CHANGES revision 269686
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 5269686Sjkim Changes between 1.0.1h and 1.0.1i [6 Aug 2014] 6269686Sjkim 7269686Sjkim *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the 8269686Sjkim SRP code can be overrun an internal buffer. Add sanity check that 9269686Sjkim g, A, B < N to SRP code. 10269686Sjkim 11269686Sjkim Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC 12269686Sjkim Group for discovering this issue. 13269686Sjkim (CVE-2014-3512) 14269686Sjkim [Steve Henson] 15269686Sjkim 16269686Sjkim *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate 17269686Sjkim TLS 1.0 instead of higher protocol versions when the ClientHello message 18269686Sjkim is badly fragmented. This allows a man-in-the-middle attacker to force a 19269686Sjkim downgrade to TLS 1.0 even if both the server and the client support a 20269686Sjkim higher protocol version, by modifying the client's TLS records. 21269686Sjkim 22269686Sjkim Thanks to David Benjamin and Adam Langley (Google) for discovering and 23269686Sjkim researching this issue. 24269686Sjkim (CVE-2014-3511) 25269686Sjkim [David Benjamin] 26269686Sjkim 27269686Sjkim *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject 28269686Sjkim to a denial of service attack. A malicious server can crash the client 29269686Sjkim with a null pointer dereference (read) by specifying an anonymous (EC)DH 30269686Sjkim ciphersuite and sending carefully crafted handshake messages. 31269686Sjkim 32269686Sjkim Thanks to Felix Gr�bert (Google) for discovering and researching this 33269686Sjkim issue. 34269686Sjkim (CVE-2014-3510) 35269686Sjkim [Emilia K�sper] 36269686Sjkim 37269686Sjkim *) By sending carefully crafted DTLS packets an attacker could cause openssl 38269686Sjkim to leak memory. This can be exploited through a Denial of Service attack. 39269686Sjkim Thanks to Adam Langley for discovering and researching this issue. 40269686Sjkim (CVE-2014-3507) 41269686Sjkim [Adam Langley] 42269686Sjkim 43269686Sjkim *) An attacker can force openssl to consume large amounts of memory whilst 44269686Sjkim processing DTLS handshake messages. This can be exploited through a 45269686Sjkim Denial of Service attack. 46269686Sjkim Thanks to Adam Langley for discovering and researching this issue. 47269686Sjkim (CVE-2014-3506) 48269686Sjkim [Adam Langley] 49269686Sjkim 50269686Sjkim *) An attacker can force an error condition which causes openssl to crash 51269686Sjkim whilst processing DTLS packets due to memory being freed twice. This 52269686Sjkim can be exploited through a Denial of Service attack. 53269686Sjkim Thanks to Adam Langley and Wan-Teh Chang for discovering and researching 54269686Sjkim this issue. 55269686Sjkim (CVE-2014-3505) 56269686Sjkim [Adam Langley] 57269686Sjkim 58269686Sjkim *) If a multithreaded client connects to a malicious server using a resumed 59269686Sjkim session and the server sends an ec point format extension it could write 60269686Sjkim up to 255 bytes to freed memory. 61269686Sjkim 62269686Sjkim Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this 63269686Sjkim issue. 64269686Sjkim (CVE-2014-3509) 65269686Sjkim [Gabor Tyukasz] 66269686Sjkim 67269686Sjkim *) A malicious server can crash an OpenSSL client with a null pointer 68269686Sjkim dereference (read) by specifying an SRP ciphersuite even though it was not 69269686Sjkim properly negotiated with the client. This can be exploited through a 70269686Sjkim Denial of Service attack. 71269686Sjkim 72269686Sjkim Thanks to Joonas Kuorilehto and Riku Hietam�ki (Codenomicon) for 73269686Sjkim discovering and researching this issue. 74269686Sjkim (CVE-2014-5139) 75269686Sjkim [Steve Henson] 76269686Sjkim 77269686Sjkim *) A flaw in OBJ_obj2txt may cause pretty printing functions such as 78269686Sjkim X509_name_oneline, X509_name_print_ex et al. to leak some information 79269686Sjkim from the stack. Applications may be affected if they echo pretty printing 80269686Sjkim output to the attacker. 81269686Sjkim 82269686Sjkim Thanks to Ivan Fratric (Google) for discovering this issue. 83269686Sjkim (CVE-2014-3508) 84269686Sjkim [Emilia K�sper, and Steve Henson] 85269686Sjkim 86269686Sjkim *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) 87269686Sjkim for corner cases. (Certain input points at infinity could lead to 88269686Sjkim bogus results, with non-infinity inputs mapped to infinity too.) 89269686Sjkim [Bodo Moeller] 90269686Sjkim 91267258Sjkim Changes between 1.0.1g and 1.0.1h [5 Jun 2014] 92267258Sjkim 93267258Sjkim *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted 94267258Sjkim handshake can force the use of weak keying material in OpenSSL 95267258Sjkim SSL/TLS clients and servers. 96267258Sjkim 97267258Sjkim Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and 98267258Sjkim researching this issue. (CVE-2014-0224) 99267258Sjkim [KIKUCHI Masashi, Steve Henson] 100267258Sjkim 101267258Sjkim *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an 102267258Sjkim OpenSSL DTLS client the code can be made to recurse eventually crashing 103267258Sjkim in a DoS attack. 104267258Sjkim 105267258Sjkim Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue. 106267258Sjkim (CVE-2014-0221) 107267258Sjkim [Imre Rad, Steve Henson] 108267258Sjkim 109267258Sjkim *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can 110267258Sjkim be triggered by sending invalid DTLS fragments to an OpenSSL DTLS 111267258Sjkim client or server. This is potentially exploitable to run arbitrary 112267258Sjkim code on a vulnerable client or server. 113267258Sjkim 114267258Sjkim Thanks to J�ri Aedla for reporting this issue. (CVE-2014-0195) 115267258Sjkim [J�ri Aedla, Steve Henson] 116267258Sjkim 117267258Sjkim *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites 118267258Sjkim are subject to a denial of service attack. 119267258Sjkim 120267258Sjkim Thanks to Felix Gr�bert and Ivan Fratric at Google for discovering 121267258Sjkim this issue. (CVE-2014-3470) 122267258Sjkim [Felix Gr�bert, Ivan Fratric, Steve Henson] 123267258Sjkim 124267258Sjkim *) Harmonize version and its documentation. -f flag is used to display 125267258Sjkim compilation flags. 126267258Sjkim [mancha <mancha1@zoho.com>] 127267258Sjkim 128267258Sjkim *) Fix eckey_priv_encode so it immediately returns an error upon a failure 129267258Sjkim in i2d_ECPrivateKey. 130267258Sjkim [mancha <mancha1@zoho.com>] 131267258Sjkim 132267258Sjkim *) Fix some double frees. These are not thought to be exploitable. 133267258Sjkim [mancha <mancha1@zoho.com>] 134267258Sjkim 135264331Sjkim Changes between 1.0.1f and 1.0.1g [7 Apr 2014] 136264331Sjkim 137264331Sjkim *) A missing bounds check in the handling of the TLS heartbeat extension 138264331Sjkim can be used to reveal up to 64k of memory to a connected client or 139264331Sjkim server. 140264331Sjkim 141264331Sjkim Thanks for Neel Mehta of Google Security for discovering this bug and to 142264331Sjkim Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for 143264331Sjkim preparing the fix (CVE-2014-0160) 144264331Sjkim [Adam Langley, Bodo Moeller] 145264331Sjkim 146264331Sjkim *) Fix for the attack described in the paper "Recovering OpenSSL 147264331Sjkim ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack" 148264331Sjkim by Yuval Yarom and Naomi Benger. Details can be obtained from: 149264331Sjkim http://eprint.iacr.org/2014/140 150264331Sjkim 151264331Sjkim Thanks to Yuval Yarom and Naomi Benger for discovering this 152264331Sjkim flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076) 153264331Sjkim [Yuval Yarom and Naomi Benger] 154264331Sjkim 155264331Sjkim *) TLS pad extension: draft-agl-tls-padding-03 156264331Sjkim 157264331Sjkim Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the 158264331Sjkim TLS client Hello record length value would otherwise be > 255 and 159264331Sjkim less that 512 pad with a dummy extension containing zeroes so it 160264331Sjkim is at least 512 bytes long. 161264331Sjkim 162264331Sjkim [Adam Langley, Steve Henson] 163264331Sjkim 164264331Sjkim Changes between 1.0.1e and 1.0.1f [6 Jan 2014] 165264331Sjkim 166264331Sjkim *) Fix for TLS record tampering bug. A carefully crafted invalid 167264331Sjkim handshake could crash OpenSSL with a NULL pointer exception. 168264331Sjkim Thanks to Anton Johansson for reporting this issues. 169264331Sjkim (CVE-2013-4353) 170264331Sjkim 171264331Sjkim *) Keep original DTLS digest and encryption contexts in retransmission 172264331Sjkim structures so we can use the previous session parameters if they need 173264331Sjkim to be resent. (CVE-2013-6450) 174264331Sjkim [Steve Henson] 175264331Sjkim 176264331Sjkim *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which 177264331Sjkim avoids preferring ECDHE-ECDSA ciphers when the client appears to be 178264331Sjkim Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for 179264331Sjkim several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug 180264331Sjkim is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing 181264331Sjkim 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer. 182264331Sjkim [Rob Stradling, Adam Langley] 183264331Sjkim 184246772Sjkim Changes between 1.0.1d and 1.0.1e [11 Feb 2013] 185246772Sjkim 186264331Sjkim *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI 187264331Sjkim supporting platforms or when small records were transferred. 188264331Sjkim [Andy Polyakov, Steve Henson] 189246772Sjkim 190246772Sjkim Changes between 1.0.1c and 1.0.1d [5 Feb 2013] 191246772Sjkim 192246772Sjkim *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 193246772Sjkim 194246772Sjkim This addresses the flaw in CBC record processing discovered by 195246772Sjkim Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 196246772Sjkim at: http://www.isg.rhul.ac.uk/tls/ 197246772Sjkim 198246772Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 199246772Sjkim Security Group at Royal Holloway, University of London 200246772Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 201246772Sjkim Emilia K�sper for the initial patch. 202246772Sjkim (CVE-2013-0169) 203246772Sjkim [Emilia K�sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 204246772Sjkim 205246772Sjkim *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode 206246772Sjkim ciphersuites which can be exploited in a denial of service attack. 207246772Sjkim Thanks go to and to Adam Langley <agl@chromium.org> for discovering 208246772Sjkim and detecting this bug and to Wolfgang Ettlinger 209246772Sjkim <wolfgang.ettlinger@gmail.com> for independently discovering this issue. 210246772Sjkim (CVE-2012-2686) 211246772Sjkim [Adam Langley] 212246772Sjkim 213246772Sjkim *) Return an error when checking OCSP signatures when key is NULL. 214246772Sjkim This fixes a DoS attack. (CVE-2013-0166) 215246772Sjkim [Steve Henson] 216246772Sjkim 217246772Sjkim *) Make openssl verify return errors. 218246772Sjkim [Chris Palmer <palmer@google.com> and Ben Laurie] 219246772Sjkim 220246772Sjkim *) Call OCSP Stapling callback after ciphersuite has been chosen, so 221246772Sjkim the right response is stapled. Also change SSL_get_certificate() 222246772Sjkim so it returns the certificate actually sent. 223246772Sjkim See http://rt.openssl.org/Ticket/Display.html?id=2836. 224246772Sjkim [Rob Stradling <rob.stradling@comodo.com>] 225246772Sjkim 226246772Sjkim *) Fix possible deadlock when decoding public keys. 227246772Sjkim [Steve Henson] 228246772Sjkim 229246772Sjkim *) Don't use TLS 1.0 record version number in initial client hello 230246772Sjkim if renegotiating. 231246772Sjkim [Steve Henson] 232246772Sjkim 233238405Sjkim Changes between 1.0.1b and 1.0.1c [10 May 2012] 234237657Sjkim 235238405Sjkim *) Sanity check record length before skipping explicit IV in TLS 236238405Sjkim 1.2, 1.1 and DTLS to fix DoS attack. 237237657Sjkim 238237657Sjkim Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 239237657Sjkim fuzzing as a service testing platform. 240237657Sjkim (CVE-2012-2333) 241237657Sjkim [Steve Henson] 242237657Sjkim 243237657Sjkim *) Initialise tkeylen properly when encrypting CMS messages. 244237657Sjkim Thanks to Solar Designer of Openwall for reporting this issue. 245237657Sjkim [Steve Henson] 246237657Sjkim 247238405Sjkim *) In FIPS mode don't try to use composite ciphers as they are not 248238405Sjkim approved. 249238405Sjkim [Steve Henson] 250237657Sjkim 251238405Sjkim Changes between 1.0.1a and 1.0.1b [26 Apr 2012] 252237657Sjkim 253238405Sjkim *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and 254238405Sjkim 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately 255238405Sjkim mean any application compiled against OpenSSL 1.0.0 headers setting 256238405Sjkim SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng 257238405Sjkim TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to 258238405Sjkim 0x10000000L Any application which was previously compiled against 259238405Sjkim OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1 260238405Sjkim will need to be recompiled as a result. Letting be results in 261238405Sjkim inability to disable specifically TLS 1.1 and in client context, 262238405Sjkim in unlike event, limit maximum offered version to TLS 1.0 [see below]. 263238405Sjkim [Steve Henson] 264237657Sjkim 265238405Sjkim *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not 266238405Sjkim disable just protocol X, but all protocols above X *if* there are 267238405Sjkim protocols *below* X still enabled. In more practical terms it means 268238405Sjkim that if application wants to disable TLS1.0 in favor of TLS1.1 and 269238405Sjkim above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass 270238405Sjkim SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to 271238405Sjkim client side. 272238405Sjkim [Andy Polyakov] 273238405Sjkim 274238405Sjkim Changes between 1.0.1 and 1.0.1a [19 Apr 2012] 275238405Sjkim 276237657Sjkim *) Check for potentially exploitable overflows in asn1_d2i_read_bio 277237657Sjkim BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 278237657Sjkim in CRYPTO_realloc_clean. 279237657Sjkim 280237657Sjkim Thanks to Tavis Ormandy, Google Security Team, for discovering this 281237657Sjkim issue and to Adam Langley <agl@chromium.org> for fixing it. 282237657Sjkim (CVE-2012-2110) 283237657Sjkim [Adam Langley (Google), Tavis Ormandy, Google Security Team] 284237657Sjkim 285238405Sjkim *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections. 286238405Sjkim [Adam Langley] 287237657Sjkim 288238405Sjkim *) Workarounds for some broken servers that "hang" if a client hello 289238405Sjkim record length exceeds 255 bytes. 290238405Sjkim 291238405Sjkim 1. Do not use record version number > TLS 1.0 in initial client 292238405Sjkim hello: some (but not all) hanging servers will now work. 293238405Sjkim 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate 294238405Sjkim the number of ciphers sent in the client hello. This should be 295238405Sjkim set to an even number, such as 50, for example by passing: 296238405Sjkim -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure. 297238405Sjkim Most broken servers should now work. 298238405Sjkim 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable 299238405Sjkim TLS 1.2 client support entirely. 300238405Sjkim [Steve Henson] 301238405Sjkim 302238405Sjkim *) Fix SEGV in Vector Permutation AES module observed in OpenSSH. 303238405Sjkim [Andy Polyakov] 304238405Sjkim 305238405Sjkim Changes between 1.0.0h and 1.0.1 [14 Mar 2012] 306238405Sjkim 307238405Sjkim *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET 308238405Sjkim STRING form instead of a DigestInfo. 