (e.g. when two threads 9464 concurrently obtain them from an external cache). 9465 The internal cache can handle only one SSL_SESSION with a given ID, 9466 so if there's a conflict, we now throw out the old one to achieve 9467 consistency. 9468 [Bodo Moeller] 9469 9470 *) Add OIDs for idea and blowfish in CBC mode. This will allow both 9471 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 9472 some routines that use cipher OIDs: some ciphers do not have OIDs 9473 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 9474 example. 9475 [Steve Henson] 9476 9477 *) Simplify the trust setting structure and code. Now we just have 9478 two sequences of OIDs for trusted and rejected settings. These will 9479 typically have values the same as the extended key usage extension 9480 and any application specific purposes. 9481 9482 The trust checking code now has a default behaviour: it will just 9483 check for an object with the same NID as the passed id. Functions can 9484 be provided to override either the default behaviour or the behaviour 9485 for a given id. SSL client, server and email already have functions 9486 in place for compatibility: they check the NID and also return "trusted" 9487 if the certificate is self signed. 9488 [Steve Henson] 9489 9490 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 9491 traditional format into an EVP_PKEY structure. 9492 [Steve Henson] 9493 9494 *) Add a password callback function PEM_cb() which either prompts for 9495 a password if usr_data is NULL or otherwise assumes it is a null 9496 terminated password. Allow passwords to be passed on command line 9497 environment or config files in a few more utilities. 9498 [Steve Henson] 9499 9500 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 9501 keys. Add some short names for PKCS#8 PBE algorithms and allow them 9502 to be specified on the command line for the pkcs8 and pkcs12 utilities. 9503 Update documentation. 9504 [Steve Henson] 9505 9506 *) Support for ASN1 "NULL" type. This could be handled before by using 9507 ASN1_TYPE but there wasn't any function that would try to read a NULL 9508 and produce an error if it couldn't. For compatibility we also have 9509 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 9510 don't allocate anything because they don't need to. 9511 [Steve Henson] 9512 9513 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 9514 for details. 9515 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 9516 9517 *) Rebuild of the memory allocation routines used by OpenSSL code and 9518 possibly others as well. The purpose is to make an interface that 9519 provide hooks so anyone can build a separate set of allocation and 9520 deallocation routines to be used by OpenSSL, for example memory 9521 pool implementations, or something else, which was previously hard 9522 since Malloc(), Realloc() and Free() were defined as macros having 9523 the values malloc, realloc and free, respectively (except for Win32 9524 compilations). The same is provided for memory debugging code. 9525 OpenSSL already comes with functionality to find memory leaks, but 9526 this gives people a chance to debug other memory problems. 9527 9528 With these changes, a new set of functions and macros have appeared: 9529 9530 CRYPTO_set_mem_debug_functions() [F] 9531 CRYPTO_get_mem_debug_functions() [F] 9532 CRYPTO_dbg_set_options() [F] 9533 CRYPTO_dbg_get_options() [F] 9534 CRYPTO_malloc_debug_init() [M] 9535 9536 The memory debug functions are NULL by default, unless the library 9537 is compiled with CRYPTO_MDEBUG or friends is defined. If someone 9538 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 9539 gives the standard debugging functions that come with OpenSSL) or 9540 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 9541 provided by the library user) must be used. When the standard 9542 debugging functions are used, CRYPTO_dbg_set_options can be used to 9543 request additional information: 9544 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 9545 the CRYPTO_MDEBUG_xxx macro when compiling the library. 9546 9547 Also, things like CRYPTO_set_mem_functions will always give the 9548 expected result (the new set of functions is used for allocation 9549 and deallocation) at all times, regardless of platform and compiler 9550 options. 9551 9552 To finish it up, some functions that were never use in any other 9553 way than through macros have a new API and new semantic: 9554 9555 CRYPTO_dbg_malloc() 9556 CRYPTO_dbg_realloc() 9557 CRYPTO_dbg_free() 9558 9559 All macros of value have retained their old syntax. 9560 [Richard Levitte and Bodo Moeller] 9561 9562 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 9563 ordering of SMIMECapabilities wasn't in "strength order" and there 9564 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 9565 algorithm. 9566 [Steve Henson] 9567 9568 *) Some ASN1 types with illegal zero length encoding (INTEGER, 9569 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 9570 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 9571 9572 *) Merge in my S/MIME library for OpenSSL. This provides a simple 9573 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 9574 functionality to handle multipart/signed properly) and a utility 9575 called 'smime' to call all this stuff. This is based on code I 9576 originally wrote for Celo who have kindly allowed it to be 9577 included in OpenSSL. 9578 [Steve Henson] 9579 9580 *) Add variants des_set_key_checked and des_set_key_unchecked of 9581 des_set_key (aka des_key_sched). Global variable des_check_key 9582 decides which of these is called by des_set_key; this way 9583 des_check_key behaves as it always did, but applications and 9584 the library itself, which was buggy for des_check_key == 1, 9585 have a cleaner way to pick the version they need. 9586 [Bodo Moeller] 9587 9588 *) New function PKCS12_newpass() which changes the password of a 9589 PKCS12 structure. 9590 [Steve Henson] 9591 9592 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 9593 dynamic mix. In both cases the ids can be used as an index into the 9594 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 9595 functions so they accept a list of the field values and the 9596 application doesn't need to directly manipulate the X509_TRUST 9597 structure. 9598 [Steve Henson] 9599 9600 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 9601 need initialising. 9602 [Steve Henson] 9603 9604 *) Modify the way the V3 extension code looks up extensions. This now 9605 works in a similar way to the object code: we have some "standard" 9606 extensions in a static table which is searched with OBJ_bsearch() 9607 and the application can add dynamic ones if needed. The file 9608 crypto/x509v3/ext_dat.h now has the info: this file needs to be 9609 updated whenever a new extension is added to the core code and kept 9610 in ext_nid order. There is a simple program 'tabtest.c' which checks 9611 this. New extensions are not added too often so this file can readily 9612 be maintained manually. 9613 9614 There are two big advantages in doing things this way. The extensions 9615 can be looked up immediately and no longer need to be "added" using 9616 X509V3_add_standard_extensions(): this function now does nothing. 9617 [Side note: I get *lots* of email saying the extension code doesn't 9618 work because people forget to call this function] 9619 Also no dynamic allocation is done unless new extensions are added: 9620 so if we don't add custom extensions there is no need to call 9621 X509V3_EXT_cleanup(). 9622 [Steve Henson] 9623 9624 *) Modify enc utility's salting as follows: make salting the default. Add a 9625 magic header, so unsalted files fail gracefully instead of just decrypting 9626 to garbage. This is because not salting is a big security hole, so people 9627 should be discouraged from doing it. 9628 [Ben Laurie] 9629 9630 *) Fixes and enhancements to the 'x509' utility. It allowed a message 9631 digest to be passed on the command line but it only used this 9632 parameter when signing a certificate. Modified so all relevant 9633 operations are affected by the digest parameter including the 9634 -fingerprint and -x509toreq options. Also -x509toreq choked if a 9635 DSA key was used because it didn't fix the digest. 9636 [Steve Henson] 9637 9638 *) Initial certificate chain verify code. Currently tests the untrusted 9639 certificates for consistency with the verify purpose (which is set 9640 when the X509_STORE_CTX structure is set up) and checks the pathlength. 9641 9642 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 9643 this is because it will reject chains with invalid extensions whereas 9644 every previous version of OpenSSL and SSLeay made no checks at all. 9645 9646 Trust code: checks the root CA for the relevant trust settings. Trust 9647 settings have an initial value consistent with the verify purpose: e.g. 9648 if the verify purpose is for SSL client use it expects the CA to be 9649 trusted for SSL client use. However the default value can be changed to 9650 permit custom trust settings: one example of this would be to only trust 9651 certificates from a specific "secure" set of CAs. 9652 9653 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 9654 which should be used for version portability: especially since the 9655 verify structure is likely to change more often now. 9656 9657 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 9658 to set them. If not set then assume SSL clients will verify SSL servers 9659 and vice versa. 9660 9661 Two new options to the verify program: -untrusted allows a set of 9662 untrusted certificates to be passed in and -purpose which sets the 9663 intended purpose of the certificate. If a purpose is set then the 9664 new chain verify code is used to check extension consistency. 9665 [Steve Henson] 9666 9667 *) Support for the authority information access extension. 9668 [Steve Henson] 9669 9670 *) Modify RSA and DSA PEM read routines to transparently handle 9671 PKCS#8 format private keys. New *_PUBKEY_* functions that handle 9672 public keys in a format compatible with certificate 9673 SubjectPublicKeyInfo structures. Unfortunately there were already 9674 functions called *_PublicKey_* which used various odd formats so 9675 these are retained for compatibility: however the DSA variants were 9676 never in a public release so they have been deleted. Changed dsa/rsa 9677 utilities to handle the new format: note no releases ever handled public 9678 keys so we should be OK. 9679 9680 The primary motivation for this change is to avoid the same fiasco 9681 that dogs private keys: there are several incompatible private key 9682 formats some of which are standard and some OpenSSL specific and 9683 require various evil hacks to allow partial transparent handling and 9684 even then it doesn't work with DER formats. Given the option anything 9685 other than PKCS#8 should be dumped: but the other formats have to 9686 stay in the name of compatibility. 9687 9688 With public keys and the benefit of hindsight one standard format 9689 is used which works with EVP_PKEY, RSA or DSA structures: though 9690 it clearly returns an error if you try to read the wrong kind of key. 9691 9692 Added a -pubkey option to the 'x509' utility to output the public key. 9693 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 9694 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 9695 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 9696 that do the same as the EVP_PKEY_assign_*() except they up the 9697 reference count of the added key (they don't "swallow" the 9698 supplied key). 9699 [Steve Henson] 9700 9701 *) Fixes to crypto/x509/by_file.c the code to read in certificates and 9702 CRLs would fail if the file contained no certificates or no CRLs: 9703 added a new function to read in both types and return the number 9704 read: this means that if none are read it will be an error. The 9705 DER versions of the certificate and CRL reader would always fail 9706 because it isn't possible to mix certificates and CRLs in DER format 9707 without choking one or the other routine. Changed this to just read 9708 a certificate: this is the best we can do. Also modified the code 9709 in apps/verify.c to take notice of return codes: it was previously 9710 attempting to read in certificates from NULL pointers and ignoring 9711 any errors: this is one reason why the cert and CRL reader seemed 9712 to work. It doesn't check return codes from the default certificate 9713 routines: these may well fail if the certificates aren't installed. 