309238405Sjkim [Steve Henson] 310238405Sjkim 311238405Sjkim *) The format used for MDC2 RSA signatures is inconsistent between EVP 312238405Sjkim and the RSA_sign/RSA_verify functions. This was made more apparent when 313238405Sjkim OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular 314238405Sjkim those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect 315238405Sjkim the correct format in RSA_verify so both forms transparently work. 316238405Sjkim [Steve Henson] 317238405Sjkim 318238405Sjkim *) Some servers which support TLS 1.0 can choke if we initially indicate 319238405Sjkim support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA 320238405Sjkim encrypted premaster secret. As a workaround use the maximum pemitted 321238405Sjkim client version in client hello, this should keep such servers happy 322238405Sjkim and still work with previous versions of OpenSSL. 323238405Sjkim [Steve Henson] 324238405Sjkim 325238405Sjkim *) Add support for TLS/DTLS heartbeats. 326238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 327238405Sjkim 328238405Sjkim *) Add support for SCTP. 329238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 330238405Sjkim 331238405Sjkim *) Improved PRNG seeding for VOS. 332238405Sjkim [Paul Green <Paul.Green@stratus.com>] 333238405Sjkim 334238405Sjkim *) Extensive assembler packs updates, most notably: 335238405Sjkim 336238405Sjkim - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support; 337238405Sjkim - x86[_64]: SSSE3 support (SHA1, vector-permutation AES); 338238405Sjkim - x86_64: bit-sliced AES implementation; 339238405Sjkim - ARM: NEON support, contemporary platforms optimizations; 340238405Sjkim - s390x: z196 support; 341238405Sjkim - *: GHASH and GF(2^m) multiplication implementations; 342238405Sjkim 343238405Sjkim [Andy Polyakov] 344238405Sjkim 345238405Sjkim *) Make TLS-SRP code conformant with RFC 5054 API cleanup 346238405Sjkim (removal of unnecessary code) 347238405Sjkim [Peter Sylvester <peter.sylvester@edelweb.fr>] 348238405Sjkim 349238405Sjkim *) Add TLS key material exporter from RFC 5705. 350238405Sjkim [Eric Rescorla] 351238405Sjkim 352238405Sjkim *) Add DTLS-SRTP negotiation from RFC 5764. 353238405Sjkim [Eric Rescorla] 354238405Sjkim 355238405Sjkim *) Add Next Protocol Negotiation, 356238405Sjkim http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be 357238405Sjkim disabled with a no-npn flag to config or Configure. Code donated 358238405Sjkim by Google. 359238405Sjkim [Adam Langley <agl@google.com> and Ben Laurie] 360238405Sjkim 361238405Sjkim *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224, 362238405Sjkim NIST-P256, NIST-P521, with constant-time single point multiplication on 363238405Sjkim typical inputs. Compiler support for the nonstandard type __uint128_t is 364238405Sjkim required to use this (present in gcc 4.4 and later, for 64-bit builds). 365238405Sjkim Code made available under Apache License version 2.0. 366238405Sjkim 367238405Sjkim Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command 368238405Sjkim line to include this in your build of OpenSSL, and run "make depend" (or 369238405Sjkim "make update"). This enables the following EC_METHODs: 370238405Sjkim 371238405Sjkim EC_GFp_nistp224_method() 372238405Sjkim EC_GFp_nistp256_method() 373238405Sjkim EC_GFp_nistp521_method() 374238405Sjkim 375238405Sjkim EC_GROUP_new_by_curve_name() will automatically use these (while 376238405Sjkim EC_GROUP_new_curve_GFp() currently prefers the more flexible 377238405Sjkim implementations). 378238405Sjkim [Emilia K�sper, Adam Langley, Bodo Moeller (Google)] 379238405Sjkim 380238405Sjkim *) Use type ossl_ssize_t instad of ssize_t which isn't available on 381238405Sjkim all platforms. Move ssize_t definition from e_os.h to the public 382238405Sjkim header file e_os2.h as it now appears in public header file cms.h 383238405Sjkim [Steve Henson] 384238405Sjkim 385238405Sjkim *) New -sigopt option to the ca, req and x509 utilities. Additional 386238405Sjkim signature parameters can be passed using this option and in 387238405Sjkim particular PSS. 388238405Sjkim [Steve Henson] 389238405Sjkim 390238405Sjkim *) Add RSA PSS signing function. This will generate and set the 391238405Sjkim appropriate AlgorithmIdentifiers for PSS based on those in the 392238405Sjkim corresponding EVP_MD_CTX structure. No application support yet. 393238405Sjkim [Steve Henson] 394238405Sjkim 395238405Sjkim *) Support for companion algorithm specific ASN1 signing routines. 396238405Sjkim New function ASN1_item_sign_ctx() signs a pre-initialised 397238405Sjkim EVP_MD_CTX structure and sets AlgorithmIdentifiers based on 398238405Sjkim the appropriate parameters. 399238405Sjkim [Steve Henson] 400238405Sjkim 401238405Sjkim *) Add new algorithm specific ASN1 verification initialisation function 402238405Sjkim to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1 403238405Sjkim handling will be the same no matter what EVP_PKEY_METHOD is used. 404238405Sjkim Add a PSS handler to support verification of PSS signatures: checked 405238405Sjkim against a number of sample certificates. 406238405Sjkim [Steve Henson] 407238405Sjkim 408238405Sjkim *) Add signature printing for PSS. Add PSS OIDs. 409238405Sjkim [Steve Henson, Martin Kaiser <lists@kaiser.cx>] 410238405Sjkim 411238405Sjkim *) Add algorithm specific signature printing. An individual ASN1 method 412238405Sjkim can now print out signatures instead of the standard hex dump. 413238405Sjkim 414238405Sjkim More complex signatures (e.g. PSS) can print out more meaningful 415238405Sjkim information. Include DSA version that prints out the signature 416238405Sjkim parameters r, s. 417238405Sjkim [Steve Henson] 418238405Sjkim 419238405Sjkim *) Password based recipient info support for CMS library: implementing 420238405Sjkim RFC3211. 421238405Sjkim [Steve Henson] 422238405Sjkim 423238405Sjkim *) Split password based encryption into PBES2 and PBKDF2 functions. This 424238405Sjkim neatly separates the code into cipher and PBE sections and is required 425238405Sjkim for some algorithms that split PBES2 into separate pieces (such as 426238405Sjkim password based CMS). 427238405Sjkim [Steve Henson] 428238405Sjkim 429238405Sjkim *) Session-handling fixes: 430238405Sjkim - Fix handling of connections that are resuming with a session ID, 431238405Sjkim but also support Session Tickets. 432238405Sjkim - Fix a bug that suppressed issuing of a new ticket if the client 433238405Sjkim presented a ticket with an expired session. 434238405Sjkim - Try to set the ticket lifetime hint to something reasonable. 435238405Sjkim - Make tickets shorter by excluding irrelevant information. 436238405Sjkim - On the client side, don't ignore renewed tickets. 437238405Sjkim [Adam Langley, Bodo Moeller (Google)] 438238405Sjkim 439238405Sjkim *) Fix PSK session representation. 440238405Sjkim [Bodo Moeller] 441238405Sjkim 442238405Sjkim *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations. 443238405Sjkim 444238405Sjkim This work was sponsored by Intel. 445238405Sjkim [Andy Polyakov] 446238405Sjkim 447238405Sjkim *) Add GCM support to TLS library. Some custom code is needed to split 448238405Sjkim the IV between the fixed (from PRF) and explicit (from TLS record) 449238405Sjkim portions. This adds all GCM ciphersuites supported by RFC5288 and 450238405Sjkim RFC5289. Generalise some AES* cipherstrings to inlclude GCM and 451238405Sjkim add a special AESGCM string for GCM only. 452238405Sjkim [Steve Henson] 453238405Sjkim 454238405Sjkim *) Expand range of ctrls for AES GCM. Permit setting invocation 455238405Sjkim field on decrypt and retrieval of invocation field only on encrypt. 456238405Sjkim [Steve Henson] 457238405Sjkim 458238405Sjkim *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support. 459238405Sjkim As required by RFC5289 these ciphersuites cannot be used if for 460238405Sjkim versions of TLS earlier than 1.2. 461238405Sjkim [Steve Henson] 462238405Sjkim 463238405Sjkim *) For FIPS capable OpenSSL interpret a NULL default public key method 464238405Sjkim as unset and return the appopriate default but do *not* set the default. 465238405Sjkim This means we can return the appopriate method in applications that 466238405Sjkim swicth between FIPS and non-FIPS modes. 467238405Sjkim [Steve Henson] 468238405Sjkim 469238405Sjkim *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an 470238405Sjkim ENGINE is used then we cannot handle that in the FIPS module so we 471238405Sjkim keep original code iff non-FIPS operations are allowed. 472238405Sjkim [Steve Henson] 473238405Sjkim 474238405Sjkim *) Add -attime option to openssl utilities. 475238405Sjkim [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson] 476238405Sjkim 477238405Sjkim *) Redirect DSA and DH operations to FIPS module in FIPS mode. 478238405Sjkim [Steve Henson] 479238405Sjkim 480238405Sjkim *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use 481238405Sjkim FIPS EC methods unconditionally for now. 482238405Sjkim [Steve Henson] 483238405Sjkim 484238405Sjkim *) New build option no-ec2m to disable characteristic 2 code. 485238405Sjkim [Steve Henson] 486238405Sjkim 487238405Sjkim *) Backport libcrypto audit of return value checking from 1.1.0-dev; not 488238405Sjkim all cases can be covered as some introduce binary incompatibilities. 489238405Sjkim [Steve Henson] 490238405Sjkim 491238405Sjkim *) Redirect RSA operations to FIPS module including keygen, 492238405Sjkim encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods. 493238405Sjkim [Steve Henson] 494238405Sjkim 495238405Sjkim *) Add similar low level API blocking to ciphers. 496238405Sjkim [Steve Henson] 497238405Sjkim 498238405Sjkim *) Low level digest APIs are not approved in FIPS mode: any attempt 499238405Sjkim to use these will cause a fatal error. Applications that *really* want 500238405Sjkim to use them can use the private_* version instead. 501238405Sjkim [Steve Henson] 502238405Sjkim 503238405Sjkim *) Redirect cipher operations to FIPS module for FIPS builds. 504238405Sjkim [Steve Henson] 505238405Sjkim 506238405Sjkim *) Redirect digest operations to FIPS module for FIPS builds. 507238405Sjkim [Steve Henson] 508238405Sjkim 509238405Sjkim *) Update build system to add "fips" flag which will link in fipscanister.o 510238405Sjkim for static and shared library builds embedding a signature if needed. 511238405Sjkim [Steve Henson] 512238405Sjkim 513238405Sjkim *) Output TLS supported curves in preference order instead of numerical 514238405Sjkim order. This is currently hardcoded for the highest order curves first. 515238405Sjkim This should be configurable so applications can judge speed vs strength. 516238405Sjkim [Steve Henson] 517238405Sjkim 518238405Sjkim *) Add TLS v1.2 server support for client authentication. 519238405Sjkim [Steve Henson] 520238405Sjkim 521238405Sjkim *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers 522238405Sjkim and enable MD5. 523238405Sjkim [Steve Henson] 524238405Sjkim 525238405Sjkim *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying 526238405Sjkim FIPS modules versions. 527238405Sjkim [Steve Henson] 528238405Sjkim 529238405Sjkim *) Add TLS v1.2 client side support for client authentication. Keep cache 530238405Sjkim of handshake records longer as we don't know the hash algorithm to use 531238405Sjkim until after the certificate request message is received. 532238405Sjkim [Steve Henson] 533238405Sjkim 534238405Sjkim *) Initial TLS v1.2 client support. Add a default signature algorithms 535238405Sjkim extension including all the algorithms we support. Parse new signature 536238405Sjkim format in client key exchange. Relax some ECC signing restrictions for 537238405Sjkim TLS v1.2 as indicated in RFC5246. 538238405Sjkim [Steve Henson] 539238405Sjkim 540238405Sjkim *) Add server support for TLS v1.2 signature algorithms extension. Switch 541238405Sjkim to new signature format when needed using client digest preference. 542238405Sjkim All server ciphersuites should now work correctly in TLS v1.2. No client 543238405Sjkim support yet and no support for client certificates. 544238405Sjkim [Steve Henson] 545238405Sjkim 546238405Sjkim *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch 547238405Sjkim to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based 548238405Sjkim ciphersuites. At present only RSA key exchange ciphersuites work with 549238405Sjkim TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete 550238405Sjkim SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods 551238405Sjkim and version checking. 552238405Sjkim [Steve Henson] 553238405Sjkim 554238405Sjkim *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled 555238405Sjkim with this defined it will not be affected by any changes to ssl internal 556238405Sjkim structures. Add several utility functions to allow openssl application 557238405Sjkim to work with OPENSSL_NO_SSL_INTERN defined. 558238405Sjkim [Steve Henson] 559238405Sjkim 560238405Sjkim *) Add SRP support. 561238405Sjkim [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie] 562238405Sjkim 563238405Sjkim *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id. 564238405Sjkim [Steve Henson] 565238405Sjkim 566238405Sjkim *) Permit abbreviated handshakes when renegotiating using the function 567238405Sjkim SSL_renegotiate_abbreviated(). 568238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 569238405Sjkim 570238405Sjkim *) Add call to ENGINE_register_all_complete() to 571238405Sjkim ENGINE_load_builtin_engines(), so some implementations get used 572238405Sjkim automatically instead of needing explicit application support. 573238405Sjkim [Steve Henson] 574238405Sjkim 575238405Sjkim *) Add support for TLS key exporter as described in RFC5705. 576238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson] 577238405Sjkim 578238405Sjkim *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only 579238405Sjkim a few changes are required: 580238405Sjkim 581238405Sjkim Add SSL_OP_NO_TLSv1_1 flag. 582238405Sjkim Add TLSv1_1 methods. 583238405Sjkim Update version checking logic to handle version 1.1. 584238405Sjkim Add explicit IV handling (ported from DTLS code). 585238405Sjkim Add command line options to s_client/s_server. 586238405Sjkim [Steve Henson] 587238405Sjkim 588264331Sjkim Changes between 1.0.0j and 1.0.0k [5 Feb 2013] 589264331Sjkim 590264331Sjkim *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 591264331Sjkim 592264331Sjkim This addresses the flaw in CBC record processing discovered by 593264331Sjkim Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 594264331Sjkim at: http://www.isg.rhul.ac.uk/tls/ 595264331Sjkim 596264331Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 597264331Sjkim Security Group at Royal Holloway, University of London 598264331Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 599264331Sjkim Emilia K�sper for the initial patch. 600264331Sjkim (CVE-2013-0169) 601264331Sjkim [Emilia K�sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 602264331Sjkim 603264331Sjkim *) Return an error when checking OCSP signatures when key is NULL. 604264331Sjkim This fixes a DoS attack. (CVE-2013-0166) 605264331Sjkim [Steve Henson] 606264331Sjkim 607264331Sjkim *) Call OCSP Stapling callback after ciphersuite has been chosen, so 608264331Sjkim the right response is stapled. Also change SSL_get_certificate() 609264331Sjkim so it returns the certificate actually sent. 610264331Sjkim See http://rt.openssl.org/Ticket/Display.html?id=2836. 611264331Sjkim (This is a backport) 612264331Sjkim [Rob Stradling <rob.stradling@comodo.com>] 613264331Sjkim 614264331Sjkim *) Fix possible deadlock when decoding public keys. 615264331Sjkim [Steve Henson] 616264331Sjkim 617264331Sjkim Changes between 1.0.0i and 1.0.0j [10 May 2012] 618264331Sjkim 619264331Sjkim [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after 620264331Sjkim OpenSSL 1.0.1.] 621264331Sjkim 622264331Sjkim *) Sanity check record length before skipping explicit IV in DTLS 623264331Sjkim to fix DoS attack. 624264331Sjkim 625264331Sjkim Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 626264331Sjkim fuzzing as a service testing platform. 627264331Sjkim (CVE-2012-2333) 628264331Sjkim [Steve Henson] 629264331Sjkim 630264331Sjkim *) Initialise tkeylen properly when encrypting CMS messages. 631264331Sjkim Thanks to Solar Designer of Openwall for reporting this issue. 632264331Sjkim [Steve Henson] 633264331Sjkim 634264331Sjkim Changes between 1.0.0h and 1.0.0i [19 Apr 2012] 635264331Sjkim 636264331Sjkim *) Check for potentially exploitable overflows in asn1_d2i_read_bio 637264331Sjkim BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 638264331Sjkim in CRYPTO_realloc_clean. 639264331Sjkim 640264331Sjkim Thanks to Tavis Ormandy, Google Security Team, for discovering this 641264331Sjkim issue and to Adam Langley <agl@chromium.org> for fixing it. 642264331Sjkim (CVE-2012-2110) 643264331Sjkim [Adam Langley (Google), Tavis Ormandy, Google Security Team] 644264331Sjkim 645238405Sjkim Changes between 1.0.0g and 1.0.0h [12 Mar 2012] 646238405Sjkim 647237657Sjkim *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 648237657Sjkim in CMS and PKCS7 code. When RSA decryption fails use a random key for 649237657Sjkim content decryption and always return the same error. Note: this attack 650237657Sjkim needs on average 2^20 messages so it only affects automated senders. The 651237657Sjkim old behaviour can be reenabled in the CMS code by setting the 652237657Sjkim CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 653237657Sjkim an MMA defence is not necessary. 654237657Sjkim Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 655237657Sjkim this issue. (CVE-2012-0884) 656237657Sjkim [Steve Henson] 657237657Sjkim 658237657Sjkim *) Fix CVE-2011-4619: make sure we really are receiving a 659237657Sjkim client hello before rejecting multiple SGC restarts. Thanks to 660237657Sjkim Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 661237657Sjkim [Steve Henson] 662237657Sjkim 663238405Sjkim Changes between 1.0.0f and 1.0.0g [18 Jan 2012] 664238405Sjkim 665238405Sjkim *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 666238405Sjkim Thanks to Antonio Martin, Enterprise Secure Access Research and 667238405Sjkim Development, Cisco Systems, Inc. for discovering this bug and 668238405Sjkim preparing a fix. (CVE-2012-0050) 669238405Sjkim [Antonio Martin] 670238405Sjkim 671238405Sjkim Changes between 1.0.0e and 1.0.0f [4 Jan 2012] 672238405Sjkim 673238405Sjkim *) Nadhem Alfardan and Kenny Paterson have discovered an extension 674238405Sjkim of the Vaudenay padding oracle attack on CBC mode encryption 675238405Sjkim which enables an efficient plaintext recovery attack against 676238405Sjkim the OpenSSL implementation of DTLS. Their attack exploits timing 677238405Sjkim differences arising during decryption processing. A research 678238405Sjkim paper describing this attack can be found at: 679238405Sjkim http://www.isg.rhul.ac.uk/~kp/dtls.pdf 680238405Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 681238405Sjkim Security Group at Royal Holloway, University of London 682238405Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 683238405Sjkim <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 684238405Sjkim for preparing the fix. (CVE-2011-4108) 685238405Sjkim [Robin Seggelmann, Michael Tuexen] 686238405Sjkim 687238405Sjkim *) Clear bytes used for block padding of SSL 3.0 records. 688238405Sjkim (CVE-2011-4576) 689238405Sjkim [Adam Langley (Google)] 690238405Sjkim 691238405Sjkim *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 692238405Sjkim Kadianakis <desnacked@gmail.com> for discovering this issue and 693238405Sjkim Adam Langley for preparing the fix. (CVE-2011-4619) 694238405Sjkim [Adam Langley (Google)] 695238405Sjkim 696238405Sjkim *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027) 697238405Sjkim [Andrey Kulikov <amdeich@gmail.com>] 698238405Sjkim 699238405Sjkim *) Prevent malformed RFC3779 data triggering an assertion failure. 700238405Sjkim Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 701238405Sjkim and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 702238405Sjkim [Rob Austein <sra@hactrn.net>] 703238405Sjkim 704238405Sjkim *) Improved PRNG seeding for VOS. 705238405Sjkim [Paul Green <Paul.Green@stratus.com>] 706238405Sjkim 707238405Sjkim *) Fix ssl_ciph.c set-up race. 708238405Sjkim [Adam Langley (Google)] 709238405Sjkim 710238405Sjkim *) Fix spurious failures in ecdsatest.c. 711238405Sjkim [Emilia K�sper (Google)] 712238405Sjkim 713238405Sjkim *) Fix the BIO_f_buffer() implementation (which was mixing different 714238405Sjkim interpretations of the '..._len' fields). 715238405Sjkim [Adam Langley (Google)] 716238405Sjkim 717238405Sjkim *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 718238405Sjkim BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 719238405Sjkim threads won't reuse the same blinding coefficients. 720238405Sjkim 721238405Sjkim This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 722238405Sjkim lock to call BN_BLINDING_invert_ex, and avoids one use of 723238405Sjkim BN_BLINDING_update for each BN_BLINDING structure (previously, 724238405Sjkim the last update always remained unused). 725238405Sjkim [Emilia K�sper (Google)] 726238405Sjkim 727238405Sjkim *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf. 728238405Sjkim [Bob Buckholz (Google)] 729238405Sjkim 730238405Sjkim Changes between 1.0.0d and 1.0.0e [6 Sep 2011] 731238405Sjkim 732238405Sjkim *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted 733238405Sjkim by initialising X509_STORE_CTX properly. (CVE-2011-3207) 734238405Sjkim [Kaspar Brand <ossl@velox.ch>] 735238405Sjkim 736238405Sjkim *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 737238405Sjkim for multi-threaded use of ECDH. (CVE-2011-3210) 738238405Sjkim [Adam Langley (Google)] 739238405Sjkim 740238405Sjkim *) Fix x509_name_ex_d2i memory leak on bad inputs. 741238405Sjkim [Bodo Moeller] 742238405Sjkim 743238405Sjkim *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check 744238405Sjkim signature public key algorithm by using OID xref utilities instead. 745238405Sjkim Before this you could only use some ECC ciphersuites with SHA1 only. 746238405Sjkim [Steve Henson] 747238405Sjkim 748238405Sjkim *) Add protection against ECDSA timing attacks as mentioned in the paper 749238405Sjkim by Billy Bob Brumley and Nicola Tuveri, see: 750238405Sjkim 751238405Sjkim http://eprint.iacr.org/2011/232.pdf 752238405Sjkim 753238405Sjkim [Billy Bob Brumley and Nicola Tuveri] 754238405Sjkim 755238405Sjkim Changes between 1.0.0c and 1.0.0d [8 Feb 2011] 756238405Sjkim 757238405Sjkim *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 758238405Sjkim [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 759238405Sjkim 760238405Sjkim *) Fix bug in string printing code: if *any* escaping is enabled we must 761238405Sjkim escape the escape character (backslash) or the resulting string is 762238405Sjkim ambiguous. 763238405Sjkim [Steve Henson] 764238405Sjkim 765238405Sjkim Changes between 1.0.0b and 1.0.0c [2 Dec 2010] 766238405Sjkim 767238405Sjkim *) Disable code workaround for ancient and obsolete Netscape browsers 768238405Sjkim and servers: an attacker can use it in a ciphersuite downgrade attack. 769238405Sjkim Thanks to Martin Rex for discovering this bug. CVE-2010-4180 770238405Sjkim [Steve Henson] 771238405Sjkim 772238405Sjkim *) Fixed J-PAKE implementation error, originally discovered by 773238405Sjkim Sebastien Martini, further info and confirmation from Stefan 774238405Sjkim Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 775238405Sjkim [Ben Laurie] 776238405Sjkim 777238405Sjkim Changes between 1.0.0a and 1.0.0b [16 Nov 2010] 778238405Sjkim 779238405Sjkim *) Fix extension code to avoid race conditions which can result in a buffer 780238405Sjkim overrun vulnerability: resumed sessions must not be modified as they can 781238405Sjkim be shared by multiple threads. CVE-2010-3864 782238405Sjkim [Steve Henson] 783238405Sjkim 784238405Sjkim *) Fix WIN32 build system to correctly link an ENGINE directory into 785238405Sjkim a DLL. 786238405Sjkim [Steve Henson] 787238405Sjkim 788238405Sjkim Changes between 1.0.0 and 1.0.0a [01 Jun 2010] 789238405Sjkim 790238405Sjkim *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover 791238405Sjkim (CVE-2010-1633) 792238405Sjkim [Steve Henson, Peter-Michael Hager <hager@dortmund.net>] 793238405Sjkim 794238405Sjkim Changes between 0.9.8n and 1.0.0 [29 Mar 2010] 795238405Sjkim 796238405Sjkim *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher 797238405Sjkim context. The operation can be customised via the ctrl mechanism in 798238405Sjkim case ENGINEs want to include additional functionality. 799238405Sjkim [Steve Henson] 800238405Sjkim 801238405Sjkim *) Tolerate yet another broken PKCS#8 key format: private key value negative. 802238405Sjkim [Steve Henson] 803238405Sjkim 804238405Sjkim *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to 805238405Sjkim output hashes compatible with older versions of OpenSSL. 806238405Sjkim [Willy Weisz <weisz@vcpc.univie.ac.at>] 807238405Sjkim 808238405Sjkim *) Fix compression algorithm handling: if resuming a session use the 809238405Sjkim compression algorithm of the resumed session instead of determining 810238405Sjkim it from client hello again. Don't allow server to change algorithm. 811238405Sjkim [Steve Henson] 812238405Sjkim 813238405Sjkim *) Add load_crls() function to apps tidying load_certs() too. Add option 814238405Sjkim to verify utility to allow additional CRLs to be included. 815238405Sjkim [Steve Henson] 816238405Sjkim 817238405Sjkim *) Update OCSP request code to permit adding custom headers to the request: 818238405Sjkim some responders need this. 819238405Sjkim [Steve Henson] 820238405Sjkim 821238405Sjkim *) The function EVP_PKEY_sign() returns <=0 on error: check return code 822238405Sjkim correctly. 823238405Sjkim [Julia Lawall <julia@diku.dk>] 824238405Sjkim 825238405Sjkim *) Update verify callback code in apps/s_cb.c and apps/verify.c, it 826238405Sjkim needlessly dereferenced structures, used obsolete functions and 827238405Sjkim didn't handle all updated verify codes correctly. 828238405Sjkim [Steve Henson] 829238405Sjkim 830238405Sjkim *) Disable MD2 in the default configuration. 831238405Sjkim [Steve Henson] 832238405Sjkim 833238405Sjkim *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to 834238405Sjkim indicate the initial BIO being pushed or popped. This makes it possible 835238405Sjkim to determine whether the BIO is the one explicitly called or as a result 836238405Sjkim of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so 837238405Sjkim it handles reference counts correctly and doesn't zero out the I/O bio 838238405Sjkim when it is not being explicitly popped. WARNING: applications which 839238405Sjkim included workarounds for the old buggy behaviour will need to be modified 840238405Sjkim or they could free up already freed BIOs. 841238405Sjkim [Steve Henson] 842238405Sjkim 843238405Sjkim *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni 844238405Sjkim renaming to all platforms (within the 0.9.8 branch, this was 845238405Sjkim done conditionally on Netware platforms to avoid a name clash). 846238405Sjkim [Guenter <lists@gknw.net>] 847238405Sjkim 848238405Sjkim *) Add ECDHE and PSK support to DTLS. 849238405Sjkim [Michael Tuexen <tuexen@fh-muenster.de>] 850238405Sjkim 851238405Sjkim *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't 852238405Sjkim be used on C++. 853238405Sjkim [Steve Henson] 854238405Sjkim 855238405Sjkim *) Add "missing" function EVP_MD_flags() (without this the only way to 856238405Sjkim retrieve a digest flags is by accessing the structure directly. Update 857238405Sjkim EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest 858238405Sjkim or cipher is registered as in the "from" argument. Print out all 859238405Sjkim registered digests in the dgst usage message instead of manually 860238405Sjkim attempting to work them out. 861238405Sjkim [Steve Henson] 862238405Sjkim 863238405Sjkim *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello: 864238405Sjkim this allows the use of compression and extensions. Change default cipher 865238405Sjkim string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2 866238405Sjkim by default unless an application cipher string requests it. 867238405Sjkim [Steve Henson] 868238405Sjkim 869238405Sjkim *) Alter match criteria in PKCS12_parse(). It used to try to use local 870238405Sjkim key ids to find matching certificates and keys but some PKCS#12 files 871238405Sjkim don't follow the (somewhat unwritten) rules and this strategy fails. 872238405Sjkim Now just gather all certificates together and the first private key 873238405Sjkim then look for the first certificate that matches the key. 874238405Sjkim [Steve Henson] 875238405Sjkim 876238405Sjkim *) Support use of registered digest and cipher names for dgst and cipher 877238405Sjkim commands instead of having to add each one as a special case. So now 878238405Sjkim you can do: 879238405Sjkim 880238405Sjkim openssl sha256 foo 881238405Sjkim 882238405Sjkim as well as: 883238405Sjkim 884238405Sjkim openssl dgst -sha256 foo 885238405Sjkim 886238405Sjkim and this works for ENGINE based algorithms too. 887238405Sjkim 888238405Sjkim [Steve Henson] 889238405Sjkim 890238405Sjkim *) Update Gost ENGINE to support parameter files. 891238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 892238405Sjkim 893238405Sjkim *) Support GeneralizedTime in ca utility. 894238405Sjkim [Oliver Martin <oliver@volatilevoid.net>, Steve Henson] 895238405Sjkim 896238405Sjkim *) Enhance the hash format used for certificate directory links. The new 897238405Sjkim form uses the canonical encoding (meaning equivalent names will work 898238405Sjkim even if they aren't identical) and uses SHA1 instead of MD5. This form 899238405Sjkim is incompatible with the older format and as a result c_rehash should 900238405Sjkim be used to rebuild symbolic links. 901238405Sjkim [Steve Henson] 902238405Sjkim 903238405Sjkim *) Make PKCS#8 the default write format for private keys, replacing the 904238405Sjkim traditional format. This form is standardised, more secure and doesn't 905238405Sjkim include an implicit MD5 dependency. 906238405Sjkim [Steve Henson] 907238405Sjkim 908238405Sjkim *) Add a $gcc_devteam_warn option to Configure. The idea is that any code 909238405Sjkim committed to OpenSSL should pass this lot as a minimum. 910238405Sjkim [Steve Henson] 911238405Sjkim 912238405Sjkim *) Add session ticket override functionality for use by EAP-FAST. 913238405Sjkim [Jouni Malinen <j@w1.fi>] 914238405Sjkim 915238405Sjkim *) Modify HMAC functions to return a value. Since these can be implemented 916238405Sjkim in an ENGINE errors can occur. 917238405Sjkim [Steve Henson] 918238405Sjkim 919238405Sjkim *) Type-checked OBJ_bsearch_ex. 920238405Sjkim [Ben Laurie] 921238405Sjkim 922238405Sjkim *) Type-checked OBJ_bsearch. Also some constification necessitated 923238405Sjkim by type-checking. Still to come: TXT_DB, bsearch(?), 924238405Sjkim OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING, 925238405Sjkim CONF_VALUE. 926238405Sjkim [Ben Laurie] 927238405Sjkim 928238405Sjkim *) New function OPENSSL_gmtime_adj() to add a specific number of days and 929238405Sjkim seconds to a tm structure directly, instead of going through OS 930238405Sjkim specific date routines. This avoids any issues with OS routines such 931238405Sjkim as the year 2038 bug. New *_adj() functions for ASN1 time structures 932238405Sjkim and X509_time_adj_ex() to cover the extended range. The existing 933238405Sjkim X509_time_adj() is still usable and will no longer have any date issues. 934238405Sjkim [Steve Henson] 935238405Sjkim 936238405Sjkim *) Delta CRL support. New use deltas option which will attempt to locate 937238405Sjkim and search any appropriate delta CRLs available. 938238405Sjkim 939238405Sjkim This work was sponsored by Google. 