9714 [Steve Henson] 9715 9716 *) Code to support otherName option in GeneralName. 9717 [Steve Henson] 9718 9719 *) First update to verify code. Change the verify utility 9720 so it warns if it is passed a self signed certificate: 9721 for consistency with the normal behaviour. X509_verify 9722 has been modified to it will now verify a self signed 9723 certificate if *exactly* the same certificate appears 9724 in the store: it was previously impossible to trust a 9725 single self signed certificate. This means that: 9726 openssl verify ss.pem 9727 now gives a warning about a self signed certificate but 9728 openssl verify -CAfile ss.pem ss.pem 9729 is OK. 9730 [Steve Henson] 9731 9732 *) For servers, store verify_result in SSL_SESSION data structure 9733 (and add it to external session representation). 9734 This is needed when client certificate verifications fails, 9735 but an application-provided verification callback (set by 9736 SSL_CTX_set_cert_verify_callback) allows accepting the session 9737 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 9738 but returns 1): When the session is reused, we have to set 9739 ssl->verify_result to the appropriate error code to avoid 9740 security holes. 9741 [Bodo Moeller, problem pointed out by Lutz Jaenicke] 9742 9743 *) Fix a bug in the new PKCS#7 code: it didn't consider the 9744 case in PKCS7_dataInit() where the signed PKCS7 structure 9745 didn't contain any existing data because it was being created. 9746 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 9747 9748 *) Add a salt to the key derivation routines in enc.c. This 9749 forms the first 8 bytes of the encrypted file. Also add a 9750 -S option to allow a salt to be input on the command line. 9751 [Steve Henson] 9752 9753 *) New function X509_cmp(). Oddly enough there wasn't a function 9754 to compare two certificates. We do this by working out the SHA1 9755 hash and comparing that. X509_cmp() will be needed by the trust 9756 code. 9757 [Steve Henson] 9758 9759 *) SSL_get1_session() is like SSL_get_session(), but increments 9760 the reference count in the SSL_SESSION returned. 9761 [Geoff Thorpe <geoff@eu.c2.net>] 9762 9763 *) Fix for 'req': it was adding a null to request attributes. 9764 Also change the X509_LOOKUP and X509_INFO code to handle 9765 certificate auxiliary information. 9766 [Steve Henson] 9767 9768 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 9769 the 'enc' command. 9770 [Steve Henson] 9771 9772 *) Add the possibility to add extra information to the memory leak 9773 detecting output, to form tracebacks, showing from where each 9774 allocation was originated: CRYPTO_push_info("constant string") adds 9775 the string plus current file name and line number to a per-thread 9776 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 9777 is like calling CYRPTO_pop_info() until the stack is empty. 9778 Also updated memory leak detection code to be multi-thread-safe. 9779 [Richard Levitte] 9780 9781 *) Add options -text and -noout to pkcs7 utility and delete the 9782 encryption options which never did anything. Update docs. 9783 [Steve Henson] 9784 9785 *) Add options to some of the utilities to allow the pass phrase 9786 to be included on either the command line (not recommended on 9787 OSes like Unix) or read from the environment. Update the 9788 manpages and fix a few bugs. 9789 [Steve Henson] 9790 9791 *) Add a few manpages for some of the openssl commands. 9792 [Steve Henson] 9793 9794 *) Fix the -revoke option in ca. It was freeing up memory twice, 9795 leaking and not finding already revoked certificates. 9796 [Steve Henson] 9797 9798 *) Extensive changes to support certificate auxiliary information. 9799 This involves the use of X509_CERT_AUX structure and X509_AUX 9800 functions. An X509_AUX function such as PEM_read_X509_AUX() 9801 can still read in a certificate file in the usual way but it 9802 will also read in any additional "auxiliary information". By 9803 doing things this way a fair degree of compatibility can be 9804 retained: existing certificates can have this information added 9805 using the new 'x509' options. 9806 9807 Current auxiliary information includes an "alias" and some trust 9808 settings. The trust settings will ultimately be used in enhanced 9809 certificate chain verification routines: currently a certificate 9810 can only be trusted if it is self signed and then it is trusted 9811 for all purposes. 9812 [Steve Henson] 9813 9814 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 9815 The problem was that one of the replacement routines had not been working 9816 since SSLeay releases. For now the offending routine has been replaced 9817 with non-optimised assembler. Even so, this now gives around 95% 9818 performance improvement for 1024 bit RSA signs. 9819 [Mark Cox] 9820 9821 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 9822 handling. Most clients have the effective key size in bits equal to 9823 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 9824 A few however don't do this and instead use the size of the decrypted key 9825 to determine the RC2 key length and the AlgorithmIdentifier to determine 9826 the effective key length. In this case the effective key length can still 9827 be 40 bits but the key length can be 168 bits for example. This is fixed 9828 by manually forcing an RC2 key into the EVP_PKEY structure because the 9829 EVP code can't currently handle unusual RC2 key sizes: it always assumes 9830 the key length and effective key length are equal. 9831 [Steve Henson] 9832 9833 *) Add a bunch of functions that should simplify the creation of 9834 X509_NAME structures. Now you should be able to do: 9835 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 9836 and have it automatically work out the correct field type and fill in 9837 the structures. The more adventurous can try: 9838 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 9839 and it will (hopefully) work out the correct multibyte encoding. 9840 [Steve Henson] 9841 9842 *) Change the 'req' utility to use the new field handling and multibyte 9843 copy routines. Before the DN field creation was handled in an ad hoc 9844 way in req, ca, and x509 which was rather broken and didn't support 9845 BMPStrings or UTF8Strings. Since some software doesn't implement 9846 BMPStrings or UTF8Strings yet, they can be enabled using the config file 9847 using the dirstring_type option. See the new comment in the default 9848 openssl.cnf for more info. 9849 [Steve Henson] 9850 9851 *) Make crypto/rand/md_rand.c more robust: 9852 - Assure unique random numbers after fork(). 9853 - Make sure that concurrent threads access the global counter and 9854 md serializably so that we never lose entropy in them 9855 or use exactly the same state in multiple threads. 9856 Access to the large state is not always serializable because 9857 the additional locking could be a performance killer, and 9858 md should be large enough anyway. 9859 [Bodo Moeller] 9860 9861 *) New file apps/app_rand.c with commonly needed functionality 9862 for handling the random seed file. 9863 9864 Use the random seed file in some applications that previously did not: 9865 ca, 9866 dsaparam -genkey (which also ignored its '-rand' option), 9867 s_client, 9868 s_server, 9869 x509 (when signing). 9870 Except on systems with /dev/urandom, it is crucial to have a random 9871 seed file at least for key creation, DSA signing, and for DH exchanges; 9872 for RSA signatures we could do without one. 9873 9874 gendh and gendsa (unlike genrsa) used to read only the first byte 9875 of each file listed in the '-rand' option. The function as previously 9876 found in genrsa is now in app_rand.c and is used by all programs 9877 that support '-rand'. 9878 [Bodo Moeller] 9879 9880 *) In RAND_write_file, use mode 0600 for creating files; 9881 don't just chmod when it may be too late. 9882 [Bodo Moeller] 9883 9884 *) Report an error from X509_STORE_load_locations 9885 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 9886 [Bill Perry] 9887 9888 *) New function ASN1_mbstring_copy() this copies a string in either 9889 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 9890 into an ASN1_STRING type. A mask of permissible types is passed 9891 and it chooses the "minimal" type to use or an error if not type 9892 is suitable. 9893 [Steve Henson] 9894 9895 *) Add function equivalents to the various macros in asn1.h. The old 9896 macros are retained with an M_ prefix. Code inside the library can 9897 use the M_ macros. External code (including the openssl utility) 9898 should *NOT* in order to be "shared library friendly". 9899 [Steve Henson] 9900 9901 *) Add various functions that can check a certificate's extensions 9902 to see if it usable for various purposes such as SSL client, 9903 server or S/MIME and CAs of these types. This is currently 9904 VERY EXPERIMENTAL but will ultimately be used for certificate chain 9905 verification. Also added a -purpose flag to x509 utility to 9906 print out all the purposes. 9907 [Steve Henson] 9908 9909 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 9910 functions. 9911 [Steve Henson] 9912 9913 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 9914 for, obtain and decode and extension and obtain its critical flag. 9915 This allows all the necessary extension code to be handled in a 9916 single function call. 9917 [Steve Henson] 9918 9919 *) RC4 tune-up featuring 30-40% performance improvement on most RISC 9920 platforms. See crypto/rc4/rc4_enc.c for further details. 9921 [Andy Polyakov] 9922 9923 *) New -noout option to asn1parse. This causes no output to be produced 9924 its main use is when combined with -strparse and -out to extract data 9925 from a file (which may not be in ASN.1 format). 9926 [Steve Henson] 9927 9928 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 9929 when producing the local key id. 9930 [Richard Levitte <levitte@stacken.kth.se>] 9931 9932 *) New option -dhparam in s_server. This allows a DH parameter file to be 9933 stated explicitly. If it is not stated then it tries the first server 9934 certificate file. The previous behaviour hard coded the filename 9935 "server.pem". 9936 [Steve Henson] 9937 9938 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 9939 a public key to be input or output. For example: 9940 openssl rsa -in key.pem -pubout -out pubkey.pem 9941 Also added necessary DSA public key functions to handle this. 9942 [Steve Henson] 9943 9944 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 9945 in the message. This was handled by allowing 9946 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 9947 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 9948 9949 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 9950 to the end of the strings whereas this didn't. This would cause problems 9951 if strings read with d2i_ASN1_bytes() were later modified. 9952 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 9953 9954 *) Fix for base64 decode bug. When a base64 bio reads only one line of 9955 data and it contains EOF it will end up returning an error. This is 9956 caused by input 46 bytes long. The cause is due to the way base64 9957 BIOs find the start of base64 encoded data. They do this by trying a 9958 trial decode on each line until they find one that works. When they 9959 do a flag is set and it starts again knowing it can pass all the 9960 data directly through the decoder. Unfortunately it doesn't reset 9961 the context it uses. This means that if EOF is reached an attempt 9962 is made to pass two EOFs through the context and this causes the 9963 resulting error. This can also cause other problems as well. As is 9964 usual with these problems it takes *ages* to find and the fix is 9965 trivial: move one line. 9966 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 9967 9968 *) Ugly workaround to get s_client and s_server working under Windows. The 9969 old code wouldn't work because it needed to select() on sockets and the 9970 tty (for keypresses and to see if data could be written). Win32 only 9971 supports select() on sockets so we select() with a 1s timeout on the 9972 sockets and then see if any characters are waiting to be read, if none 9973 are present then we retry, we also assume we can always write data to 9974 the tty. This isn't nice because the code then blocks until we've 9975 received a complete line of data and it is effectively polling the 9976 keyboard at 1s intervals: however it's quite a bit better than not 9977 working at all :-) A dedicated Windows application might handle this 9978 with an event loop for example. 9979 [Steve Henson] 9980 9981 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 9982 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 9983 will be called when RSA_sign() and RSA_verify() are used. This is useful 9984 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 9985 For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 9986 should *not* be used: RSA_sign() and RSA_verify() must be used instead. 9987 This necessitated the support of an extra signature type NID_md5_sha1 9988 for SSL signatures and modifications to the SSL library to use it instead 9989 of calling RSA_public_decrypt() and RSA_private_encrypt(). 9990 [Steve Henson] 9991 9992 *) Add new -verify -CAfile and -CApath options to the crl program, these 9993 will lookup a CRL issuers certificate and verify the signature in a 9994 similar way to the verify program. Tidy up the crl program so it 9995 no longer accesses structures directly. Make the ASN1 CRL parsing a bit 9996 less strict. It will now permit CRL extensions even if it is not 9997 a V2 CRL: this will allow it to tolerate some broken CRLs. 9998 [Steve Henson] 9999 10000 *) Initialize all non-automatic variables each time one of the openssl 10001 sub-programs is started (this is necessary as they may be started 10002 multiple times from the "OpenSSL>" prompt). 10003 [Lennart Bang, Bodo Moeller] 10004 10005 *) Preliminary compilation option RSA_NULL which disables RSA crypto without 10006 removing all other RSA functionality (this is what NO_RSA does). This 10007 is so (for example) those in the US can disable those operations covered 10008 by the RSA patent while allowing storage and parsing of RSA keys and RSA 10009 key generation. 10010 [Steve Henson] 10011 10012 *) Non-copying interface to BIO pairs. 10013 (still largely untested) 10014 [Bodo Moeller] 10015 10016 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 10017 ASCII string. This was handled independently in various places before. 10018 [Steve Henson] 10019 10020 *) New functions UTF8_getc() and UTF8_putc() that parse and generate 10021 UTF8 strings a character at a time. 10022 [Steve Henson] 10023 10024 *) Use client_version from client hello to select the protocol 10025 (s23_srvr.c) and for RSA client key exchange verification 10026 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 10027 [Bodo Moeller] 10028 10029 *) Add various utility functions to handle SPKACs, these were previously 10030 handled by poking round in the structure internals. Added new function 10031 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 10032 print, verify and generate SPKACs. Based on an original idea from 10033 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 10034 [Steve Henson] 10035 10036 *) RIPEMD160 is operational on all platforms and is back in 'make test'. 10037 [Andy Polyakov] 10038 10039 *) Allow the config file extension section to be overwritten on the 10040 command line. Based on an original idea from Massimiliano Pala 10041 <madwolf@comune.modena.it>. The new option is called -extensions 10042 and can be applied to ca, req and x509. Also -reqexts to override 10043 the request extensions in req and -crlexts to override the crl extensions 10044 in ca. 10045 [Steve Henson] 10046 10047 *) Add new feature to the SPKAC handling in ca. Now you can include 10048 the same field multiple times by preceding it by "XXXX." for example: 10049 1.OU="Unit name 1" 10050 2.OU="Unit name 2" 10051 this is the same syntax as used in the req config file. 10052 [Steve Henson] 10053 10054 *) Allow certificate extensions to be added to certificate requests. These 10055 are specified in a 'req_extensions' option of the req section of the 10056 config file. They can be printed out with the -text option to req but 10057 are otherwise ignored at present. 10058 [Steve Henson] 10059 10060 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 10061 data read consists of only the final block it would not decrypted because 10062 EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 10063 A misplaced 'break' also meant the decrypted final block might not be 10064 copied until the next read. 10065 [Steve Henson] 10066 10067 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 10068 a few extra parameters to the DH structure: these will be useful if 10069 for example we want the value of 'q' or implement X9.42 DH. 10070 [Steve Henson] 10071 10072 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 10073 provides hooks that allow the default DSA functions or functions on a 10074 "per key" basis to be replaced. This allows hardware acceleration and 10075 hardware key storage to be handled without major modification to the 10076 library. Also added low level modexp hooks and CRYPTO_EX structure and 10077 associated functions. 10078 [Steve Henson] 10079 10080 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 10081 as "read only": it can't be written to and the buffer it points to will 10082 not be freed. Reading from a read only BIO is much more efficient than 10083 a normal memory BIO. This was added because there are several times when 10084 an area of memory needs to be read from a BIO. The previous method was 10085 to create a memory BIO and write the data to it, this results in two 10086 copies of the data and an O(n^2) reading algorithm. There is a new 10087 function BIO_new_mem_buf() which creates a read only memory BIO from 10088 an area of memory. Also modified the PKCS#7 routines to use read only 10089 memory BIOs. 10090 [Steve Henson] 10091 10092 *) Bugfix: ssl23_get_client_hello did not work properly when called in 10093 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 10094 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 10095 but a retry condition occured while trying to read the rest. 10096 [Bodo Moeller] 10097 10098 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 10099 NID_pkcs7_encrypted by default: this was wrong since this should almost 10100 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 10101 the encrypted data type: this is a more sensible place to put it and it 10102 allows the PKCS#12 code to be tidied up that duplicated this 10103 functionality. 10104 [Steve Henson] 10105 10106 *) Changed obj_dat.pl script so it takes its input and output files on 10107 the command line. This should avoid shell escape redirection problems 10108 under Win32. 10109 [Steve Henson] 10110 10111 *) Initial support for certificate extension requests, these are included 10112 in things like Xenroll certificate requests. Included functions to allow 10113 extensions to be obtained and added. 10114 [Steve Henson] 10115 10116 *) -crlf option to s_client and s_server for sending newlines as 10117 CRLF (as required by many protocols). 10118 [Bodo Moeller] 10119 10120 Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 10121 10122 *) Install libRSAglue.a when OpenSSL is built with RSAref. 10123 [Ralf S. Engelschall] 10124 10125 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 10126 [Andrija Antonijevic <TheAntony2@bigfoot.com>] 10127 10128 *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 10129 program. 10130 [Steve Henson] 10131 10132 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 10133 DH parameters/keys (q is lost during that conversion, but the resulting 10134 DH parameters contain its length). 10135 10136 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 10137 much faster than DH_generate_parameters (which creates parameters 10138 where p = 2*q + 1), and also the smaller q makes DH computations 10139 much more efficient (160-bit exponentiation instead of 1024-bit 10140 exponentiation); so this provides a convenient way to support DHE 10141 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 10142 utter importance to use 10143 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 10144 or 10145 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 10146 when such DH parameters are used, because otherwise small subgroup 10147 attacks may become possible! 10148 [Bodo Moeller] 10149 10150 *) Avoid memory leak in i2d_DHparams. 10151 [Bodo Moeller] 10152 10153 *) Allow the -k option to be used more than once in the enc program: 10154 this allows the same encrypted message to be read by multiple recipients. 10155 [Steve Henson] 10156 10157 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 10158 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 10159 it will always use the numerical form of the OID, even if it has a short 10160 or long name. 10161 [Steve Henson] 10162 10163 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 10164 method only got called if p,q,dmp1,dmq1,iqmp components were present, 10165 otherwise bn_mod_exp was called. In the case of hardware keys for example 10166 no private key components need be present and it might store extra data 10167 in the RSA structure, which cannot be accessed from bn_mod_exp. 10168 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 10169 private key operations. 10170 [Steve Henson] 10171 10172 *) Added support for SPARC Linux. 10173 [Andy Polyakov] 10174 10175 *) pem_password_cb function type incompatibly changed from 10176 typedef int pem_password_cb(char *buf, int size, int rwflag); 10177 to 10178 ....(char *buf, int size, int rwflag, void *userdata); 10179 so that applications can pass data to their callbacks: 10180 The PEM[_ASN1]_{read,write}... functions and macros now take an 10181 additional void * argument, which is just handed through whenever 10182 the password callback is called. 10183 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 10184 10185 New function SSL_CTX_set_default_passwd_cb_userdata. 10186 10187 Compatibility note: As many C implementations push function arguments 10188 onto the stack in reverse order, the new library version is likely to 10189 interoperate with programs that have been compiled with the old 10190 pem_password_cb definition (PEM_whatever takes some data that 10191 happens to be on the stack as its last argument, and the callback 10192 just ignores this garbage); but there is no guarantee whatsoever that 10193 this will work. 10194 10195 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 10196 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 10197 problems not only on Windows, but also on some Unix platforms. 10198 To avoid problematic command lines, these definitions are now in an 10199 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 10200 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 10201 [Bodo Moeller] 10202 10203 *) MIPS III/IV assembler module is reimplemented. 10204 [Andy Polyakov] 10205 10206 *) More DES library cleanups: remove references to srand/rand and 10207 delete an unused file. 10208 [Ulf M��ller] 10209 10210 *) Add support for the the free Netwide assembler (NASM) under Win32, 10211 since not many people have MASM (ml) and it can be hard to obtain. 10212 This is currently experimental but it seems to work OK and pass all 10213 the tests. Check out INSTALL.W32 for info. 10214 [Steve Henson] 10215 10216 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 10217 without temporary keys kept an extra copy of the server key, 10218 and connections with temporary keys did not free everything in case 10219 of an error. 10220 [Bodo Moeller] 10221 10222 *) New function RSA_check_key and new openssl rsa option -check 10223 for verifying the consistency of RSA keys. 10224 [Ulf Moeller, Bodo Moeller] 10225 10226 *) Various changes to make Win32 compile work: 10227 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 10228 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 10229 comparison" warnings. 10230 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 10231 [Steve Henson] 10232 10233 *) Add a debugging option to PKCS#5 v2 key generation function: when 10234 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 10235 derived keys are printed to stderr. 10236 [Steve Henson] 10237 10238 *) Copy the flags in ASN1_STRING_dup(). 