940238405Sjkim [Steve Henson] 941238405Sjkim 942238405Sjkim *) Support for CRLs partitioned by reason code. Reorganise CRL processing 943238405Sjkim code and add additional score elements. Validate alternate CRL paths 944238405Sjkim as part of the CRL checking and indicate a new error "CRL path validation 945238405Sjkim error" in this case. Applications wanting additional details can use 946238405Sjkim the verify callback and check the new "parent" field. If this is not 947238405Sjkim NULL CRL path validation is taking place. Existing applications wont 948238405Sjkim see this because it requires extended CRL support which is off by 949238405Sjkim default. 950238405Sjkim 951238405Sjkim This work was sponsored by Google. 952238405Sjkim [Steve Henson] 953238405Sjkim 954238405Sjkim *) Support for freshest CRL extension. 955238405Sjkim 956238405Sjkim This work was sponsored by Google. 957238405Sjkim [Steve Henson] 958238405Sjkim 959238405Sjkim *) Initial indirect CRL support. Currently only supported in the CRLs 960238405Sjkim passed directly and not via lookup. Process certificate issuer 961238405Sjkim CRL entry extension and lookup CRL entries by bother issuer name 962238405Sjkim and serial number. Check and process CRL issuer entry in IDP extension. 963238405Sjkim 964238405Sjkim This work was sponsored by Google. 965238405Sjkim [Steve Henson] 966238405Sjkim 967238405Sjkim *) Add support for distinct certificate and CRL paths. The CRL issuer 968238405Sjkim certificate is validated separately in this case. Only enabled if 969238405Sjkim an extended CRL support flag is set: this flag will enable additional 970238405Sjkim CRL functionality in future. 971238405Sjkim 972238405Sjkim This work was sponsored by Google. 973238405Sjkim [Steve Henson] 974238405Sjkim 975238405Sjkim *) Add support for policy mappings extension. 976238405Sjkim 977238405Sjkim This work was sponsored by Google. 978238405Sjkim [Steve Henson] 979238405Sjkim 980238405Sjkim *) Fixes to pathlength constraint, self issued certificate handling, 981238405Sjkim policy processing to align with RFC3280 and PKITS tests. 982238405Sjkim 983238405Sjkim This work was sponsored by Google. 984238405Sjkim [Steve Henson] 985238405Sjkim 986238405Sjkim *) Support for name constraints certificate extension. DN, email, DNS 987238405Sjkim and URI types are currently supported. 988238405Sjkim 989238405Sjkim This work was sponsored by Google. 990238405Sjkim [Steve Henson] 991238405Sjkim 992238405Sjkim *) To cater for systems that provide a pointer-based thread ID rather 993238405Sjkim than numeric, deprecate the current numeric thread ID mechanism and 994238405Sjkim replace it with a structure and associated callback type. This 995238405Sjkim mechanism allows a numeric "hash" to be extracted from a thread ID in 996238405Sjkim either case, and on platforms where pointers are larger than 'long', 997238405Sjkim mixing is done to help ensure the numeric 'hash' is usable even if it 998238405Sjkim can't be guaranteed unique. The default mechanism is to use "&errno" 999238405Sjkim as a pointer-based thread ID to distinguish between threads. 1000238405Sjkim 1001238405Sjkim Applications that want to provide their own thread IDs should now use 1002238405Sjkim CRYPTO_THREADID_set_callback() to register a callback that will call 1003238405Sjkim either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer(). 1004238405Sjkim 1005238405Sjkim Note that ERR_remove_state() is now deprecated, because it is tied 1006238405Sjkim to the assumption that thread IDs are numeric. ERR_remove_state(0) 1007238405Sjkim to free the current thread's error state should be replaced by 1008238405Sjkim ERR_remove_thread_state(NULL). 1009238405Sjkim 1010238405Sjkim (This new approach replaces the functions CRYPTO_set_idptr_callback(), 1011238405Sjkim CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in 1012238405Sjkim OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an 1013238405Sjkim application was previously providing a numeric thread callback that 1014238405Sjkim was inappropriate for distinguishing threads, then uniqueness might 1015238405Sjkim have been obtained with &errno that happened immediately in the 1016238405Sjkim intermediate development versions of OpenSSL; this is no longer the 1017238405Sjkim case, the numeric thread callback will now override the automatic use 1018238405Sjkim of &errno.) 1019238405Sjkim [Geoff Thorpe, with help from Bodo Moeller] 1020238405Sjkim 1021238405Sjkim *) Initial support for different CRL issuing certificates. This covers a 1022238405Sjkim simple case where the self issued certificates in the chain exist and 1023238405Sjkim the real CRL issuer is higher in the existing chain. 1024238405Sjkim 1025238405Sjkim This work was sponsored by Google. 1026238405Sjkim [Steve Henson] 1027238405Sjkim 1028238405Sjkim *) Removed effectively defunct crypto/store from the build. 1029238405Sjkim [Ben Laurie] 1030238405Sjkim 1031238405Sjkim *) Revamp of STACK to provide stronger type-checking. Still to come: 1032238405Sjkim TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE, 1033238405Sjkim ASN1_STRING, CONF_VALUE. 1034238405Sjkim [Ben Laurie] 1035238405Sjkim 1036238405Sjkim *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer 1037238405Sjkim RAM on SSL connections. This option can save about 34k per idle SSL. 1038238405Sjkim [Nick Mathewson] 1039238405Sjkim 1040238405Sjkim *) Revamp of LHASH to provide stronger type-checking. Still to come: 1041238405Sjkim STACK, TXT_DB, bsearch, qsort. 1042238405Sjkim [Ben Laurie] 1043238405Sjkim 1044238405Sjkim *) Initial support for Cryptographic Message Syntax (aka CMS) based 1045238405Sjkim on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility, 1046238405Sjkim support for data, signedData, compressedData, digestedData and 1047238405Sjkim encryptedData, envelopedData types included. Scripts to check against 1048238405Sjkim RFC4134 examples draft and interop and consistency checks of many 1049238405Sjkim content types and variants. 1050238405Sjkim [Steve Henson] 1051238405Sjkim 1052238405Sjkim *) Add options to enc utility to support use of zlib compression BIO. 1053238405Sjkim [Steve Henson] 1054238405Sjkim 1055238405Sjkim *) Extend mk1mf to support importing of options and assembly language 1056238405Sjkim files from Configure script, currently only included in VC-WIN32. 1057238405Sjkim The assembly language rules can now optionally generate the source 1058238405Sjkim files from the associated perl scripts. 1059238405Sjkim [Steve Henson] 1060238405Sjkim 1061238405Sjkim *) Implement remaining functionality needed to support GOST ciphersuites. 1062238405Sjkim Interop testing has been performed using CryptoPro implementations. 1063238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 1064238405Sjkim 1065238405Sjkim *) s390x assembler pack. 1066238405Sjkim [Andy Polyakov] 1067238405Sjkim 1068238405Sjkim *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU 1069238405Sjkim "family." 1070238405Sjkim [Andy Polyakov] 1071238405Sjkim 1072238405Sjkim *) Implement Opaque PRF Input TLS extension as specified in 1073238405Sjkim draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an 1074238405Sjkim official specification yet and no extension type assignment by 1075238405Sjkim IANA exists, this extension (for now) will have to be explicitly 1076238405Sjkim enabled when building OpenSSL by providing the extension number 1077238405Sjkim to use. For example, specify an option 1078238405Sjkim 1079238405Sjkim -DTLSEXT_TYPE_opaque_prf_input=0x9527 1080238405Sjkim 1081238405Sjkim to the "config" or "Configure" script to enable the extension, 1082238405Sjkim assuming extension number 0x9527 (which is a completely arbitrary 1083238405Sjkim and unofficial assignment based on the MD5 hash of the Internet 1084238405Sjkim Draft). Note that by doing so, you potentially lose 1085238405Sjkim interoperability with other TLS implementations since these might 1086238405Sjkim be using the same extension number for other purposes. 1087238405Sjkim 1088238405Sjkim SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the 1089238405Sjkim opaque PRF input value to use in the handshake. This will create 1090238405Sjkim an interal copy of the length-'len' string at 'src', and will 1091238405Sjkim return non-zero for success. 1092238405Sjkim 1093238405Sjkim To get more control and flexibility, provide a callback function 1094238405Sjkim by using 1095238405Sjkim 1096238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb) 1097238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg) 1098238405Sjkim 1099238405Sjkim where 1100238405Sjkim 1101238405Sjkim int (*cb)(SSL *, void *peerinput, size_t len, void *arg); 1102238405Sjkim void *arg; 1103238405Sjkim 1104238405Sjkim Callback function 'cb' will be called in handshakes, and is 1105238405Sjkim expected to use SSL_set_tlsext_opaque_prf_input() as appropriate. 1106238405Sjkim Argument 'arg' is for application purposes (the value as given to 1107238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly 1108238405Sjkim be provided to the callback function). The callback function 1109238405Sjkim has to return non-zero to report success: usually 1 to use opaque 1110238405Sjkim PRF input just if possible, or 2 to enforce use of the opaque PRF 1111238405Sjkim input. In the latter case, the library will abort the handshake 1112238405Sjkim if opaque PRF input is not successfully negotiated. 1113238405Sjkim 1114238405Sjkim Arguments 'peerinput' and 'len' given to the callback function 1115238405Sjkim will always be NULL and 0 in the case of a client. A server will 1116238405Sjkim see the client's opaque PRF input through these variables if 1117238405Sjkim available (NULL and 0 otherwise). Note that if the server 1118238405Sjkim provides an opaque PRF input, the length must be the same as the 1119238405Sjkim length of the client's opaque PRF input. 1120238405Sjkim 1121238405Sjkim Note that the callback function will only be called when creating 1122238405Sjkim a new session (session resumption can resume whatever was 1123238405Sjkim previously negotiated), and will not be called in SSL 2.0 1124238405Sjkim handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or 1125238405Sjkim SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended 1126238405Sjkim for applications that need to enforce opaque PRF input. 1127238405Sjkim 1128238405Sjkim [Bodo Moeller] 1129238405Sjkim 1130238405Sjkim *) Update ssl code to support digests other than SHA1+MD5 for handshake 1131238405Sjkim MAC. 1132238405Sjkim 1133238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 1134238405Sjkim 1135238405Sjkim *) Add RFC4507 support to OpenSSL. This includes the corrections in 1136238405Sjkim RFC4507bis. The encrypted ticket format is an encrypted encoded 1137238405Sjkim SSL_SESSION structure, that way new session features are automatically 1138238405Sjkim supported. 1139238405Sjkim 1140238405Sjkim If a client application caches session in an SSL_SESSION structure 1141238405Sjkim support is transparent because tickets are now stored in the encoded 1142238405Sjkim SSL_SESSION. 1143238405Sjkim 1144238405Sjkim The SSL_CTX structure automatically generates keys for ticket 1145238405Sjkim protection in servers so again support should be possible 1146238405Sjkim with no application modification. 1147238405Sjkim 1148238405Sjkim If a client or server wishes to disable RFC4507 support then the option 1149238405Sjkim SSL_OP_NO_TICKET can be set. 1150238405Sjkim 1151238405Sjkim Add a TLS extension debugging callback to allow the contents of any client 1152238405Sjkim or server extensions to be examined. 1153238405Sjkim 1154238405Sjkim This work was sponsored by Google. 1155238405Sjkim [Steve Henson] 1156238405Sjkim 1157238405Sjkim *) Final changes to avoid use of pointer pointer casts in OpenSSL. 1158238405Sjkim OpenSSL should now compile cleanly on gcc 4.2 1159238405Sjkim [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson] 1160238405Sjkim 1161238405Sjkim *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC 1162238405Sjkim support including streaming MAC support: this is required for GOST 1163238405Sjkim ciphersuite support. 1164238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson] 1165238405Sjkim 1166238405Sjkim *) Add option -stream to use PKCS#7 streaming in smime utility. New 1167238405Sjkim function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream() 1168238405Sjkim to output in BER and PEM format. 1169238405Sjkim [Steve Henson] 1170238405Sjkim 1171238405Sjkim *) Experimental support for use of HMAC via EVP_PKEY interface. This 1172238405Sjkim allows HMAC to be handled via the EVP_DigestSign*() interface. The 1173238405Sjkim EVP_PKEY "key" in this case is the HMAC key, potentially allowing 1174238405Sjkim ENGINE support for HMAC keys which are unextractable. New -mac and 1175238405Sjkim -macopt options to dgst utility. 1176238405Sjkim [Steve Henson] 1177238405Sjkim 1178238405Sjkim *) New option -sigopt to dgst utility. Update dgst to use 1179238405Sjkim EVP_Digest{Sign,Verify}*. These two changes make it possible to use 1180238405Sjkim alternative signing paramaters such as X9.31 or PSS in the dgst 1181238405Sjkim utility. 1182238405Sjkim [Steve Henson] 1183238405Sjkim 1184238405Sjkim *) Change ssl_cipher_apply_rule(), the internal function that does 1185238405Sjkim the work each time a ciphersuite string requests enabling 1186238405Sjkim ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or 1187238405Sjkim removing ("!foo+bar") a class of ciphersuites: Now it maintains 1188238405Sjkim the order of disabled ciphersuites such that those ciphersuites 1189238405Sjkim that most recently went from enabled to disabled not only stay 1190238405Sjkim in order with respect to each other, but also have higher priority 1191238405Sjkim than other disabled ciphersuites the next time ciphersuites are 1192238405Sjkim enabled again. 1193238405Sjkim 1194238405Sjkim This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable 1195238405Sjkim the same ciphersuites as with "HIGH" alone, but in a specific 1196238405Sjkim order where the PSK ciphersuites come first (since they are the 1197238405Sjkim most recently disabled ciphersuites when "HIGH" is parsed). 1198238405Sjkim 1199238405Sjkim Also, change ssl_create_cipher_list() (using this new 1200238405Sjkim funcionality) such that between otherwise identical 1201238405Sjkim cihpersuites, ephemeral ECDH is preferred over ephemeral DH in 1202238405Sjkim the default order. 1203238405Sjkim [Bodo Moeller] 1204238405Sjkim 1205238405Sjkim *) Change ssl_create_cipher_list() so that it automatically 1206238405Sjkim arranges the ciphersuites in reasonable order before starting 1207238405Sjkim to process the rule string. Thus, the definition for "DEFAULT" 1208238405Sjkim (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but 1209238405Sjkim remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH". 1210238405Sjkim This makes it much easier to arrive at a reasonable default order 1211238405Sjkim in applications for which anonymous ciphers are OK (meaning 1212238405Sjkim that you can't actually use DEFAULT). 1213238405Sjkim [Bodo Moeller; suggested by Victor Duchovni] 1214238405Sjkim 1215238405Sjkim *) Split the SSL/TLS algorithm mask (as used for ciphersuite string 1216238405Sjkim processing) into multiple integers instead of setting 1217238405Sjkim "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK", 1218238405Sjkim "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer. 1219238405Sjkim (These masks as well as the individual bit definitions are hidden 1220238405Sjkim away into the non-exported interface ssl/ssl_locl.h, so this 1221238405Sjkim change to the definition of the SSL_CIPHER structure shouldn't 1222238405Sjkim affect applications.) This give us more bits for each of these 1223238405Sjkim categories, so there is no longer a need to coagulate AES128 and 1224238405Sjkim AES256 into a single algorithm bit, and to coagulate Camellia128 1225238405Sjkim and Camellia256 into a single algorithm bit, which has led to all 1226238405Sjkim kinds of kludges. 