10239 [Roman E. Pavlov <pre@mo.msk.ru>] 10240 10241 *) The x509 application mishandled signing requests containing DSA 10242 keys when the signing key was also DSA and the parameters didn't match. 10243 10244 It was supposed to omit the parameters when they matched the signing key: 10245 the verifying software was then supposed to automatically use the CA's 10246 parameters if they were absent from the end user certificate. 10247 10248 Omitting parameters is no longer recommended. The test was also 10249 the wrong way round! This was probably due to unusual behaviour in 10250 EVP_cmp_parameters() which returns 1 if the parameters match. 10251 This meant that parameters were omitted when they *didn't* match and 10252 the certificate was useless. Certificates signed with 'ca' didn't have 10253 this bug. 10254 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 10255 10256 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 10257 The interface is as follows: 10258 Applications can use 10259 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 10260 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 10261 "off" is now the default. 10262 The library internally uses 10263 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 10264 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 10265 to disable memory-checking temporarily. 10266 10267 Some inconsistent states that previously were possible (and were 10268 even the default) are now avoided. 10269 10270 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 10271 with each memory chunk allocated; this is occasionally more helpful 10272 than just having a counter. 10273 10274 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 10275 10276 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 10277 extensions. 10278 [Bodo Moeller] 10279 10280 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 10281 which largely parallels "options", but is for changing API behaviour, 10282 whereas "options" are about protocol behaviour. 10283 Initial "mode" flags are: 10284 10285 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 10286 a single record has been written. 10287 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 10288 retries use the same buffer location. 10289 (But all of the contents must be 10290 copied!) 10291 [Bodo Moeller] 10292 10293 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 10294 worked. 10295 10296 *) Fix problems with no-hmac etc. 10297 [Ulf M��ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 10298 10299 *) New functions RSA_get_default_method(), RSA_set_method() and 10300 RSA_get_method(). These allows replacement of RSA_METHODs without having 10301 to mess around with the internals of an RSA structure. 10302 [Steve Henson] 10303 10304 *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 10305 Also really enable memory leak checks in openssl.c and in some 10306 test programs. 10307 [Chad C. Mulligan, Bodo Moeller] 10308 10309 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 10310 up the length of negative integers. This has now been simplified to just 10311 store the length when it is first determined and use it later, rather 10312 than trying to keep track of where data is copied and updating it to 10313 point to the end. 10314 [Steve Henson, reported by Brien Wheeler 10315 <bwheeler@authentica-security.com>] 10316 10317 *) Add a new function PKCS7_signatureVerify. This allows the verification 10318 of a PKCS#7 signature but with the signing certificate passed to the 10319 function itself. This contrasts with PKCS7_dataVerify which assumes the 10320 certificate is present in the PKCS#7 structure. This isn't always the 10321 case: certificates can be omitted from a PKCS#7 structure and be 10322 distributed by "out of band" means (such as a certificate database). 10323 [Steve Henson] 10324 10325 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 10326 function prototypes in pem.h, also change util/mkdef.pl to add the 10327 necessary function names. 10328 [Steve Henson] 10329 10330 *) mk1mf.pl (used by Windows builds) did not properly read the 10331 options set by Configure in the top level Makefile, and Configure 10332 was not even able to write more than one option correctly. 10333 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 10334 [Bodo Moeller] 10335 10336 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 10337 file to be loaded from a BIO or FILE pointer. The BIO version will 10338 for example allow memory BIOs to contain config info. 10339 [Steve Henson] 10340 10341 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 10342 Whoever hopes to achieve shared-library compatibility across versions 10343 must use this, not the compile-time macro. 10344 (Exercise 0.9.4: Which is the minimum library version required by 10345 such programs?) 10346 Note: All this applies only to multi-threaded programs, others don't 10347 need locks. 10348 [Bodo Moeller] 10349 10350 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 10351 through a BIO pair triggered the default case, i.e. 10352 SSLerr(...,SSL_R_UNKNOWN_STATE). 10353 [Bodo Moeller] 10354 10355 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 10356 can use the SSL library even if none of the specific BIOs is 10357 appropriate. 10358 [Bodo Moeller] 10359 10360 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 10361 for the encoded length. 10362 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 10363 10364 *) Add initial documentation of the X509V3 functions. 10365 [Steve Henson] 10366 10367 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 10368 PEM_write_bio_PKCS8PrivateKey() that are equivalent to 10369 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 10370 secure PKCS#8 private key format with a high iteration count. 10371 [Steve Henson] 10372 10373 *) Fix determination of Perl interpreter: A perl or perl5 10374 _directory_ in $PATH was also accepted as the interpreter. 10375 [Ralf S. Engelschall] 10376 10377 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 10378 wrong with it but it was very old and did things like calling 10379 PEM_ASN1_read() directly and used MD5 for the hash not to mention some 10380 unusual formatting. 10381 [Steve Henson] 10382 10383 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 10384 to use the new extension code. 10385 [Steve Henson] 10386 10387 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 10388 with macros. This should make it easier to change their form, add extra 10389 arguments etc. Fix a few PEM prototypes which didn't have cipher as a 10390 constant. 10391 [Steve Henson] 10392 10393 *) Add to configuration table a new entry that can specify an alternative 10394 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 10395 according to Mark Crispin <MRC@Panda.COM>. 10396 [Bodo Moeller] 10397 10398#if 0 10399 *) DES CBC did not update the IV. Weird. 10400 [Ben Laurie] 10401#else 10402 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 10403 Changing the behaviour of the former might break existing programs -- 10404 where IV updating is needed, des_ncbc_encrypt can be used. 10405#endif 10406 10407 *) When bntest is run from "make test" it drives bc to check its 10408 calculations, as well as internally checking them. If an internal check 10409 fails, it needs to cause bc to give a non-zero result or make test carries 10410 on without noticing the failure. Fixed. 10411 [Ben Laurie] 10412 10413 *) DES library cleanups. 10414 [Ulf M��ller] 10415 10416 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 10417 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 10418 ciphers. NOTE: although the key derivation function has been verified 10419 against some published test vectors it has not been extensively tested 10420 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 10421 of v2.0. 10422 [Steve Henson] 10423 10424 *) Instead of "mkdir -p", which is not fully portable, use new 10425 Perl script "util/mkdir-p.pl". 10426 [Bodo Moeller] 10427 10428 *) Rewrite the way password based encryption (PBE) is handled. It used to 10429 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 10430 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 10431 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 10432 the 'parameter' field of the AlgorithmIdentifier is passed to the 10433 underlying key generation function so it must do its own ASN1 parsing. 10434 This has also changed the EVP_PBE_CipherInit() function which now has a 10435 'parameter' argument instead of literal salt and iteration count values 10436 and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 10437 [Steve Henson] 10438 10439 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 10440 and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 10441 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 10442 KEY" because this clashed with PKCS#8 unencrypted string. Since this 10443 value was just used as a "magic string" and not used directly its 10444 value doesn't matter. 10445 [Steve Henson] 10446 10447 *) Introduce some semblance of const correctness to BN. Shame C doesn't 10448 support mutable. 10449 [Ben Laurie] 10450 10451 *) "linux-sparc64" configuration (ultrapenguin). 10452 [Ray Miller <ray.miller@oucs.ox.ac.uk>] 10453 "linux-sparc" configuration. 10454 [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 10455 10456 *) config now generates no-xxx options for missing ciphers. 10457 [Ulf M��ller] 10458 10459 *) Support the EBCDIC character set (work in progress). 10460 File ebcdic.c not yet included because it has a different license. 10461 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 10462 10463 *) Support BS2000/OSD-POSIX. 10464 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 10465 10466 *) Make callbacks for key generation use void * instead of char *. 10467 [Ben Laurie] 10468 10469 *) Make S/MIME samples compile (not yet tested). 10470 [Ben Laurie] 10471 10472 *) Additional typesafe stacks. 10473 [Ben Laurie] 10474 10475 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 10476 [Bodo Moeller] 10477 10478 10479 Changes between 0.9.3 and 0.9.3a [29 May 1999] 10480 10481 *) New configuration variant "sco5-gcc". 10482 10483 *) Updated some demos. 10484 [Sean O Riordain, Wade Scholine] 10485 10486 *) Add missing BIO_free at exit of pkcs12 application. 10487 [Wu Zhigang] 10488 10489 *) Fix memory leak in conf.c. 10490 [Steve Henson] 10491 10492 *) Updates for Win32 to assembler version of MD5. 10493 [Steve Henson] 10494 10495 *) Set #! path to perl in apps/der_chop to where we found it 10496 instead of using a fixed path. 10497 [Bodo Moeller] 10498 10499 *) SHA library changes for irix64-mips4-cc. 10500 [Andy Polyakov] 10501 10502 *) Improvements for VMS support. 10503 [Richard Levitte] 10504 10505 10506 Changes between 0.9.2b and 0.9.3 [24 May 1999] 10507 10508 *) Bignum library bug fix. IRIX 6 passes "make test" now! 10509 This also avoids the problems with SC4.2 and unpatched SC5. 10510 [Andy Polyakov <appro@fy.chalmers.se>] 10511 10512 *) New functions sk_num, sk_value and sk_set to replace the previous macros. 10513 These are required because of the typesafe stack would otherwise break 10514 existing code. If old code used a structure member which used to be STACK 10515 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 10516 sk_num or sk_value it would produce an error because the num, data members 10517 are not present in STACK_OF. Now it just produces a warning. sk_set 10518 replaces the old method of assigning a value to sk_value 10519 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 10520 that does this will no longer work (and should use sk_set instead) but 10521 this could be regarded as a "questionable" behaviour anyway. 10522 [Steve Henson] 10523 10524 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 10525 correctly handle encrypted S/MIME data. 10526 [Steve Henson] 10527 10528 *) Change type of various DES function arguments from des_cblock 10529 (which means, in function argument declarations, pointer to char) 10530 to des_cblock * (meaning pointer to array with 8 char elements), 10531 which allows the compiler to do more typechecking; it was like 10532 that back in SSLeay, but with lots of ugly casts. 10533 10534 Introduce new type const_des_cblock. 10535 [Bodo Moeller] 10536 10537 *) Reorganise the PKCS#7 library and get rid of some of the more obvious 10538 problems: find RecipientInfo structure that matches recipient certificate 10539 and initialise the ASN1 structures properly based on passed cipher. 10540 [Steve Henson] 10541 10542 *) Belatedly make the BN tests actually check the results. 10543 [Ben Laurie] 10544 10545 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 10546 to and from BNs: it was completely broken. New compilation option 10547 NEG_PUBKEY_BUG to allow for some broken certificates that encode public 10548 key elements as negative integers. 10549 [Steve Henson] 10550 10551 *) Reorganize and speed up MD5. 10552 [Andy Polyakov <appro@fy.chalmers.se>] 10553 10554 *) VMS support. 10555 [Richard Levitte <richard@levitte.org>] 10556 10557 *) New option -out to asn1parse to allow the parsed structure to be 10558 output to a file. This is most useful when combined with the -strparse 10559 option to examine the output of things like OCTET STRINGS. 10560 [Steve Henson] 10561 10562 *) Make SSL library a little more fool-proof by not requiring any longer 10563 that SSL_set_{accept,connect}_state be called before 10564 SSL_{accept,connect} may be used (SSL_set_..._state is omitted 10565 in many applications because usually everything *appeared* to work as 10566 intended anyway -- now it really works as intended). 10567 [Bodo Moeller] 10568 10569 *) Move openssl.cnf out of lib/. 10570 [Ulf M��ller] 10571 10572 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 10573 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 10574 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 10575 [Ralf S. Engelschall] 10576 10577 *) Various fixes to the EVP and PKCS#7 code. It may now be able to 10578 handle PKCS#7 enveloped data properly. 10579 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 10580 10581 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 10582 copying pointers. The cert_st handling is changed by this in 10583 various ways (and thus what used to be known as ctx->default_cert 10584 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 10585 any longer when s->cert does not give us what we need). 10586 ssl_cert_instantiate becomes obsolete by this change. 10587 As soon as we've got the new code right (possibly it already is?), 10588 we have solved a couple of bugs of the earlier code where s->cert 10589 was used as if it could not have been shared with other SSL structures. 10590 10591 Note that using the SSL API in certain dirty ways now will result 10592 in different behaviour than observed with earlier library versions: 10593 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 10594 does not influence s as it used to. 10595 10596 In order to clean up things more thoroughly, inside SSL_SESSION 10597 we don't use CERT any longer, but a new structure SESS_CERT 10598 that holds per-session data (if available); currently, this is 10599 the peer's certificate chain and, for clients, the server's certificate 10600 and temporary key. CERT holds only those values that can have 10601 meaningful defaults in an SSL_CTX. 10602 [Bodo Moeller] 10603 10604 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 10605 from the internal representation. Various PKCS#7 fixes: remove some 10606 evil casts and set the enc_dig_alg field properly based on the signing 10607 key type. 10608 [Steve Henson] 10609 10610 *) Allow PKCS#12 password to be set from the command line or the 10611 environment. Let 'ca' get its config file name from the environment 10612 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 10613 and 'x509'). 10614 [Steve Henson] 10615 10616 *) Allow certificate policies extension to use an IA5STRING for the 10617 organization field. This is contrary to the PKIX definition but 10618 VeriSign uses it and IE5 only recognises this form. Document 'x509' 10619 extension option. 10620 [Steve Henson] 10621 10622 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 10623 without disallowing inline assembler and the like for non-pedantic builds. 10624 [Ben Laurie] 10625 10626 *) Support Borland C++ builder. 10627 [Janez Jere <jj@void.si>, modified by Ulf M��ller] 10628 10629 *) Support Mingw32. 10630 [Ulf M��ller] 10631 10632 *) SHA-1 cleanups and performance enhancements. 10633 [Andy Polyakov <appro@fy.chalmers.se>] 10634 10635 *) Sparc v8plus assembler for the bignum library. 10636 [Andy Polyakov <appro@fy.chalmers.se>] 10637 10638 *) Accept any -xxx and +xxx compiler options in Configure. 10639 [Ulf M��ller] 10640 10641 *) Update HPUX configuration. 10642 [Anonymous] 10643 10644 *) Add missing sk_<type>_unshift() function to safestack.h 10645 [Ralf S. Engelschall] 10646 10647 *) New function SSL_CTX_use_certificate_chain_file that sets the 10648 "extra_cert"s in addition to the certificate. (This makes sense 10649 only for "PEM" format files, as chains as a whole are not 10650 DER-encoded.) 10651 [Bodo Moeller] 10652 10653 *) Support verify_depth from the SSL API. 10654 x509_vfy.c had what can be considered an off-by-one-error: 10655 Its depth (which was not part of the external interface) 10656 was actually counting the number of certificates in a chain; 10657 now it really counts the depth. 10658 [Bodo Moeller] 10659 10660 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 10661 instead of X509err, which often resulted in confusing error 10662 messages since the error codes are not globally unique 10663 (e.g. an alleged error in ssl3_accept when a certificate 10664 didn't match the private key). 10665 10666 *) New function SSL_CTX_set_session_id_context that allows to set a default 10667 value (so that you don't need SSL_set_session_id_context for each 10668 connection using the SSL_CTX). 10669 [Bodo Moeller] 10670 10671 *) OAEP decoding bug fix. 10672 [Ulf M��ller] 10673 10674 *) Support INSTALL_PREFIX for package builders, as proposed by 10675 David Harris. 10676 [Bodo Moeller] 10677 10678 *) New Configure options "threads" and "no-threads". For systems 10679 where the proper compiler options are known (currently Solaris 10680 and Linux), "threads" is the default. 10681 [Bodo Moeller] 10682 10683 *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 10684 [Bodo Moeller] 10685 10686 *) Install various scripts to $(OPENSSLDIR)/misc, not to 10687 $(INSTALLTOP)/bin -- they shouldn't clutter directories 10688 such as /usr/local/bin. 10689 [Bodo Moeller] 10690 10691 *) "make linux-shared" to build shared libraries. 10692 [Niels Poppe <niels@netbox.org>] 10693 10694 *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 10695 [Ulf M��ller] 10696 10697 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 10698 extension adding in x509 utility. 10699 [Steve Henson] 10700 10701 *) Remove NOPROTO sections and error code comments. 10702 [Ulf M��ller] 10703 10704 *) Partial rewrite of the DEF file generator to now parse the ANSI 10705 prototypes. 10706 [Steve Henson] 10707 10708 *) New Configure options --prefix=DIR and --openssldir=DIR. 10709 [Ulf M��ller] 10710 10711 *) Complete rewrite of the error code script(s). It is all now handled 10712 by one script at the top level which handles error code gathering, 10713 header rewriting and C source file generation. It should be much better 10714 than the old method: it now uses a modified version of Ulf's parser to 10715 read the ANSI prototypes in all header files (thus the old K&R definitions 10716 aren't needed for error creation any more) and do a better job of 10717 translating function codes into names. The old 'ASN1 error code imbedded 10718 in a comment' is no longer necessary and it doesn't use .err files which 10719 have now been deleted. Also the error code call doesn't have to appear all 10720 on one line (which resulted in some large lines...). 10721 [Steve Henson] 10722 10723 *) Change #include filenames from <foo.h> to <openssl/foo.h>. 10724 [Bodo Moeller] 10725 10726 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 10727 0 (which usually indicates a closed connection), but continue reading. 10728 [Bodo Moeller] 10729 10730 *) Fix some race conditions. 10731 [Bodo Moeller] 10732 10733 *) Add support for CRL distribution points extension. Add Certificate 10734 Policies and CRL distribution points documentation. 10735 [Steve Henson] 10736 10737 *) Move the autogenerated header file parts to crypto/opensslconf.h. 10738 [Ulf M��ller] 10739 10740 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 10741 8 of keying material. Merlin has also confirmed interop with this fix 10742 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 10743 [Merlin Hughes <merlin@baltimore.ie>] 10744 10745 *) Fix lots of warnings. 10746 [Richard Levitte <levitte@stacken.kth.se>] 10747 10748 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 10749 the directory spec didn't end with a LIST_SEPARATOR_CHAR. 10750 [Richard Levitte <levitte@stacken.kth.se>] 10751 10752 *) Fix problems with sizeof(long) == 8. 10753 [Andy Polyakov <appro@fy.chalmers.se>] 10754 10755 *) Change functions to ANSI C. 10756 [Ulf M��ller] 10757 10758 *) Fix typos in error codes. 10759 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M��ller] 10760 10761 *) Remove defunct assembler files from Configure. 10762 [Ulf M��ller] 10763 10764 *) SPARC v8 assembler BIGNUM implementation. 10765 [Andy Polyakov <appro@fy.chalmers.se>] 10766 10767 *) Support for Certificate Policies extension: both print and set. 10768 Various additions to support the r2i method this uses. 10769 [Steve Henson] 10770 10771 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 10772 return a const string when you are expecting an allocated buffer. 10773 [Ben Laurie] 10774 10775 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 10776 types DirectoryString and DisplayText. 10777 [Steve Henson] 10778 10779 *) Add code to allow r2i extensions to access the configuration database, 10780 add an LHASH database driver and add several ctx helper functions. 10781 [Steve Henson] 10782 10783 *) Fix an evil bug in bn_expand2() which caused various BN functions to 10784 fail when they extended the size of a BIGNUM. 10785 [Steve Henson] 10786 10787 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 10788 support typesafe stack. 10789 [Steve Henson] 10790 10791 *) Fix typo in SSL_[gs]et_options(). 10792 [Nils Frostberg <nils@medcom.se>] 10793 10794 *) Delete various functions and files that belonged to the (now obsolete) 10795 old X509V3 handling code. 10796 [Steve Henson] 10797 10798 *) New Configure option "rsaref". 10799 [Ulf M��ller] 10800 10801 *) Don't auto-generate pem.h. 10802 [Bodo Moeller] 10803 10804 *) Introduce type-safe ASN.1 SETs. 10805 [Ben Laurie] 10806 10807 *) Convert various additional casted stacks to type-safe STACK_OF() variants. 10808 [Ben Laurie, Ralf S. Engelschall, Steve Henson] 10809 10810 *) Introduce type-safe STACKs. This will almost certainly break lots of code 10811 that links with OpenSSL (well at least cause lots of warnings), but fear 10812 not: the conversion is trivial, and it eliminates loads of evil casts. A 10813 few STACKed things have been converted already. Feel free to convert more. 10814 In the fullness of time, I'll do away with the STACK type altogether. 10815 [Ben Laurie] 10816 10817 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 10818 specified in <certfile> by updating the entry in the index.txt file. 10819 This way one no longer has to edit the index.txt file manually for 10820 revoking a certificate. The -revoke option does the gory details now. 10821 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 10822 10823 *) Fix `openssl crl -noout -text' combination where `-noout' killed the 10824 `-text' option at all and this way the `-noout -text' combination was 10825 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 10826 [Ralf S. Engelschall] 10827 10828 *) Make sure a corresponding plain text error message exists for the 10829 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 10830 verify callback function determined that a certificate was revoked. 10831 [Ralf S. Engelschall] 10832 10833 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 10834 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 10835 all available cipers including rc5, which was forgotten until now. 10836 In order to let the testing shell script know which algorithms 10837 are available, a new (up to now undocumented) command 10838 "openssl list-cipher-commands" is used. 10839 [Bodo Moeller] 10840 10841 *) Bugfix: s_client occasionally would sleep in select() when 10842 it should have checked SSL_pending() first. 