1227238405Sjkim 1228238405Sjkim Thus, among other things, the kludge introduced in 0.9.7m and 1229238405Sjkim 0.9.8e for masking out AES256 independently of AES128 or masking 1230238405Sjkim out Camellia256 independently of AES256 is not needed here in 0.9.9. 1231238405Sjkim 1232238405Sjkim With the change, we also introduce new ciphersuite aliases that 1233238405Sjkim so far were missing: "AES128", "AES256", "CAMELLIA128", and 1234238405Sjkim "CAMELLIA256". 1235238405Sjkim [Bodo Moeller] 1236238405Sjkim 1237238405Sjkim *) Add support for dsa-with-SHA224 and dsa-with-SHA256. 1238238405Sjkim Use the leftmost N bytes of the signature input if the input is 1239238405Sjkim larger than the prime q (with N being the size in bytes of q). 1240238405Sjkim [Nils Larsch] 1241238405Sjkim 1242238405Sjkim *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses 1243238405Sjkim it yet and it is largely untested. 1244238405Sjkim [Steve Henson] 1245238405Sjkim 1246238405Sjkim *) Add support for the ecdsa-with-SHA224/256/384/512 signature types. 1247238405Sjkim [Nils Larsch] 1248238405Sjkim 1249238405Sjkim *) Initial incomplete changes to avoid need for function casts in OpenSSL 1250238405Sjkim some compilers (gcc 4.2 and later) reject their use. Safestack is 1251238405Sjkim reimplemented. Update ASN1 to avoid use of legacy functions. 1252238405Sjkim [Steve Henson] 1253238405Sjkim 1254238405Sjkim *) Win32/64 targets are linked with Winsock2. 1255238405Sjkim [Andy Polyakov] 1256238405Sjkim 1257238405Sjkim *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected 1258238405Sjkim to external functions. This can be used to increase CRL handling 1259238405Sjkim efficiency especially when CRLs are very large by (for example) storing 1260238405Sjkim the CRL revoked certificates in a database. 1261238405Sjkim [Steve Henson] 1262238405Sjkim 1263238405Sjkim *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so 1264238405Sjkim new CRLs added to a directory can be used. New command line option 1265238405Sjkim -verify_return_error to s_client and s_server. This causes real errors 1266238405Sjkim to be returned by the verify callback instead of carrying on no matter 1267238405Sjkim what. This reflects the way a "real world" verify callback would behave. 1268238405Sjkim [Steve Henson] 1269238405Sjkim 1270238405Sjkim *) GOST engine, supporting several GOST algorithms and public key formats. 1271238405Sjkim Kindly donated by Cryptocom. 1272238405Sjkim [Cryptocom] 1273238405Sjkim 1274238405Sjkim *) Partial support for Issuing Distribution Point CRL extension. CRLs 1275238405Sjkim partitioned by DP are handled but no indirect CRL or reason partitioning 1276238405Sjkim (yet). Complete overhaul of CRL handling: now the most suitable CRL is 1277238405Sjkim selected via a scoring technique which handles IDP and AKID in CRLs. 1278238405Sjkim [Steve Henson] 1279238405Sjkim 1280238405Sjkim *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which 1281238405Sjkim will ultimately be used for all verify operations: this will remove the 1282238405Sjkim X509_STORE dependency on certificate verification and allow alternative 1283238405Sjkim lookup methods. X509_STORE based implementations of these two callbacks. 1284238405Sjkim [Steve Henson] 1285238405Sjkim 1286238405Sjkim *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names. 1287238405Sjkim Modify get_crl() to find a valid (unexpired) CRL if possible. 1288238405Sjkim [Steve Henson] 1289238405Sjkim 1290238405Sjkim *) New function X509_CRL_match() to check if two CRLs are identical. Normally 1291238405Sjkim this would be called X509_CRL_cmp() but that name is already used by 1292238405Sjkim a function that just compares CRL issuer names. Cache several CRL 1293238405Sjkim extensions in X509_CRL structure and cache CRLDP in X509. 1294238405Sjkim [Steve Henson] 1295238405Sjkim 1296238405Sjkim *) Store a "canonical" representation of X509_NAME structure (ASN1 Name) 1297238405Sjkim this maps equivalent X509_NAME structures into a consistent structure. 1298238405Sjkim Name comparison can then be performed rapidly using memcmp(). 1299238405Sjkim [Steve Henson] 1300238405Sjkim 1301238405Sjkim *) Non-blocking OCSP request processing. Add -timeout option to ocsp 1302238405Sjkim utility. 1303238405Sjkim [Steve Henson] 1304238405Sjkim 1305238405Sjkim *) Allow digests to supply their own micalg string for S/MIME type using 1306238405Sjkim the ctrl EVP_MD_CTRL_MICALG. 1307238405Sjkim [Steve Henson] 1308238405Sjkim 1309238405Sjkim *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the 1310238405Sjkim EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN 1311238405Sjkim ctrl. It can then customise the structure before and/or after signing 1312238405Sjkim if necessary. 1313238405Sjkim [Steve Henson] 1314238405Sjkim 1315238405Sjkim *) New function OBJ_add_sigid() to allow application defined signature OIDs 1316238405Sjkim to be added to OpenSSLs internal tables. New function OBJ_sigid_free() 1317238405Sjkim to free up any added signature OIDs. 1318238405Sjkim [Steve Henson] 1319238405Sjkim 1320238405Sjkim *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(), 1321238405Sjkim EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal 1322238405Sjkim digest and cipher tables. New options added to openssl utility: 1323238405Sjkim list-message-digest-algorithms and list-cipher-algorithms. 1324238405Sjkim [Steve Henson] 1325238405Sjkim 1326238405Sjkim *) Change the array representation of binary polynomials: the list 1327238405Sjkim of degrees of non-zero coefficients is now terminated with -1. 1328238405Sjkim Previously it was terminated with 0, which was also part of the 1329238405Sjkim value; thus, the array representation was not applicable to 1330238405Sjkim polynomials where t^0 has coefficient zero. This change makes 1331238405Sjkim the array representation useful in a more general context. 1332238405Sjkim [Douglas Stebila] 1333238405Sjkim 1334238405Sjkim *) Various modifications and fixes to SSL/TLS cipher string 1335238405Sjkim handling. For ECC, the code now distinguishes between fixed ECDH 1336238405Sjkim with RSA certificates on the one hand and with ECDSA certificates 1337238405Sjkim on the other hand, since these are separate ciphersuites. The 1338238405Sjkim unused code for Fortezza ciphersuites has been removed. 1339238405Sjkim 1340238405Sjkim For consistency with EDH, ephemeral ECDH is now called "EECDH" 1341238405Sjkim (not "ECDHE"). For consistency with the code for DH 1342238405Sjkim certificates, use of ECDH certificates is now considered ECDH 1343238405Sjkim authentication, not RSA or ECDSA authentication (the latter is 1344238405Sjkim merely the CA's signing algorithm and not actively used in the 1345238405Sjkim protocol). 1346238405Sjkim 1347238405Sjkim The temporary ciphersuite alias "ECCdraft" is no longer 1348238405Sjkim available, and ECC ciphersuites are no longer excluded from "ALL" 1349238405Sjkim and "DEFAULT". The following aliases now exist for RFC 4492 1350238405Sjkim ciphersuites, most of these by analogy with the DH case: 1351238405Sjkim 1352238405Sjkim kECDHr - ECDH cert, signed with RSA 1353238405Sjkim kECDHe - ECDH cert, signed with ECDSA 1354238405Sjkim kECDH - ECDH cert (signed with either RSA or ECDSA) 1355238405Sjkim kEECDH - ephemeral ECDH 1356238405Sjkim ECDH - ECDH cert or ephemeral ECDH 1357238405Sjkim 1358238405Sjkim aECDH - ECDH cert 1359238405Sjkim aECDSA - ECDSA cert 1360238405Sjkim ECDSA - ECDSA cert 1361238405Sjkim 1362238405Sjkim AECDH - anonymous ECDH 1363238405Sjkim EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH") 1364238405Sjkim 1365238405Sjkim [Bodo Moeller] 1366238405Sjkim 1367238405Sjkim *) Add additional S/MIME capabilities for AES and GOST ciphers if supported. 1368238405Sjkim Use correct micalg parameters depending on digest(s) in signed message. 1369238405Sjkim [Steve Henson] 1370238405Sjkim 1371238405Sjkim *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process 1372238405Sjkim an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code. 1373238405Sjkim [Steve Henson] 1374238405Sjkim 1375238405Sjkim *) Initial engine support for EVP_PKEY_METHOD. New functions to permit 1376238405Sjkim an engine to register a method. Add ENGINE lookups for methods and 1377238405Sjkim functional reference processing. 1378238405Sjkim [Steve Henson] 1379238405Sjkim 1380238405Sjkim *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of 1381238405Sjkim EVP_{Sign,Verify}* which allow an application to customise the signature 1382238405Sjkim process. 1383238405Sjkim [Steve Henson] 1384238405Sjkim 1385238405Sjkim *) New -resign option to smime utility. This adds one or more signers 1386238405Sjkim to an existing PKCS#7 signedData structure. Also -md option to use an 1387238405Sjkim alternative message digest algorithm for signing. 1388238405Sjkim [Steve Henson] 1389238405Sjkim 1390238405Sjkim *) Tidy up PKCS#7 routines and add new functions to make it easier to 1391238405Sjkim create PKCS7 structures containing multiple signers. Update smime 1392238405Sjkim application to support multiple signers. 1393238405Sjkim [Steve Henson] 1394238405Sjkim 1395238405Sjkim *) New -macalg option to pkcs12 utility to allow setting of an alternative 1396238405Sjkim digest MAC. 1397238405Sjkim [Steve Henson] 1398238405Sjkim 1399238405Sjkim *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC. 1400238405Sjkim Reorganize PBE internals to lookup from a static table using NIDs, 1401238405Sjkim add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl: 1402238405Sjkim EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative 1403238405Sjkim PRF which will be automatically used with PBES2. 1404238405Sjkim [Steve Henson] 1405238405Sjkim 1406238405Sjkim *) Replace the algorithm specific calls to generate keys in "req" with the 1407238405Sjkim new API. 1408238405Sjkim [Steve Henson] 1409238405Sjkim 1410238405Sjkim *) Update PKCS#7 enveloped data routines to use new API. This is now 1411238405Sjkim supported by any public key method supporting the encrypt operation. A 1412238405Sjkim ctrl is added to allow the public key algorithm to examine or modify 1413238405Sjkim the PKCS#7 RecipientInfo structure if it needs to: for RSA this is 1414238405Sjkim a no op. 1415238405Sjkim [Steve Henson] 1416238405Sjkim 1417238405Sjkim *) Add a ctrl to asn1 method to allow a public key algorithm to express 1418238405Sjkim a default digest type to use. In most cases this will be SHA1 but some 1419238405Sjkim algorithms (such as GOST) need to specify an alternative digest. The 1420238405Sjkim return value indicates how strong the prefernce is 1 means optional and 1421238405Sjkim 2 is mandatory (that is it is the only supported type). Modify 1422238405Sjkim ASN1_item_sign() to accept a NULL digest argument to indicate it should 1423238405Sjkim use the default md. Update openssl utilities to use the default digest 1424238405Sjkim type for signing if it is not explicitly indicated. 1425238405Sjkim [Steve Henson] 1426238405Sjkim 1427238405Sjkim *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New 1428238405Sjkim EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant 1429238405Sjkim signing method from the key type. This effectively removes the link 1430238405Sjkim between digests and public key types. 1431238405Sjkim [Steve Henson] 1432238405Sjkim 1433238405Sjkim *) Add an OID cross reference table and utility functions. Its purpose is to 1434238405Sjkim translate between signature OIDs such as SHA1WithrsaEncryption and SHA1, 1435238405Sjkim rsaEncryption. This will allow some of the algorithm specific hackery 1436238405Sjkim needed to use the correct OID to be removed. 1437238405Sjkim [Steve Henson] 1438238405Sjkim 1439238405Sjkim *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO 1440238405Sjkim structures for PKCS7_sign(). They are now set up by the relevant public 1441238405Sjkim key ASN1 method. 1442238405Sjkim [Steve Henson] 1443238405Sjkim 1444238405Sjkim *) Add provisional EC pkey method with support for ECDSA and ECDH. 1445238405Sjkim [Steve Henson] 1446238405Sjkim 1447238405Sjkim *) Add support for key derivation (agreement) in the API, DH method and 1448238405Sjkim pkeyutl. 1449238405Sjkim [Steve Henson] 1450238405Sjkim 1451238405Sjkim *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support 1452238405Sjkim public and private key formats. As a side effect these add additional 1453238405Sjkim command line functionality not previously available: DSA signatures can be 1454238405Sjkim generated and verified using pkeyutl and DH key support and generation in 1455238405Sjkim pkey, genpkey. 1456238405Sjkim [Steve Henson] 1457238405Sjkim 1458238405Sjkim *) BeOS support. 1459238405Sjkim [Oliver Tappe <zooey@hirschkaefer.de>] 1460238405Sjkim 1461238405Sjkim *) New make target "install_html_docs" installs HTML renditions of the 1462238405Sjkim manual pages. 1463238405Sjkim [Oliver Tappe <zooey@hirschkaefer.de>] 1464238405Sjkim 1465238405Sjkim *) New utility "genpkey" this is analagous to "genrsa" etc except it can 1466238405Sjkim generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to 1467238405Sjkim support key and parameter generation and add initial key generation 1468238405Sjkim functionality for RSA. 1469238405Sjkim [Steve Henson] 1470238405Sjkim 1471238405Sjkim *) Add functions for main EVP_PKEY_method operations. The undocumented 1472238405Sjkim functions EVP_PKEY_{encrypt,decrypt} have been renamed to 1473238405Sjkim EVP_PKEY_{encrypt,decrypt}_old. 1474238405Sjkim [Steve Henson] 1475238405Sjkim 1476238405Sjkim *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public 1477238405Sjkim key API, doesn't do much yet. 1478238405Sjkim [Steve Henson] 1479238405Sjkim 1480238405Sjkim *) New function EVP_PKEY_asn1_get0_info() to retrieve information about 1481238405Sjkim public key algorithms. New option to openssl utility: 1482238405Sjkim "list-public-key-algorithms" to print out info. 1483238405Sjkim [Steve Henson] 1484238405Sjkim 1485238405Sjkim *) Implement the Supported Elliptic Curves Extension for 1486238405Sjkim ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1487238405Sjkim [Douglas Stebila] 1488238405Sjkim 1489238405Sjkim *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or 1490238405Sjkim EVP_CIPHER structures to avoid later problems in EVP_cleanup(). 1491238405Sjkim [Steve Henson] 1492238405Sjkim 1493238405Sjkim *) New utilities pkey and pkeyparam. These are similar to algorithm specific 1494238405Sjkim utilities such as rsa, dsa, dsaparam etc except they process any key 1495238405Sjkim type. 1496238405Sjkim [Steve Henson] 1497238405Sjkim 1498238405Sjkim *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New 1499238405Sjkim functions EVP_PKEY_print_public(), EVP_PKEY_print_private(), 1500238405Sjkim EVP_PKEY_print_param() to print public key data from an EVP_PKEY 1501238405Sjkim structure. 1502238405Sjkim [Steve Henson] 1503238405Sjkim 1504238405Sjkim *) Initial support for pluggable public key ASN1. 1505238405Sjkim De-spaghettify the public key ASN1 handling. Move public and private 1506238405Sjkim key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate 1507238405Sjkim algorithm specific handling to a single module within the relevant 1508238405Sjkim algorithm directory. Add functions to allow (near) opaque processing 1509238405Sjkim of public and private key structures. 1510238405Sjkim [Steve Henson] 1511238405Sjkim 1512238405Sjkim *) Implement the Supported Point Formats Extension for 1513238405Sjkim ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1514238405Sjkim [Douglas Stebila] 1515238405Sjkim 1516238405Sjkim *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members 1517238405Sjkim for the psk identity [hint] and the psk callback functions to the 1518238405Sjkim SSL_SESSION, SSL and SSL_CTX structure. 