10843 [Bodo Moeller] 10844 10845 *) New functions DSA_do_sign and DSA_do_verify to provide access to 10846 the raw DSA values prior to ASN.1 encoding. 10847 [Ulf M��ller] 10848 10849 *) Tweaks to Configure 10850 [Niels Poppe <niels@netbox.org>] 10851 10852 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 10853 yet... 10854 [Steve Henson] 10855 10856 *) New variables $(RANLIB) and $(PERL) in the Makefiles. 10857 [Ulf M��ller] 10858 10859 *) New config option to avoid instructions that are illegal on the 80386. 10860 The default code is faster, but requires at least a 486. 10861 [Ulf M��ller] 10862 10863 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 10864 SSL2_SERVER_VERSION (not used at all) macros, which are now the 10865 same as SSL2_VERSION anyway. 10866 [Bodo Moeller] 10867 10868 *) New "-showcerts" option for s_client. 10869 [Bodo Moeller] 10870 10871 *) Still more PKCS#12 integration. Add pkcs12 application to openssl 10872 application. Various cleanups and fixes. 10873 [Steve Henson] 10874 10875 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 10876 modify error routines to work internally. Add error codes and PBE init 10877 to library startup routines. 10878 [Steve Henson] 10879 10880 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 10881 packing functions to asn1 and evp. Changed function names and error 10882 codes along the way. 10883 [Steve Henson] 10884 10885 *) PKCS12 integration: and so it begins... First of several patches to 10886 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 10887 objects to objects.h 10888 [Steve Henson] 10889 10890 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 10891 and display support for Thawte strong extranet extension. 10892 [Steve Henson] 10893 10894 *) Add LinuxPPC support. 10895 [Jeff Dubrule <igor@pobox.org>] 10896 10897 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 10898 bn_div_words in alpha.s. 10899 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 10900 10901 *) Make sure the RSA OAEP test is skipped under -DRSAref because 10902 OAEP isn't supported when OpenSSL is built with RSAref. 10903 [Ulf Moeller <ulf@fitug.de>] 10904 10905 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 10906 so they no longer are missing under -DNOPROTO. 10907 [Soren S. Jorvang <soren@t.dk>] 10908 10909 10910 Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 10911 10912 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 10913 doesn't work when the session is reused. Coming soon! 10914 [Ben Laurie] 10915 10916 *) Fix a security hole, that allows sessions to be reused in the wrong 10917 context thus bypassing client cert protection! All software that uses 10918 client certs and session caches in multiple contexts NEEDS PATCHING to 10919 allow session reuse! A fuller solution is in the works. 10920 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 10921 10922 *) Some more source tree cleanups (removed obsolete files 10923 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 10924 permission on "config" script to be executable) and a fix for the INSTALL 10925 document. 10926 [Ulf Moeller <ulf@fitug.de>] 10927 10928 *) Remove some legacy and erroneous uses of malloc, free instead of 10929 Malloc, Free. 10930 [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 10931 10932 *) Make rsa_oaep_test return non-zero on error. 10933 [Ulf Moeller <ulf@fitug.de>] 10934 10935 *) Add support for native Solaris shared libraries. Configure 10936 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 10937 if someone would make that last step automatic. 10938 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 10939 10940 *) ctx_size was not built with the right compiler during "make links". Fixed. 10941 [Ben Laurie] 10942 10943 *) Change the meaning of 'ALL' in the cipher list. It now means "everything 10944 except NULL ciphers". This means the default cipher list will no longer 10945 enable NULL ciphers. They need to be specifically enabled e.g. with 10946 the string "DEFAULT:eNULL". 10947 [Steve Henson] 10948 10949 *) Fix to RSA private encryption routines: if p < q then it would 10950 occasionally produce an invalid result. This will only happen with 10951 externally generated keys because OpenSSL (and SSLeay) ensure p > q. 10952 [Steve Henson] 10953 10954 *) Be less restrictive and allow also `perl util/perlpath.pl 10955 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 10956 because this way one can also use an interpreter named `perl5' (which is 10957 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 10958 installed as `perl'). 10959 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 10960 10961 *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 10962 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 10963 10964 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 10965 advapi32.lib to Win32 build and change the pem test comparision 10966 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 10967 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 10968 and crypto/des/ede_cbcm_enc.c. 10969 [Steve Henson] 10970 10971 *) DES quad checksum was broken on big-endian architectures. Fixed. 10972 [Ben Laurie] 10973 10974 *) Comment out two functions in bio.h that aren't implemented. Fix up the 10975 Win32 test batch file so it (might) work again. The Win32 test batch file 10976 is horrible: I feel ill.... 10977 [Steve Henson] 10978 10979 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 10980 in e_os.h. Audit of header files to check ANSI and non ANSI 10981 sections: 10 functions were absent from non ANSI section and not exported 10982 from Windows DLLs. Fixed up libeay.num for new functions. 10983 [Steve Henson] 10984 10985 *) Make `openssl version' output lines consistent. 10986 [Ralf S. Engelschall] 10987 10988 *) Fix Win32 symbol export lists for BIO functions: Added 10989 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 10990 to ms/libeay{16,32}.def. 10991 [Ralf S. Engelschall] 10992 10993 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 10994 fine under Unix and passes some trivial tests I've now added. But the 10995 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 10996 added to make sure no one expects that this stuff really works in the 10997 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 10998 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 10999 openssl_bio.xs. 11000 [Ralf S. Engelschall] 11001 11002 *) Fix the generation of two part addresses in perl. 11003 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 11004 11005 *) Add config entry for Linux on MIPS. 11006 [John Tobey <jtobey@channel1.com>] 11007 11008 *) Make links whenever Configure is run, unless we are on Windoze. 11009 [Ben Laurie] 11010 11011 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 11012 Currently only issuerAltName and AuthorityKeyIdentifier make any sense 11013 in CRLs. 11014 [Steve Henson] 11015 11016 *) Add a useful kludge to allow package maintainers to specify compiler and 11017 other platforms details on the command line without having to patch the 11018 Configure script everytime: One now can use ``perl Configure 11019 <id>:<details>'', i.e. platform ids are allowed to have details appended 11020 to them (seperated by colons). This is treated as there would be a static 11021 pre-configured entry in Configure's %table under key <id> with value 11022 <details> and ``perl Configure <id>'' is called. So, when you want to 11023 perform a quick test-compile under FreeBSD 3.1 with pgcc and without 11024 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 11025 now, which overrides the FreeBSD-elf entry on-the-fly. 11026 [Ralf S. Engelschall] 11027 11028 *) Disable new TLS1 ciphersuites by default: they aren't official yet. 11029 [Ben Laurie] 11030 11031 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 11032 on the `perl Configure ...' command line. This way one can compile 11033 OpenSSL libraries with Position Independent Code (PIC) which is needed 11034 for linking it into DSOs. 11035 [Ralf S. Engelschall] 11036 11037 *) Remarkably, export ciphers were totally broken and no-one had noticed! 11038 Fixed. 11039 [Ben Laurie] 11040 11041 *) Cleaned up the LICENSE document: The official contact for any license 11042 questions now is the OpenSSL core team under openssl-core@openssl.org. 11043 And add a paragraph about the dual-license situation to make sure people 11044 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 11045 to the OpenSSL toolkit. 11046 [Ralf S. Engelschall] 11047 11048 *) General source tree makefile cleanups: Made `making xxx in yyy...' 11049 display consistent in the source tree and replaced `/bin/rm' by `rm'. 11050 Additonally cleaned up the `make links' target: Remove unnecessary 11051 semicolons, subsequent redundant removes, inline point.sh into mklink.sh 11052 to speed processing and no longer clutter the display with confusing 11053 stuff. Instead only the actually done links are displayed. 11054 [Ralf S. Engelschall] 11055 11056 *) Permit null encryption ciphersuites, used for authentication only. It used 11057 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 11058 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 11059 encryption. 11060 [Ben Laurie] 11061 11062 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 11063 signed attributes when verifying signatures (this would break them), 11064 the detached data encoding was wrong and public keys obtained using 11065 X509_get_pubkey() weren't freed. 11066 [Steve Henson] 11067 11068 *) Add text documentation for the BUFFER functions. Also added a work around 11069 to a Win95 console bug. This was triggered by the password read stuff: the 11070 last character typed gets carried over to the next fread(). If you were 11071 generating a new cert request using 'req' for example then the last 11072 character of the passphrase would be CR which would then enter the first 11073 field as blank. 11074 [Steve Henson] 11075 11076 *) Added the new `Includes OpenSSL Cryptography Software' button as 11077 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 11078 button and can be used by applications based on OpenSSL to show the 11079 relationship to the OpenSSL project. 11080 [Ralf S. Engelschall] 11081 11082 *) Remove confusing variables in function signatures in files 11083 ssl/ssl_lib.c and ssl/ssl.h. 11084 [Lennart Bong <lob@kulthea.stacken.kth.se>] 11085 11086 *) Don't install bss_file.c under PREFIX/include/ 11087 [Lennart Bong <lob@kulthea.stacken.kth.se>] 11088 11089 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 11090 functions that return function pointers and has support for NT specific 11091 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 11092 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 11093 unsigned to signed types: this was killing the Win32 compile. 11094 [Steve Henson] 11095 11096 *) Add new certificate file to stack functions, 11097 SSL_add_dir_cert_subjects_to_stack() and 11098 SSL_add_file_cert_subjects_to_stack(). These largely supplant 11099 SSL_load_client_CA_file(), and can be used to add multiple certs easily 11100 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 11101 This means that Apache-SSL and similar packages don't have to mess around 11102 to add as many CAs as they want to the preferred list. 11103 [Ben Laurie] 11104 11105 *) Experiment with doxygen documentation. Currently only partially applied to 11106 ssl/ssl_lib.c. 11107 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 11108 openssl.doxy as the configuration file. 11109 [Ben Laurie] 11110 11111 *) Get rid of remaining C++-style comments which strict C compilers hate. 11112 [Ralf S. Engelschall, pointed out by Carlos Amengual] 11113 11114 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 11115 compiled in by default: it has problems with large keys. 11116 [Steve Henson] 11117 11118 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 11119 DH private keys and/or callback functions which directly correspond to 11120 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 11121 is needed for applications which have to configure certificates on a 11122 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 11123 (e.g. s_server). 