1519238405Sjkim 1520238405Sjkim New ciphersuites: 1521238405Sjkim PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA, 1522238405Sjkim PSK-AES256-CBC-SHA 1523238405Sjkim 1524238405Sjkim New functions: 1525238405Sjkim SSL_CTX_use_psk_identity_hint 1526238405Sjkim SSL_get_psk_identity_hint 1527238405Sjkim SSL_get_psk_identity 1528238405Sjkim SSL_use_psk_identity_hint 1529238405Sjkim 1530238405Sjkim [Mika Kousa and Pasi Eronen of Nokia Corporation] 1531238405Sjkim 1532238405Sjkim *) Add RFC 3161 compliant time stamp request creation, response generation 1533238405Sjkim and response verification functionality. 1534238405Sjkim [Zolt�n Gl�zik <zglozik@opentsa.org>, The OpenTSA Project] 1535238405Sjkim 1536238405Sjkim *) Add initial support for TLS extensions, specifically for the server_name 1537238405Sjkim extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 1538238405Sjkim have new members for a host name. The SSL data structure has an 1539238405Sjkim additional member SSL_CTX *initial_ctx so that new sessions can be 1540238405Sjkim stored in that context to allow for session resumption, even after the 1541238405Sjkim SSL has been switched to a new SSL_CTX in reaction to a client's 1542238405Sjkim server_name extension. 1543238405Sjkim 1544238405Sjkim New functions (subject to change): 1545238405Sjkim 1546238405Sjkim SSL_get_servername() 1547238405Sjkim SSL_get_servername_type() 1548238405Sjkim SSL_set_SSL_CTX() 1549238405Sjkim 1550238405Sjkim New CTRL codes and macros (subject to change): 1551238405Sjkim 1552238405Sjkim SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 1553238405Sjkim - SSL_CTX_set_tlsext_servername_callback() 1554238405Sjkim SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 1555238405Sjkim - SSL_CTX_set_tlsext_servername_arg() 1556238405Sjkim SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 1557238405Sjkim 1558238405Sjkim openssl s_client has a new '-servername ...' option. 1559238405Sjkim 1560238405Sjkim openssl s_server has new options '-servername_host ...', '-cert2 ...', 1561238405Sjkim '-key2 ...', '-servername_fatal' (subject to change). This allows 1562238405Sjkim testing the HostName extension for a specific single host name ('-cert' 1563238405Sjkim and '-key' remain fallbacks for handshakes without HostName 1564238405Sjkim negotiation). If the unrecogninzed_name alert has to be sent, this by 1565238405Sjkim default is a warning; it becomes fatal with the '-servername_fatal' 1566238405Sjkim option. 1567238405Sjkim 1568238405Sjkim [Peter Sylvester, Remy Allais, Christophe Renou] 1569238405Sjkim 1570238405Sjkim *) Whirlpool hash implementation is added. 1571238405Sjkim [Andy Polyakov] 1572238405Sjkim 1573238405Sjkim *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to 1574238405Sjkim bn(64,32). Because of instruction set limitations it doesn't have 1575238405Sjkim any negative impact on performance. This was done mostly in order 1576238405Sjkim to make it possible to share assembler modules, such as bn_mul_mont 1577238405Sjkim implementations, between 32- and 64-bit builds without hassle. 1578238405Sjkim [Andy Polyakov] 1579238405Sjkim 1580238405Sjkim *) Move code previously exiled into file crypto/ec/ec2_smpt.c 1581238405Sjkim to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP 1582238405Sjkim macro. 1583238405Sjkim [Bodo Moeller] 1584238405Sjkim 1585238405Sjkim *) New candidate for BIGNUM assembler implementation, bn_mul_mont, 1586238405Sjkim dedicated Montgomery multiplication procedure, is introduced. 1587238405Sjkim BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher 1588238405Sjkim "64-bit" performance on certain 32-bit targets. 1589238405Sjkim [Andy Polyakov] 1590238405Sjkim 1591238405Sjkim *) New option SSL_OP_NO_COMP to disable use of compression selectively 1592238405Sjkim in SSL structures. New SSL ctrl to set maximum send fragment size. 1593238405Sjkim Save memory by seeting the I/O buffer sizes dynamically instead of 1594238405Sjkim using the maximum available value. 1595238405Sjkim [Steve Henson] 1596238405Sjkim 1597238405Sjkim *) New option -V for 'openssl ciphers'. This prints the ciphersuite code 1598238405Sjkim in addition to the text details. 1599238405Sjkim [Bodo Moeller] 1600238405Sjkim 1601238405Sjkim *) Very, very preliminary EXPERIMENTAL support for printing of general 1602238405Sjkim ASN1 structures. This currently produces rather ugly output and doesn't 1603238405Sjkim handle several customised structures at all. 1604238405Sjkim [Steve Henson] 1605238405Sjkim 1606238405Sjkim *) Integrated support for PVK file format and some related formats such 1607238405Sjkim as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support 1608238405Sjkim these in the 'rsa' and 'dsa' utilities. 1609238405Sjkim [Steve Henson] 1610238405Sjkim 1611238405Sjkim *) Support for PKCS#1 RSAPublicKey format on rsa utility command line. 1612238405Sjkim [Steve Henson] 1613238405Sjkim 1614238405Sjkim *) Remove the ancient ASN1_METHOD code. This was only ever used in one 1615238405Sjkim place for the (very old) "NETSCAPE" format certificates which are now 1616238405Sjkim handled using new ASN1 code equivalents. 1617238405Sjkim [Steve Henson] 1618238405Sjkim 1619238405Sjkim *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD 1620238405Sjkim pointer and make the SSL_METHOD parameter in SSL_CTX_new, 1621238405Sjkim SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'. 1622238405Sjkim [Nils Larsch] 1623238405Sjkim 1624238405Sjkim *) Modify CRL distribution points extension code to print out previously 1625238405Sjkim unsupported fields. Enhance extension setting code to allow setting of 1626238405Sjkim all fields. 1627238405Sjkim [Steve Henson] 1628238405Sjkim 1629238405Sjkim *) Add print and set support for Issuing Distribution Point CRL extension. 1630238405Sjkim [Steve Henson] 1631238405Sjkim 1632238405Sjkim *) Change 'Configure' script to enable Camellia by default. 1633238405Sjkim [NTT] 1634238405Sjkim 1635264331Sjkim Changes between 0.9.8x and 0.9.8y [5 Feb 2013] 1636264331Sjkim 1637264331Sjkim *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 1638264331Sjkim 1639264331Sjkim This addresses the flaw in CBC record processing discovered by 1640264331Sjkim Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 1641264331Sjkim at: http://www.isg.rhul.ac.uk/tls/ 1642264331Sjkim 1643264331Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 1644264331Sjkim Security Group at Royal Holloway, University of London 1645264331Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 1646264331Sjkim Emilia K�sper for the initial patch. 1647264331Sjkim (CVE-2013-0169) 1648264331Sjkim [Emilia K�sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 1649264331Sjkim 1650264331Sjkim *) Return an error when checking OCSP signatures when key is NULL. 1651264331Sjkim This fixes a DoS attack. (CVE-2013-0166) 1652264331Sjkim [Steve Henson] 1653264331Sjkim 1654264331Sjkim *) Call OCSP Stapling callback after ciphersuite has been chosen, so 1655264331Sjkim the right response is stapled. Also change SSL_get_certificate() 1656264331Sjkim so it returns the certificate actually sent. 1657264331Sjkim See http://rt.openssl.org/Ticket/Display.html?id=2836. 1658264331Sjkim (This is a backport) 1659264331Sjkim [Rob Stradling <rob.stradling@comodo.com>] 1660264331Sjkim 1661264331Sjkim *) Fix possible deadlock when decoding public keys. 1662264331Sjkim [Steve Henson] 1663264331Sjkim 1664264331Sjkim Changes between 0.9.8w and 0.9.8x [10 May 2012] 1665264331Sjkim 1666264331Sjkim *) Sanity check record length before skipping explicit IV in DTLS 1667264331Sjkim to fix DoS attack. 1668264331Sjkim 1669264331Sjkim Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 1670264331Sjkim fuzzing as a service testing platform. 1671264331Sjkim (CVE-2012-2333) 1672264331Sjkim [Steve Henson] 1673264331Sjkim 1674264331Sjkim *) Initialise tkeylen properly when encrypting CMS messages. 1675264331Sjkim Thanks to Solar Designer of Openwall for reporting this issue. 1676264331Sjkim [Steve Henson] 1677264331Sjkim 1678264331Sjkim Changes between 0.9.8v and 0.9.8w [23 Apr 2012] 1679264331Sjkim 1680264331Sjkim *) The fix for CVE-2012-2110 did not take into account that the 1681264331Sjkim 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an 1682264331Sjkim int in OpenSSL 0.9.8, making it still vulnerable. Fix by 1683264331Sjkim rejecting negative len parameter. (CVE-2012-2131) 1684264331Sjkim [Tomas Hoger <thoger@redhat.com>] 1685264331Sjkim 1686264331Sjkim Changes between 0.9.8u and 0.9.8v [19 Apr 2012] 1687264331Sjkim 1688264331Sjkim *) Check for potentially exploitable overflows in asn1_d2i_read_bio 1689264331Sjkim BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 1690264331Sjkim in CRYPTO_realloc_clean. 1691264331Sjkim 1692264331Sjkim Thanks to Tavis Ormandy, Google Security Team, for discovering this 1693264331Sjkim issue and to Adam Langley <agl@chromium.org> for fixing it. 1694264331Sjkim (CVE-2012-2110) 1695264331Sjkim [Adam Langley (Google), Tavis Ormandy, Google Security Team] 1696264331Sjkim 1697264331Sjkim Changes between 0.9.8t and 0.9.8u [12 Mar 2012] 1698264331Sjkim 1699264331Sjkim *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 1700264331Sjkim in CMS and PKCS7 code. When RSA decryption fails use a random key for 1701264331Sjkim content decryption and always return the same error. Note: this attack 1702264331Sjkim needs on average 2^20 messages so it only affects automated senders. The 1703264331Sjkim old behaviour can be reenabled in the CMS code by setting the 1704264331Sjkim CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 1705264331Sjkim an MMA defence is not necessary. 1706264331Sjkim Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 1707264331Sjkim this issue. (CVE-2012-0884) 1708264331Sjkim [Steve Henson] 1709264331Sjkim 1710264331Sjkim *) Fix CVE-2011-4619: make sure we really are receiving a 1711264331Sjkim client hello before rejecting multiple SGC restarts. Thanks to 1712264331Sjkim Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 1713264331Sjkim [Steve Henson] 1714264331Sjkim 1715237657Sjkim Changes between 0.9.8s and 0.9.8t [18 Jan 2012] 1716237657Sjkim 1717237657Sjkim *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 1718237657Sjkim Thanks to Antonio Martin, Enterprise Secure Access Research and 1719237657Sjkim Development, Cisco Systems, Inc. for discovering this bug and 1720237657Sjkim preparing a fix. (CVE-2012-0050) 1721237657Sjkim [Antonio Martin] 1722264331Sjkim 1723237657Sjkim Changes between 0.9.8r and 0.9.8s [4 Jan 2012] 1724237657Sjkim 1725237657Sjkim *) Nadhem Alfardan and Kenny Paterson have discovered an extension 1726237657Sjkim of the Vaudenay padding oracle attack on CBC mode encryption 1727237657Sjkim which enables an efficient plaintext recovery attack against 1728237657Sjkim the OpenSSL implementation of DTLS. Their attack exploits timing 1729237657Sjkim differences arising during decryption processing. A research 1730237657Sjkim paper describing this attack can be found at: 1731237657Sjkim http://www.isg.rhul.ac.uk/~kp/dtls.pdf 1732237657Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 1733237657Sjkim Security Group at Royal Holloway, University of London 1734237657Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 1735237657Sjkim <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 1736237657Sjkim for preparing the fix. (CVE-2011-4108) 1737237657Sjkim [Robin Seggelmann, Michael Tuexen] 1738237657Sjkim 1739237657Sjkim *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109) 1740237657Sjkim [Ben Laurie, Kasper <ekasper@google.com>] 1741237657Sjkim 1742237657Sjkim *) Clear bytes used for block padding of SSL 3.0 records. 1743237657Sjkim (CVE-2011-4576) 1744237657Sjkim [Adam Langley (Google)] 1745237657Sjkim 1746237657Sjkim *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 1747237657Sjkim Kadianakis <desnacked@gmail.com> for discovering this issue and 1748237657Sjkim Adam Langley for preparing the fix. (CVE-2011-4619) 1749237657Sjkim [Adam Langley (Google)] 1750237657Sjkim 1751237657Sjkim *) Prevent malformed RFC3779 data triggering an assertion failure. 1752237657Sjkim Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 1753237657Sjkim and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 1754237657Sjkim [Rob Austein <sra@hactrn.net>] 1755237657Sjkim 1756237657Sjkim *) Fix ssl_ciph.c set-up race. 1757237657Sjkim [Adam Langley (Google)] 1758237657Sjkim 1759237657Sjkim *) Fix spurious failures in ecdsatest.c. 1760237657Sjkim [Emilia K�sper (Google)] 1761237657Sjkim 1762237657Sjkim *) Fix the BIO_f_buffer() implementation (which was mixing different 1763237657Sjkim interpretations of the '..._len' fields). 1764237657Sjkim [Adam Langley (Google)] 1765237657Sjkim 1766237657Sjkim *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 1767237657Sjkim BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 1768237657Sjkim threads won't reuse the same blinding coefficients. 1769237657Sjkim 1770237657Sjkim This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 1771237657Sjkim lock to call BN_BLINDING_invert_ex, and avoids one use of 1772237657Sjkim BN_BLINDING_update for each BN_BLINDING structure (previously, 1773237657Sjkim the last update always remained unused). 1774237657Sjkim [Emilia K�sper (Google)] 1775237657Sjkim 1776237657Sjkim *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 1777237657Sjkim for multi-threaded use of ECDH. 1778237657Sjkim [Adam Langley (Google)] 1779237657Sjkim 1780237657Sjkim *) Fix x509_name_ex_d2i memory leak on bad inputs. 1781237657Sjkim [Bodo Moeller] 1782237657Sjkim 1783237657Sjkim *) Add protection against ECDSA timing attacks as mentioned in the paper 1784237657Sjkim by Billy Bob Brumley and Nicola Tuveri, see: 1785237657Sjkim 1786237657Sjkim http://eprint.iacr.org/2011/232.pdf 1787237657Sjkim 1788237657Sjkim [Billy Bob Brumley and Nicola Tuveri] 1789237657Sjkim 1790237657Sjkim Changes between 0.9.8q and 0.9.8r [8 Feb 2011] 1791237657Sjkim 1792237657Sjkim *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 1793237657Sjkim [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 1794237657Sjkim 1795237657Sjkim *) Fix bug in string printing code: if *any* escaping is enabled we must 1796237657Sjkim escape the escape character (backslash) or the resulting string is 1797237657Sjkim ambiguous. 1798237657Sjkim [Steve Henson] 1799237657Sjkim 1800216166Ssimon Changes between 0.9.8p and 0.9.8q [2 Dec 2010] 1801216166Ssimon 1802216166Ssimon *) Disable code workaround for ancient and obsolete Netscape browsers 1803216166Ssimon and servers: an attacker can use it in a ciphersuite downgrade attack. 1804216166Ssimon Thanks to Martin Rex for discovering this bug. CVE-2010-4180 1805216166Ssimon [Steve Henson] 1806216166Ssimon 1807216166Ssimon *) Fixed J-PAKE implementation error, originally discovered by 1808216166Ssimon Sebastien Martini, further info and confirmation from Stefan 1809216166Ssimon Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 1810216166Ssimon [Ben Laurie] 1811216166Ssimon 1812215697Ssimon Changes between 0.9.8o and 0.9.8p [16 Nov 2010] 1813215697Ssimon 1814215697Ssimon *) Fix extension code to avoid race conditions which can result in a buffer 1815215697Ssimon overrun vulnerability: resumed sessions must not be modified as they can 1816215697Ssimon be shared by multiple threads. CVE-2010-3864 1817215697Ssimon [Steve Henson] 1818215697Ssimon 1819215697Ssimon *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939 1820215697Ssimon [Steve Henson] 1821215697Ssimon 1822215697Ssimon *) Don't reencode certificate when calculating signature: cache and use 1823215697Ssimon the original encoding instead. This makes signature verification of 1824215697Ssimon some broken encodings work correctly. 1825215697Ssimon [Steve Henson] 1826215697Ssimon 1827215697Ssimon *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT 1828215697Ssimon is also one of the inputs. 1829215697Ssimon [Emilia K�sper <emilia.kasper@esat.kuleuven.be> (Google)] 1830215697Ssimon 1831215697Ssimon *) Don't repeatedly append PBE algorithms to table if they already exist. 1832215697Ssimon Sort table on each new add. This effectively makes the table read only 1833215697Ssimon after all algorithms are added and subsequent calls to PKCS12_pbe_add 1834215697Ssimon etc are non-op. 1835215697Ssimon [Steve Henson] 1836215697Ssimon 1837215697Ssimon Changes between 0.9.8n and 0.9.8o [01 Jun 2010] 1838215697Ssimon 1839215697Ssimon [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after 1840215697Ssimon OpenSSL 1.0.0.] 1841215697Ssimon 1842215697Ssimon *) Correct a typo in the CMS ASN1 module which can result in invalid memory 1843215697Ssimon access or freeing data twice (CVE-2010-0742) 1844215697Ssimon [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>] 1845215697Ssimon 1846215697Ssimon *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more 1847215697Ssimon common in certificates and some applications which only call 1848215697Ssimon SSL_library_init and not OpenSSL_add_all_algorithms() will fail. 1849215697Ssimon [Steve Henson] 1850215697Ssimon 1851215697Ssimon *) VMS fixes: 1852215697Ssimon Reduce copying into .apps and .test in makevms.com 1853215697Ssimon Don't try to use blank CA certificate in CA.com 1854215697Ssimon Allow use of C files from original directories in maketests.com 1855215697Ssimon [Steven M. Schweda" <sms@antinode.info>] 1856215697Ssimon 1857206046Ssimon Changes between 0.9.8m and 0.9.8n [24 Mar 2010] 1858206046Ssimon 1859206046Ssimon *) When rejecting SSL/TLS records due to an incorrect version number, never 1860206046Ssimon update s->server with a new major version number. As of 1861206046Ssimon - OpenSSL 0.9.8m if 'short' is a 16-bit type, 1862206046Ssimon - OpenSSL 0.9.8f if 'short' is longer than 16 bits, 1863206046Ssimon the previous behavior could result in a read attempt at NULL when 1864206046Ssimon receiving specific incorrect SSL/TLS records once record payload 1865206046Ssimon protection is active. (CVE-2010-0740) 1866206046Ssimon [Bodo Moeller, Adam Langley <agl@chromium.org>] 1867206046Ssimon 1868206046Ssimon *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL 1869206046Ssimon could be crashed if the relevant tables were not present (e.g. chrooted). 1870206046Ssimon [Tomas Hoger <thoger@redhat.com>] 1871206046Ssimon 1872205128Ssimon Changes between 0.9.8l and 0.9.8m [25 Feb 2010] 1873205128Ssimon 1874205128Ssimon *) Always check bn_wexpend() return values for failure. (CVE-2009-3245) 1875205128Ssimon [Martin Olsson, Neel Mehta] 1876205128Ssimon 1877205128Ssimon *) Fix X509_STORE locking: Every 'objs' access requires a lock (to 1878205128Ssimon accommodate for stack sorting, always a write lock!). 1879205128Ssimon [Bodo Moeller] 1880205128Ssimon 1881205128Ssimon *) On some versions of WIN32 Heap32Next is very slow. This can cause 1882205128Ssimon excessive delays in the RAND_poll(): over a minute. As a workaround 1883205128Ssimon include a time check in the inner Heap32Next loop too. 1884205128Ssimon [Steve Henson] 1885205128Ssimon 1886205128Ssimon *) The code that handled flushing of data in SSL/TLS originally used the 1887205128Ssimon BIO_CTRL_INFO ctrl to see if any data was pending first. This caused 1888205128Ssimon the problem outlined in PR#1949. The fix suggested there however can 1889205128Ssimon trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions 1890205128Ssimon of Apache). So instead simplify the code to flush unconditionally. 1891205128Ssimon This should be fine since flushing with no data to flush is a no op. 1892205128Ssimon [Steve Henson] 1893205128Ssimon 1894205128Ssimon *) Handle TLS versions 2.0 and later properly and correctly use the 1895205128Ssimon highest version of TLS/SSL supported. Although TLS >= 2.0 is some way 1896205128Ssimon off ancient servers have a habit of sticking around for a while... 1897205128Ssimon [Steve Henson] 1898205128Ssimon 1899205128Ssimon *) Modify compression code so it frees up structures without using the 1900205128Ssimon ex_data callbacks. This works around a problem where some applications 1901205128Ssimon call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when 1902205128Ssimon restarting) then use compression (e.g. SSL with compression) later. 1903205128Ssimon This results in significant per-connection memory leaks and 1904205128Ssimon has caused some security issues including CVE-2008-1678 and 1905205128Ssimon CVE-2009-4355. 1906205128Ssimon [Steve Henson] 1907205128Ssimon 1908205128Ssimon *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't 1909205128Ssimon change when encrypting or decrypting. 1910205128Ssimon [Bodo Moeller] 1911205128Ssimon 1912205128Ssimon *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to 1913205128Ssimon connect and renegotiate with servers which do not support RI. 1914205128Ssimon Until RI is more widely deployed this option is enabled by default. 1915205128Ssimon [Steve Henson] 1916205128Ssimon 1917205128Ssimon *) Add "missing" ssl ctrls to clear options and mode. 1918205128Ssimon [Steve Henson] 1919205128Ssimon 1920205128Ssimon *) If client attempts to renegotiate and doesn't support RI respond with 1921205128Ssimon a no_renegotiation alert as required by RFC5746. Some renegotiating 1922205128Ssimon TLS clients will continue a connection gracefully when they receive 1923205128Ssimon the alert. Unfortunately OpenSSL mishandled this alert and would hang 1924205128Ssimon waiting for a server hello which it will never receive. Now we treat a 1925205128Ssimon received no_renegotiation alert as a fatal error. This is because 1926205128Ssimon applications requesting a renegotiation might well expect it to succeed 1927205128Ssimon and would have no code in place to handle the server denying it so the 1928205128Ssimon only safe thing to do is to terminate the connection. 1929205128Ssimon [Steve Henson] 1930205128Ssimon 1931205128Ssimon *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if 1932205128Ssimon peer supports secure renegotiation and 0 otherwise. Print out peer 1933205128Ssimon renegotiation support in s_client/s_server. 1934205128Ssimon [Steve Henson] 1935205128Ssimon 1936205128Ssimon *) Replace the highly broken and deprecated SPKAC certification method with 1937205128Ssimon the updated NID creation version. This should correctly handle UTF8. 1938205128Ssimon [Steve Henson] 1939205128Ssimon 1940205128Ssimon *) Implement RFC5746. Re-enable renegotiation but require the extension 1941205128Ssimon as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION 1942205128Ssimon turns out to be a bad idea. It has been replaced by 1943205128Ssimon SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with 1944205128Ssimon SSL_CTX_set_options(). This is really not recommended unless you 1945205128Ssimon know what you are doing. 1946205128Ssimon [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson] 1947205128Ssimon 1948205128Ssimon *) Fixes to stateless session resumption handling. Use initial_ctx when 1949205128Ssimon issuing and attempting to decrypt tickets in case it has changed during 1950205128Ssimon servername handling. Use a non-zero length session ID when attempting 1951205128Ssimon stateless session resumption: this makes it possible to determine if 1952205128Ssimon a resumption has occurred immediately after receiving server hello 1953205128Ssimon (several places in OpenSSL subtly assume this) instead of later in 1954205128Ssimon the handshake. 1955205128Ssimon [Steve Henson] 1956205128Ssimon 1957205128Ssimon *) The functions ENGINE_ctrl(), OPENSSL_isservice(), 1958205128Ssimon CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error 1959205128Ssimon fixes for a few places where the return code is not checked 1960205128Ssimon correctly. 1961205128Ssimon [Julia Lawall <julia@diku.dk>] 1962205128Ssimon 1963205128Ssimon *) Add --strict-warnings option to Configure script to include devteam 1964205128Ssimon warnings in other configurations. 1965205128Ssimon [Steve Henson] 1966205128Ssimon 1967205128Ssimon *) Add support for --libdir option and LIBDIR variable in makefiles. This 1968205128Ssimon makes it possible to install openssl libraries in locations which 1969205128Ssimon have names other than "lib", for example "/usr/lib64" which some 1970205128Ssimon systems need. 1971205128Ssimon [Steve Henson, based on patch from Jeremy Utley] 1972205128Ssimon 1973205128Ssimon *) Don't allow the use of leading 0x80 in OIDs. This is a violation of 1974205128Ssimon X690 8.9.12 and can produce some misleading textual output of OIDs. 1975205128Ssimon [Steve Henson, reported by Dan Kaminsky] 1976205128Ssimon 1977205128Ssimon *) Delete MD2 from algorithm tables. This follows the recommendation in 1978205128Ssimon several standards that it is not used in new applications due to 1979205128Ssimon several cryptographic weaknesses. For binary compatibility reasons 1980205128Ssimon the MD2 API is still compiled in by default. 1981205128Ssimon [Steve Henson] 1982205128Ssimon 1983205128Ssimon *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved 1984205128Ssimon and restored. 1985205128Ssimon [Steve Henson] 1986205128Ssimon 1987205128Ssimon *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and 1988205128Ssimon OPENSSL_asc2uni conditionally on Netware platforms to avoid a name 1989205128Ssimon clash. 1990205128Ssimon [Guenter <lists@gknw.net>] 1991205128Ssimon 1992205128Ssimon *) Fix the server certificate chain building code to use X509_verify_cert(), 1993205128Ssimon it used to have an ad-hoc builder which was unable to cope with anything 1994205128Ssimon other than a simple chain. 1995205128Ssimon [David Woodhouse <dwmw2@infradead.org>, Steve Henson] 1996205128Ssimon 1997205128Ssimon *) Don't check self signed certificate signatures in X509_verify_cert() 1998205128Ssimon by default (a flag can override this): it just wastes time without 1999205128Ssimon adding any security. As a useful side effect self signed root CAs 2000205128Ssimon with non-FIPS digests are now usable in FIPS mode. 2001205128Ssimon [Steve Henson] 2002205128Ssimon 2003205128Ssimon *) In dtls1_process_out_of_seq_message() the check if the current message 2004205128Ssimon is already buffered was missing. For every new message was memory 2005205128Ssimon allocated, allowing an attacker to perform an denial of service attack 2006205128Ssimon with sending out of seq handshake messages until there is no memory 2007205128Ssimon left. Additionally every future messege was buffered, even if the 2008205128Ssimon sequence number made no sense and would be part of another handshake. 2009205128Ssimon So only messages with sequence numbers less than 10 in advance will be 2010205128Ssimon buffered. (CVE-2009-1378) 2011205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 2012205128Ssimon 2013205128Ssimon *) Records are buffered if they arrive with a future epoch to be 2014205128Ssimon processed after finishing the corresponding handshake. There is 2015205128Ssimon currently no limitation to this buffer allowing an attacker to perform 2016205128Ssimon a DOS attack with sending records with future epochs until there is no 2017205128Ssimon memory left. This patch adds the pqueue_size() function to detemine 2018205128Ssimon the size of a buffer and limits the record buffer to 100 entries. 2019205128Ssimon (CVE-2009-1377) 2020205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 2021205128Ssimon 2022205128Ssimon *) Keep a copy of frag->msg_header.frag_len so it can be used after the 2023205128Ssimon parent structure is freed. (CVE-2009-1379) 2024205128Ssimon [Daniel Mentz] 2025205128Ssimon 2026205128Ssimon *) Handle non-blocking I/O properly in SSL_shutdown() call. 2027205128Ssimon [Darryl Miles <darryl-mailinglists@netbauds.net>] 2028205128Ssimon 2029205128Ssimon *) Add 2.5.4.* OIDs 2030205128Ssimon [Ilya O. <vrghost@gmail.com>] 2031205128Ssimon 2032205128Ssimon Changes between 0.9.8k and 0.9.8l [5 Nov 2009] 2033205128Ssimon 2034205128Ssimon *) Disable renegotiation completely - this fixes a severe security 2035205128Ssimon problem (CVE-2009-3555) at the cost of breaking all 2036205128Ssimon renegotiation. Renegotiation can be re-enabled by setting 2037205128Ssimon SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at 2038205128Ssimon run-time. This is really not recommended unless you know what 2039205128Ssimon you're doing. 2040205128Ssimon [Ben Laurie] 2041205128Ssimon 2042194206Ssimon Changes between 0.9.8j and 0.9.8k [25 Mar 2009] 2043167612Ssimon 2044194206Ssimon *) Don't set val to NULL when freeing up structures, it is freed up by 2045194206Ssimon underlying code. If sizeof(void *) > sizeof(long) this can result in 2046194206Ssimon zeroing past the valid field. (CVE-2009-0789) 2047194206Ssimon [Paolo Ganci <Paolo.Ganci@AdNovum.CH>] 2048194206Ssimon 2049194206Ssimon *) Fix bug where return value of CMS_SignerInfo_verify_content() was not 2050194206Ssimon checked correctly. This would allow some invalid signed attributes to 2051194206Ssimon appear to verify correctly. (CVE-2009-0591) 2052194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 2053194206Ssimon 2054194206Ssimon *) Reject UniversalString and BMPString types with invalid lengths. This 2055194206Ssimon prevents a crash in ASN1_STRING_print_ex() which assumes the strings have 2056194206Ssimon a legal length. (CVE-2009-0590) 2057194206Ssimon [Steve Henson] 2058194206Ssimon 2059194206Ssimon *) Set S/MIME signing as the default purpose rather than setting it 2060194206Ssimon unconditionally. This allows applications to override it at the store 2061194206Ssimon level. 2062194206Ssimon [Steve Henson] 2063194206Ssimon 2064194206Ssimon *) Permit restricted recursion of ASN1 strings. This is needed in practice 2065194206Ssimon to handle some structures. 2066194206Ssimon [Steve Henson] 2067194206Ssimon 2068194206Ssimon *) Improve efficiency of mem_gets: don't search whole buffer each time 2069194206Ssimon for a '\n' 2070194206Ssimon [Jeremy Shapiro <jnshapir@us.ibm.com>] 2071194206Ssimon 2072194206Ssimon *) New -hex option for openssl rand. 2073194206Ssimon [Matthieu Herrb] 2074194206Ssimon 2075194206Ssimon *) Print out UTF8String and NumericString when parsing ASN1. 2076194206Ssimon [Steve Henson] 2077194206Ssimon 2078194206Ssimon *) Support NumericString type for name components. 2079194206Ssimon [Steve Henson] 2080194206Ssimon 2081194206Ssimon *) Allow CC in the environment to override the automatically chosen 2082194206Ssimon compiler. Note that nothing is done to ensure flags work with the 2083194206Ssimon chosen compiler. 2084194206Ssimon [Ben Laurie] 2085194206Ssimon 2086194206Ssimon Changes between 0.9.8i and 0.9.8j [07 Jan 2009] 2087194206Ssimon 2088194206Ssimon *) Properly check EVP_VerifyFinal() and similar return values 2089194206Ssimon (CVE-2008-5077). 2090194206Ssimon [Ben Laurie, Bodo Moeller, Google Security Team] 2091194206Ssimon 2092194206Ssimon *) Enable TLS extensions by default. 2093194206Ssimon [Ben Laurie] 2094194206Ssimon 2095194206Ssimon *) Allow the CHIL engine to be loaded, whether the application is 2096194206Ssimon multithreaded or not. (This does not release the developer from the 2097194206Ssimon obligation to set up the dynamic locking callbacks.) 2098194206Ssimon [Sander Temme <sander@temme.net>] 2099194206Ssimon 2100194206Ssimon *) Use correct exit code if there is an error in dgst command. 2101194206Ssimon [Steve Henson; problem pointed out by Roland Dirlewanger] 2102194206Ssimon 2103194206Ssimon *) Tweak Configure so that you need to say "experimental-jpake" to enable 2104194206Ssimon JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. 