11124 For the RSA certificate situation is makes no difference, but 11125 for the DSA certificate situation this fixes the "no shared cipher" 11126 problem where the OpenSSL cipher selection procedure failed because the 11127 temporary keys were not overtaken from the context and the API provided 11128 no way to reconfigure them. 11129 The new functions now let applications reconfigure the stuff and they 11130 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 11131 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 11132 non-public-API function ssl_cert_instantiate() is used as a helper 11133 function and also to reduce code redundancy inside ssl_rsa.c. 11134 [Ralf S. Engelschall] 11135 11136 *) Move s_server -dcert and -dkey options out of the undocumented feature 11137 area because they are useful for the DSA situation and should be 11138 recognized by the users. 11139 [Ralf S. Engelschall] 11140 11141 *) Fix the cipher decision scheme for export ciphers: the export bits are 11142 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 11143 SSL_EXP_MASK. So, the original variable has to be used instead of the 11144 already masked variable. 11145 [Richard Levitte <levitte@stacken.kth.se>] 11146 11147 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 11148 [Richard Levitte <levitte@stacken.kth.se>] 11149 11150 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 11151 from `int' to `unsigned int' because it's a length and initialized by 11152 EVP_DigestFinal() which expects an `unsigned int *'. 11153 [Richard Levitte <levitte@stacken.kth.se>] 11154 11155 *) Don't hard-code path to Perl interpreter on shebang line of Configure 11156 script. Instead use the usual Shell->Perl transition trick. 11157 [Ralf S. Engelschall] 11158 11159 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 11160 (in addition to RSA certificates) to match the behaviour of `openssl dsa 11161 -noout -modulus' as it's already the case for `openssl rsa -noout 11162 -modulus'. For RSA the -modulus is the real "modulus" while for DSA 11163 currently the public key is printed (a decision which was already done by 11164 `openssl dsa -modulus' in the past) which serves a similar purpose. 11165 Additionally the NO_RSA no longer completely removes the whole -modulus 11166 option; it now only avoids using the RSA stuff. Same applies to NO_DSA 11167 now, too. 11168 [Ralf S. Engelschall] 11169 11170 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 11171 BIO. See the source (crypto/evp/bio_ok.c) for more info. 11172 [Arne Ansper <arne@ats.cyber.ee>] 11173 11174 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 11175 to be added. Now both 'req' and 'ca' can use new objects defined in the 11176 config file. 11177 [Steve Henson] 11178 11179 *) Add cool BIO that does syslog (or event log on NT). 11180 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 11181 11182 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 11183 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 11184 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 11185 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 11186 [Ben Laurie] 11187 11188 *) Add preliminary config info for new extension code. 11189 [Steve Henson] 11190 11191 *) Make RSA_NO_PADDING really use no padding. 11192 [Ulf Moeller <ulf@fitug.de>] 11193 11194 *) Generate errors when private/public key check is done. 11195 [Ben Laurie] 11196 11197 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 11198 for some CRL extensions and new objects added. 11199 [Steve Henson] 11200 11201 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 11202 key usage extension and fuller support for authority key id. 11203 [Steve Henson] 11204 11205 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 11206 padding method for RSA, which is recommended for new applications in PKCS 11207 #1 v2.0 (RFC 2437, October 1998). 11208 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 11209 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 11210 against Bleichbacher's attack on RSA. 11211 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 11212 Ben Laurie] 11213 11214 *) Updates to the new SSL compression code 11215 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 11216 11217 *) Fix so that the version number in the master secret, when passed 11218 via RSA, checks that if TLS was proposed, but we roll back to SSLv3 11219 (because the server will not accept higher), that the version number 11220 is 0x03,0x01, not 0x03,0x00 11221 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 11222 11223 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 11224 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 11225 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 11226 [Steve Henson] 11227 11228 *) Support for RAW extensions where an arbitrary extension can be 11229 created by including its DER encoding. See apps/openssl.cnf for 11230 an example. 11231 [Steve Henson] 11232 11233 *) Make sure latest Perl versions don't interpret some generated C array 11234 code as Perl array code in the crypto/err/err_genc.pl script. 11235 [Lars Weber <3weber@informatik.uni-hamburg.de>] 11236 11237 *) Modify ms/do_ms.bat to not generate assembly language makefiles since 11238 not many people have the assembler. Various Win32 compilation fixes and 11239 update to the INSTALL.W32 file with (hopefully) more accurate Win32 11240 build instructions. 11241 [Steve Henson] 11242 11243 *) Modify configure script 'Configure' to automatically create crypto/date.h 11244 file under Win32 and also build pem.h from pem.org. New script 11245 util/mkfiles.pl to create the MINFO file on environments that can't do a 11246 'make files': perl util/mkfiles.pl >MINFO should work. 11247 [Steve Henson] 11248 11249 *) Major rework of DES function declarations, in the pursuit of correctness 11250 and purity. As a result, many evil casts evaporated, and some weirdness, 11251 too. You may find this causes warnings in your code. Zapping your evil 11252 casts will probably fix them. Mostly. 11253 [Ben Laurie] 11254 11255 *) Fix for a typo in asn1.h. Bug fix to object creation script 11256 obj_dat.pl. It considered a zero in an object definition to mean 11257 "end of object": none of the objects in objects.h have any zeros 11258 so it wasn't spotted. 11259 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 11260 11261 *) Add support for Triple DES Cipher Block Chaining with Output Feedback 11262 Masking (CBCM). In the absence of test vectors, the best I have been able 11263 to do is check that the decrypt undoes the encrypt, so far. Send me test 11264 vectors if you have them. 11265 [Ben Laurie] 11266 11267 *) Correct calculation of key length for export ciphers (too much space was 11268 allocated for null ciphers). This has not been tested! 11269 [Ben Laurie] 11270 11271 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 11272 message is now correct (it understands "crypto" and "ssl" on its 11273 command line). There is also now an "update" option. This will update 11274 the util/ssleay.num and util/libeay.num files with any new functions. 11275 If you do a: 11276 perl util/mkdef.pl crypto ssl update 11277 it will update them. 11278 [Steve Henson] 11279 11280 *) Overhauled the Perl interface (perl/*): 11281 - ported BN stuff to OpenSSL's different BN library 11282 - made the perl/ source tree CVS-aware 11283 - renamed the package from SSLeay to OpenSSL (the files still contain 11284 their history because I've copied them in the repository) 11285 - removed obsolete files (the test scripts will be replaced 11286 by better Test::Harness variants in the future) 11287 [Ralf S. Engelschall] 11288 11289 *) First cut for a very conservative source tree cleanup: 11290 1. merge various obsolete readme texts into doc/ssleay.txt 11291 where we collect the old documents and readme texts. 11292 2. remove the first part of files where I'm already sure that we no 11293 longer need them because of three reasons: either they are just temporary 11294 files which were left by Eric or they are preserved original files where 11295 I've verified that the diff is also available in the CVS via "cvs diff 11296 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 11297 the crypto/md/ stuff). 11298 [Ralf S. Engelschall] 11299 11300 *) More extension code. Incomplete support for subject and issuer alt 11301 name, issuer and authority key id. Change the i2v function parameters 11302 and add an extra 'crl' parameter in the X509V3_CTX structure: guess 11303 what that's for :-) Fix to ASN1 macro which messed up 11304 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 11305 [Steve Henson] 11306 11307 *) Preliminary support for ENUMERATED type. This is largely copied from the 11308 INTEGER code. 11309 [Steve Henson] 11310 11311 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 11312 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 11313 11314 *) Make sure `make rehash' target really finds the `openssl' program. 11315 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 11316 11317 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 11318 like to hear about it if this slows down other processors. 11319 [Ben Laurie] 11320 11321 *) Add CygWin32 platform information to Configure script. 11322 [Alan Batie <batie@aahz.jf.intel.com>] 11323 11324 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 11325 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 11326 11327 *) New program nseq to manipulate netscape certificate sequences 11328 [Steve Henson] 11329 11330 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 11331 few typos. 11332 [Steve Henson] 11333 11334 *) Fixes to BN code. Previously the default was to define BN_RECURSION 11335 but the BN code had some problems that would cause failures when 11336 doing certificate verification and some other functions. 11337 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 11338 11339 *) Add ASN1 and PEM code to support netscape certificate sequences. 11340 [Steve Henson] 11341 11342 *) Add ASN1 and PEM code to support netscape certificate sequences. 11343 [Steve Henson] 11344 11345 *) Add several PKIX and private extended key usage OIDs. 11346 [Steve Henson] 11347 11348 *) Modify the 'ca' program to handle the new extension code. Modify 11349 openssl.cnf for new extension format, add comments. 11350 [Steve Henson] 11351 11352 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 11353 and add a sample to openssl.cnf so req -x509 now adds appropriate 11354 CA extensions. 11355 [Steve Henson] 11356 11357 *) Continued X509 V3 changes. Add to other makefiles, integrate with the 11358 error code, add initial support to X509_print() and x509 application. 11359 [Steve Henson] 11360 11361 *) Takes a deep breath and start addding X509 V3 extension support code. Add 11362 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 11363 stuff is currently isolated and isn't even compiled yet. 11364 [Steve Henson] 11365 11366 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 11367 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 11368 Removed the versions check from X509 routines when loading extensions: 11369 this allows certain broken certificates that don't set the version 11370 properly to be processed. 11371 [Steve Henson] 11372 11373 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 11374 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 11375 can still be regenerated with "make depend". 11376 [Ben Laurie] 11377 11378 *) Spelling mistake in C version of CAST-128. 11379 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 11380 11381 *) Changes to the error generation code. The perl script err-code.pl 11382 now reads in the old error codes and retains the old numbers, only 11383 adding new ones if necessary. It also only changes the .err files if new 11384 codes are added. The makefiles have been modified to only insert errors 11385 when needed (to avoid needlessly modifying header files). This is done 11386 by only inserting errors if the .err file is newer than the auto generated 11387 C file. To rebuild all the error codes from scratch (the old behaviour) 11388 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 11389 or delete all the .err files. 11390 [Steve Henson] 11391 11392 *) CAST-128 was incorrectly implemented for short keys. The C version has 11393 been fixed, but is untested. The assembler versions are also fixed, but 11394 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 11395 to regenerate it if needed. 11396 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 11397 Hagino <itojun@kame.net>] 11398 11399 *) File was opened incorrectly in randfile.c. 11400 [Ulf M��ller <ulf@fitug.de>] 11401 11402 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 11403 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 11404 GeneralizedTime. ASN1_TIME is the proper type used in certificates et 11405 al: it's just almost always a UTCTime. Note this patch adds new error 11406 codes so do a "make errors" if there are problems. 11407 [Steve Henson] 11408 11409 *) Correct Linux 1 recognition in config. 11410 [Ulf M��ller <ulf@fitug.de>] 11411 11412 *) Remove pointless MD5 hash when using DSA keys in ca. 11413 [Anonymous <nobody@replay.com>] 11414 11415 *) Generate an error if given an empty string as a cert directory. Also 11416 generate an error if handed NULL (previously returned 0 to indicate an 11417 error, but didn't set one). 11418 [Ben Laurie, reported by Anonymous <nobody@replay.com>] 11419 11420 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 11421 [Ben Laurie] 11422 11423 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 11424 parameters. This was causing a warning which killed off the Win32 compile. 11425 [Steve Henson] 11426 11427 *) Remove C++ style comments from crypto/bn/bn_local.h. 11428 [Neil Costigan <neil.costigan@celocom.com>] 11429 11430 *) The function OBJ_txt2nid was broken. It was supposed to return a nid 11431 based on a text string, looking up short and long names and finally 11432 "dot" format. The "dot" format stuff didn't work. Added new function 11433 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 11434 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 11435 OID is not part of the table. 11436 [Steve Henson] 11437 11438 *) Add prototypes to X509 lookup/verify methods, fixing a bug in 11439 X509_LOOKUP_by_alias(). 11440 [Ben Laurie] 11441 11442 *) Sort openssl functions by name. 11443 [Ben Laurie] 11444 11445 *) Get the gendsa program working (hopefully) and add it to app list. Remove 11446 encryption from sample DSA keys (in case anyone is interested the password 11447 was "1234"). 11448 [Steve Henson] 11449 11450 *) Make _all_ *_free functions accept a NULL pointer. 11451 [Frans Heymans <fheymans@isaserver.be>] 11452 11453 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 11454 NULL pointers. 11455 [Anonymous <nobody@replay.com>] 11456 11457 *) s_server should send the CAfile as acceptable CAs, not its own cert. 11458 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 11459 11460 *) Don't blow it for numeric -newkey arguments to apps/req. 11461 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 11462 11463 *) Temp key "for export" tests were wrong in s3_srvr.c. 11464 [Anonymous <nobody@replay.com>] 11465 11466 *) Add prototype for temp key callback functions 11467 SSL_CTX_set_tmp_{rsa,dh}_callback(). 11468 [Ben Laurie] 11469 11470 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 11471 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 11472 [Steve Henson] 11473 11474 *) X509_name_add_entry() freed the wrong thing after an error. 11475 [Arne Ansper <arne@ats.cyber.ee>] 11476 11477 *) rsa_eay.c would attempt to free a NULL context. 11478 [Arne Ansper <arne@ats.cyber.ee>] 11479 11480 *) BIO_s_socket() had a broken should_retry() on Windoze. 11481 [Arne Ansper <arne@ats.cyber.ee>] 11482 11483 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 11484 [Arne Ansper <arne@ats.cyber.ee>] 11485 11486 *) Make sure the already existing X509_STORE->depth variable is initialized 11487 in X509_STORE_new(), but document the fact that this variable is still 11488 unused in the certificate verification process. 11489 [Ralf S. Engelschall] 11490 11491 *) Fix the various library and apps files to free up pkeys obtained from 11492 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 11493 [Steve Henson] 11494 11495 *) Fix reference counting in X509_PUBKEY_get(). This makes 11496 demos/maurice/example2.c work, amongst others, probably. 11497 [Steve Henson and Ben Laurie] 11498 11499 *) First cut of a cleanup for apps/. First the `ssleay' program is now named 11500 `openssl' and second, the shortcut symlinks for the `openssl <command>' 11501 are no longer created. This way we have a single and consistent command 11502 line interface `openssl <command>', similar to `cvs <command>'. 11503 [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 11504 11505 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 11506 BIT STRING wrapper always have zero unused bits. 11507 [Steve Henson] 11508 11509 *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 11510 [Steve Henson] 11511 11512 *) Make the top-level INSTALL documentation easier to understand. 11513 [Paul Sutton] 11514 11515 *) Makefiles updated to exit if an error occurs in a sub-directory 11516 make (including if user presses ^C) [Paul Sutton] 11517 11518 *) Make Montgomery context stuff explicit in RSA data structure. 11519 [Ben Laurie] 11520 11521 *) Fix build order of pem and err to allow for generated pem.h. 11522 [Ben Laurie] 11523 11524 *) Fix renumbering bug in X509_NAME_delete_entry(). 11525 [Ben Laurie] 11526 11527 *) Enhanced the err-ins.pl script so it makes the error library number 11528 global and can add a library name. This is needed for external ASN1 and 11529 other error libraries. 11530 [Steve Henson] 11531 11532 *) Fixed sk_insert which never worked properly. 11533 [Steve Henson] 11534 11535 *) Fix ASN1 macros so they can handle indefinite length construted 11536 EXPLICIT tags. Some non standard certificates use these: they can now 11537 be read in. 11538 [Steve Henson] 11539 11540 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 11541 into a single doc/ssleay.txt bundle. This way the information is still 11542 preserved but no longer messes up this directory. Now it's new room for 11543 the new set of documenation files. 11544 [Ralf S. Engelschall] 11545 11546 *) SETs were incorrectly DER encoded. This was a major pain, because they 11547 shared code with SEQUENCEs, which aren't coded the same. This means that 11548 almost everything to do with SETs or SEQUENCEs has either changed name or 11549 number of arguments. 11550 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 11551 11552 *) Fix test data to work with the above. 11553 [Ben Laurie] 11554 11555 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 11556 was already fixed by Eric for 0.9.1 it seems. 11557 [Ben Laurie - pointed out by Ulf M��ller <ulf@fitug.de>] 11558 11559 *) Autodetect FreeBSD3. 11560 [Ben Laurie] 11561 11562 *) Fix various bugs in Configure. This affects the following platforms: 11563 nextstep 11564 ncr-scde 11565 unixware-2.0 11566 unixware-2.0-pentium 11567 sco5-cc. 11568 [Ben Laurie] 11569 11570 *) Eliminate generated files from CVS. Reorder tests to regenerate files 11571 before they are needed. 11572 [Ben Laurie] 11573 11574 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 11575 [Ben Laurie] 11576 11577 11578 Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 11579 11580 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 11581 changed SSLeay to OpenSSL in version strings. 11582 [Ralf S. Engelschall] 11583 11584 *) Some fixups to the top-level documents. 11585 [Paul Sutton] 11586 11587 *) Fixed the nasty bug where rsaref.h was not found under compile-time 11588 because the symlink to include/ was missing. 11589 [Ralf S. Engelschall] 11590 11591 *) Incorporated the popular no-RSA/DSA-only patches 11592 which allow to compile a RSA-free SSLeay. 11593 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 11594 11595 *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 11596 when "ssleay" is still not found. 11597 [Ralf S. Engelschall] 11598 11599 *) Added more platforms to Configure: Cray T3E, HPUX 11, 11600 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 11601 11602 *) Updated the README file. 11603 [Ralf S. Engelschall] 11604 11605 *) Added various .cvsignore files in the CVS repository subdirs 11606 to make a "cvs update" really silent. 11607 [Ralf S. Engelschall] 11608 11609 *) Recompiled the error-definition header files and added 11610 missing symbols to the Win32 linker tables. 11611 [Ralf S. Engelschall] 11612 11613 *) Cleaned up the top-level documents; 11614 o new files: CHANGES and LICENSE 11615 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 11616 o merged COPYRIGHT into LICENSE 11617 o removed obsolete TODO file 11618 o renamed MICROSOFT to INSTALL.W32 11619 [Ralf S. Engelschall] 11620 11621 *) Removed dummy files from the 0.9.1b source tree: 11622 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 11623 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 11624 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 11625 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 11626 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 11627 [Ralf S. Engelschall] 11628 11629 *) Added various platform portability fixes. 11630 [Mark J. Cox] 11631 11632 *) The Genesis of the OpenSSL rpject: 11633 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 11634 Young and Tim J. Hudson created while they were working for C2Net until 11635 summer 1998. 11636 [The OpenSSL Project] 11637 11638 11639 Changes between 0.9.0b and 0.9.1b [not released] 11640 11641 *) Updated a few CA certificates under certs/ 11642 [Eric A. Young] 11643 11644 *) Changed some BIGNUM api stuff. 11645 [Eric A. Young] 11646 11647 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 11648 DGUX x86, Linux Alpha, etc. 11649 [Eric A. Young] 11650 11651 *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 11652 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 11653 available). 11654 [Eric A. Young] 11655 11656 *) Add -strparse option to asn1pars program which parses nested 11657 binary structures 11658 [Dr Stephen Henson <shenson@bigfoot.com>] 11659 11660 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 11661 [Eric A. Young] 11662 11663 *) DSA fix for "ca" program. 11664 [Eric A. Young] 11665 11666 *) Added "-genkey" option to "dsaparam" program. 11667 [Eric A. Young] 11668 11669 *) Added RIPE MD160 (rmd160) message digest. 11670 [Eric A. Young] 11671 11672 *) Added -a (all) option to "ssleay version" command. 11673 [Eric A. Young] 11674 11675 *) Added PLATFORM define which is the id given to Configure. 11676 [Eric A. Young] 11677 11678 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 11679 [Eric A. Young] 11680 11681 *) Extended the ASN.1 parser routines. 11682 [Eric A. Young] 11683 11684 *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 11685 [Eric A. Young] 11686 11687 *) Added a BN_CTX to the BN library. 11688 [Eric A. Young] 11689 11690 *) Fixed the weak key values in DES library 11691 [Eric A. Young] 11692 11693 *) Changed API in EVP library for cipher aliases. 11694 [Eric A. Young] 11695 11696 *) Added support for RC2/64bit cipher. 11697 [Eric A. Young] 11698 11699 *) Converted the lhash library to the crypto/mem.c functions. 11700 [Eric A. Young] 11701 11702 *) Added more recognized ASN.1 object ids. 11703 [Eric A. Young] 11704 11705 *) Added more RSA padding checks for SSL/TLS. 11706 [Eric A. Young] 11707 11708 *) Added BIO proxy/filter functionality. 11709 [Eric A. Young] 11710 11711 *) Added extra_certs to SSL_CTX which can be used 11712 send extra CA certificates to the client in the CA cert chain sending 11713 process. It can be configured with SSL_CTX_add_extra_chain_cert(). 11714 [Eric A. Young] 11715 11716 *) Now Fortezza is denied in the authentication phase because 11717 this is key exchange mechanism is not supported by SSLeay at all. 11718 [Eric A. Young] 11719 11720 *) Additional PKCS1 checks. 11721 [Eric A. Young] 11722 11723 *) Support the string "TLSv1" for all TLS v1 ciphers. 11724 [Eric A. Young] 11725 11726 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 11727 ex_data index of the SSL context in the X509_STORE_CTX ex_data. 11728 [Eric A. Young] 11729 11730 *) Fixed a few memory leaks. 11731 [Eric A. Young] 11732 11733 *) Fixed various code and comment typos. 11734 [Eric A. Young] 11735 11736 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 11737 bytes sent in the client random. 11738 [Edward Bishop <ebishop@spyglass.com>] 11739 11740