2105194206Ssimon [Bodo Moeller] 2106194206Ssimon 2107194206Ssimon *) Add experimental JPAKE support, including demo authentication in 2108194206Ssimon s_client and s_server. 2109194206Ssimon [Ben Laurie] 2110194206Ssimon 2111194206Ssimon *) Set the comparison function in v3_addr_canonize(). 2112194206Ssimon [Rob Austein <sra@hactrn.net>] 2113194206Ssimon 2114194206Ssimon *) Add support for XMPP STARTTLS in s_client. 2115194206Ssimon [Philip Paeps <philip@freebsd.org>] 2116194206Ssimon 2117194206Ssimon *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior 2118194206Ssimon to ensure that even with this option, only ciphersuites in the 2119194206Ssimon server's preference list will be accepted. (Note that the option 2120194206Ssimon applies only when resuming a session, so the earlier behavior was 2121194206Ssimon just about the algorithm choice for symmetric cryptography.) 2122194206Ssimon [Bodo Moeller] 2123194206Ssimon 2124194206Ssimon Changes between 0.9.8h and 0.9.8i [15 Sep 2008] 2125194206Ssimon 2126205128Ssimon *) Fix NULL pointer dereference if a DTLS server received 2127205128Ssimon ChangeCipherSpec as first record (CVE-2009-1386). 2128205128Ssimon [PR #1679] 2129205128Ssimon 2130194206Ssimon *) Fix a state transitition in s3_srvr.c and d1_srvr.c 2131194206Ssimon (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...). 2132194206Ssimon [Nagendra Modadugu] 2133194206Ssimon 2134194206Ssimon *) The fix in 0.9.8c that supposedly got rid of unsafe 2135194206Ssimon double-checked locking was incomplete for RSA blinding, 2136194206Ssimon addressing just one layer of what turns out to have been 2137194206Ssimon doubly unsafe triple-checked locking. 2138194206Ssimon 2139194206Ssimon So now fix this for real by retiring the MONT_HELPER macro 2140194206Ssimon in crypto/rsa/rsa_eay.c. 2141194206Ssimon 2142194206Ssimon [Bodo Moeller; problem pointed out by Marius Schilder] 2143194206Ssimon 2144194206Ssimon *) Various precautionary measures: 2145194206Ssimon 2146194206Ssimon - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). 2147194206Ssimon 2148194206Ssimon - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). 2149194206Ssimon (NB: This would require knowledge of the secret session ticket key 2150194206Ssimon to exploit, in which case you'd be SOL either way.) 2151194206Ssimon 2152194206Ssimon - Change bn_nist.c so that it will properly handle input BIGNUMs 2153194206Ssimon outside the expected range. 2154194206Ssimon 2155194206Ssimon - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG 2156194206Ssimon builds. 2157194206Ssimon 2158194206Ssimon [Neel Mehta, Bodo Moeller] 2159194206Ssimon 2160194206Ssimon *) Allow engines to be "soft loaded" - i.e. optionally don't die if 2161194206Ssimon the load fails. Useful for distros. 2162194206Ssimon [Ben Laurie and the FreeBSD team] 2163194206Ssimon 2164194206Ssimon *) Add support for Local Machine Keyset attribute in PKCS#12 files. 2165194206Ssimon [Steve Henson] 2166194206Ssimon 2167194206Ssimon *) Fix BN_GF2m_mod_arr() top-bit cleanup code. 2168194206Ssimon [Huang Ying] 2169194206Ssimon 2170194206Ssimon *) Expand ENGINE to support engine supplied SSL client certificate functions. 2171194206Ssimon 2172194206Ssimon This work was sponsored by Logica. 2173194206Ssimon [Steve Henson] 2174194206Ssimon 2175194206Ssimon *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows 2176194206Ssimon keystores. Support for SSL/TLS client authentication too. 2177194206Ssimon Not compiled unless enable-capieng specified to Configure. 2178194206Ssimon 2179194206Ssimon This work was sponsored by Logica. 2180194206Ssimon [Steve Henson] 2181194206Ssimon 2182194206Ssimon *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using 2183194206Ssimon ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain 2184194206Ssimon attribute creation routines such as certifcate requests and PKCS#12 2185194206Ssimon files. 2186194206Ssimon [Steve Henson] 2187194206Ssimon 2188194206Ssimon Changes between 0.9.8g and 0.9.8h [28 May 2008] 2189194206Ssimon 2190194206Ssimon *) Fix flaw if 'Server Key exchange message' is omitted from a TLS 2191194206Ssimon handshake which could lead to a cilent crash as found using the 2192194206Ssimon Codenomicon TLS test suite (CVE-2008-1672) 2193194206Ssimon [Steve Henson, Mark Cox] 2194194206Ssimon 2195194206Ssimon *) Fix double free in TLS server name extensions which could lead to 2196194206Ssimon a remote crash found by Codenomicon TLS test suite (CVE-2008-0891) 2197194206Ssimon [Joe Orton] 2198194206Ssimon 2199194206Ssimon *) Clear error queue in SSL_CTX_use_certificate_chain_file() 2200194206Ssimon 2201194206Ssimon Clear the error queue to ensure that error entries left from 2202194206Ssimon older function calls do not interfere with the correct operation. 2203194206Ssimon [Lutz Jaenicke, Erik de Castro Lopo] 2204194206Ssimon 2205194206Ssimon *) Remove root CA certificates of commercial CAs: 2206194206Ssimon 2207194206Ssimon The OpenSSL project does not recommend any specific CA and does not 2208194206Ssimon have any policy with respect to including or excluding any CA. 2209194206Ssimon Therefore it does not make any sense to ship an arbitrary selection 2210194206Ssimon of root CA certificates with the OpenSSL software. 2211194206Ssimon [Lutz Jaenicke] 2212194206Ssimon 2213194206Ssimon *) RSA OAEP patches to fix two separate invalid memory reads. 2214194206Ssimon The first one involves inputs when 'lzero' is greater than 2215194206Ssimon 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes 2216194206Ssimon before the beginning of from). The second one involves inputs where 2217194206Ssimon the 'db' section contains nothing but zeroes (there is a one-byte 2218194206Ssimon invalid read after the end of 'db'). 2219194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 2220194206Ssimon 2221194206Ssimon *) Partial backport from 0.9.9-dev: 2222194206Ssimon 2223194206Ssimon Introduce bn_mul_mont (dedicated Montgomery multiplication 2224194206Ssimon procedure) as a candidate for BIGNUM assembler implementation. 2225194206Ssimon While 0.9.9-dev uses assembler for various architectures, only 2226194206Ssimon x86_64 is available by default here in the 0.9.8 branch, and 2227194206Ssimon 32-bit x86 is available through a compile-time setting. 2228194206Ssimon 2229194206Ssimon To try the 32-bit x86 assembler implementation, use Configure 2230194206Ssimon option "enable-montasm" (which exists only for this backport). 2231194206Ssimon 2232194206Ssimon As "enable-montasm" for 32-bit x86 disclaims code stability 2233194206Ssimon anyway, in this constellation we activate additional code 2234194206Ssimon backported from 0.9.9-dev for further performance improvements, 2235194206Ssimon namely BN_from_montgomery_word. (To enable this otherwise, 2236194206Ssimon e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".) 2237194206Ssimon 2238194206Ssimon [Andy Polyakov (backport partially by Bodo Moeller)] 2239194206Ssimon 2240194206Ssimon *) Add TLS session ticket callback. This allows an application to set 2241194206Ssimon TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed 2242194206Ssimon values. This is useful for key rollover for example where several key 2243194206Ssimon sets may exist with different names. 2244194206Ssimon [Steve Henson] 2245194206Ssimon 2246194206Ssimon *) Reverse ENGINE-internal logic for caching default ENGINE handles. 2247194206Ssimon This was broken until now in 0.9.8 releases, such that the only way 2248194206Ssimon a registered ENGINE could be used (assuming it initialises 2249194206Ssimon successfully on the host) was to explicitly set it as the default 2250194206Ssimon for the relevant algorithms. This is in contradiction with 0.9.7 2251194206Ssimon behaviour and the documentation. With this fix, when an ENGINE is 2252194206Ssimon registered into a given algorithm's table of implementations, the 2253194206Ssimon 'uptodate' flag is reset so that auto-discovery will be used next 2254194206Ssimon time a new context for that algorithm attempts to select an 2255194206Ssimon implementation. 2256194206Ssimon [Ian Lister (tweaked by Geoff Thorpe)] 2257194206Ssimon 2258194206Ssimon *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 2259194206Ssimon implemention in the following ways: 2260194206Ssimon 2261194206Ssimon Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be 2262194206Ssimon hard coded. 2263194206Ssimon 2264194206Ssimon Lack of BER streaming support means one pass streaming processing is 2265194206Ssimon only supported if data is detached: setting the streaming flag is 2266194206Ssimon ignored for embedded content. 2267194206Ssimon 2268194206Ssimon CMS support is disabled by default and must be explicitly enabled 2269194206Ssimon with the enable-cms configuration option. 2270194206Ssimon [Steve Henson] 2271194206Ssimon 2272194206Ssimon *) Update the GMP engine glue to do direct copies between BIGNUM and 2273194206Ssimon mpz_t when openssl and GMP use the same limb size. Otherwise the 2274194206Ssimon existing "conversion via a text string export" trick is still used. 2275194206Ssimon [Paul Sheer <paulsheer@gmail.com>] 2276194206Ssimon 2277194206Ssimon *) Zlib compression BIO. This is a filter BIO which compressed and 2278194206Ssimon uncompresses any data passed through it. 2279194206Ssimon [Steve Henson] 2280194206Ssimon 2281194206Ssimon *) Add AES_wrap_key() and AES_unwrap_key() functions to implement 2282194206Ssimon RFC3394 compatible AES key wrapping. 2283194206Ssimon [Steve Henson] 2284194206Ssimon 2285194206Ssimon *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): 2286194206Ssimon sets string data without copying. X509_ALGOR_set0() and 2287194206Ssimon X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) 2288194206Ssimon data. Attribute function X509at_get0_data_by_OBJ(): retrieves data 2289194206Ssimon from an X509_ATTRIBUTE structure optionally checking it occurs only 2290194206Ssimon once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied 2291194206Ssimon data. 2292194206Ssimon [Steve Henson] 2293194206Ssimon 2294194206Ssimon *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() 2295194206Ssimon to get the expected BN_FLG_CONSTTIME behavior. 2296194206Ssimon [Bodo Moeller (Google)] 2297194206Ssimon 2298194206Ssimon *) Netware support: 2299194206Ssimon 2300194206Ssimon - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets 2301194206Ssimon - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) 2302194206Ssimon - added some more tests to do_tests.pl 2303194206Ssimon - fixed RunningProcess usage so that it works with newer LIBC NDKs too 2304194206Ssimon - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency 2305194206Ssimon - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, 2306194206Ssimon netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc 2307194206Ssimon - various changes to netware.pl to enable gcc-cross builds on Win32 2308194206Ssimon platform 2309194206Ssimon - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) 2310194206Ssimon - various changes to fix missing prototype warnings 2311194206Ssimon - fixed x86nasm.pl to create correct asm files for NASM COFF output 2312194206Ssimon - added AES, WHIRLPOOL and CPUID assembler code to build files 2313194206Ssimon - added missing AES assembler make rules to mk1mf.pl 2314194206Ssimon - fixed order of includes in apps/ocsp.c so that e_os.h settings apply 2315194206Ssimon [Guenter Knauf <eflash@gmx.net>] 2316194206Ssimon 2317194206Ssimon *) Implement certificate status request TLS extension defined in RFC3546. 2318194206Ssimon A client can set the appropriate parameters and receive the encoded 2319194206Ssimon OCSP response via a callback. A server can query the supplied parameters 2320194206Ssimon and set the encoded OCSP response in the callback. Add simplified examples 2321194206Ssimon to s_client and s_server. 2322194206Ssimon [Steve Henson] 2323194206Ssimon 2324194206Ssimon Changes between 0.9.8f and 0.9.8g [19 Oct 2007] 2325194206Ssimon 2326194206Ssimon *) Fix various bugs: 2327194206Ssimon + Binary incompatibility of ssl_ctx_st structure 2328194206Ssimon + DTLS interoperation with non-compliant servers 2329194206Ssimon + Don't call get_session_cb() without proposed session 2330194206Ssimon + Fix ia64 assembler code 2331194206Ssimon [Andy Polyakov, Steve Henson] 2332194206Ssimon 2333194206Ssimon Changes between 0.9.8e and 0.9.8f [11 Oct 2007] 2334194206Ssimon 2335194206Ssimon *) DTLS Handshake overhaul. There were longstanding issues with 2336194206Ssimon OpenSSL DTLS implementation, which were making it impossible for 2337194206Ssimon RFC 4347 compliant client to communicate with OpenSSL server. 2338194206Ssimon Unfortunately just fixing these incompatibilities would "cut off" 2339194206Ssimon pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e 2340194206Ssimon server keeps tolerating non RFC compliant syntax. The opposite is 2341194206Ssimon not true, 0.9.8f client can not communicate with earlier server. 2342194206Ssimon This update even addresses CVE-2007-4995. 2343194206Ssimon [Andy Polyakov] 2344194206Ssimon 2345194206Ssimon *) Changes to avoid need for function casts in OpenSSL: some compilers 2346194206Ssimon (gcc 4.2 and later) reject their use. 2347194206Ssimon [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, 2348194206Ssimon Steve Henson] 2349194206Ssimon 2350194206Ssimon *) Add RFC4507 support to OpenSSL. This includes the corrections in 2351194206Ssimon RFC4507bis. The encrypted ticket format is an encrypted encoded 2352194206Ssimon SSL_SESSION structure, that way new session features are automatically 2353194206Ssimon supported. 2354194206Ssimon 2355194206Ssimon If a client application caches session in an SSL_SESSION structure 2356194206Ssimon support is transparent because tickets are now stored in the encoded 2357194206Ssimon SSL_SESSION. 2358194206Ssimon 2359194206Ssimon The SSL_CTX structure automatically generates keys for ticket 2360194206Ssimon protection in servers so again support should be possible 2361194206Ssimon with no application modification. 2362194206Ssimon 2363194206Ssimon If a client or server wishes to disable RFC4507 support then the option 2364194206Ssimon SSL_OP_NO_TICKET can be set. 2365194206Ssimon 2366194206Ssimon Add a TLS extension debugging callback to allow the contents of any client 2367194206Ssimon or server extensions to be examined. 2368194206Ssimon 2369194206Ssimon This work was sponsored by Google. 2370194206Ssimon [Steve Henson] 2371194206Ssimon 2372194206Ssimon *) Add initial support for TLS extensions, specifically for the server_name 2373194206Ssimon extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 2374194206Ssimon have new members for a host name. The SSL data structure has an 2375194206Ssimon additional member SSL_CTX *initial_ctx so that new sessions can be 2376194206Ssimon stored in that context to allow for session resumption, even after the 2377194206Ssimon SSL has been switched to a new SSL_CTX in reaction to a client's 2378194206Ssimon server_name extension. 2379194206Ssimon 2380194206Ssimon New functions (subject to change): 2381194206Ssimon 2382194206Ssimon SSL_get_servername() 2383194206Ssimon SSL_get_servername_type() 2384194206Ssimon SSL_set_SSL_CTX() 2385194206Ssimon 2386194206Ssimon New CTRL codes and macros (subject to change): 2387194206Ssimon 2388194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 2389194206Ssimon - SSL_CTX_set_tlsext_servername_callback() 2390194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 2391194206Ssimon - SSL_CTX_set_tlsext_servername_arg() 2392194206Ssimon SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 2393194206Ssimon 2394194206Ssimon openssl s_client has a new '-servername ...' option. 2395194206Ssimon 2396194206Ssimon openssl s_server has new options '-servername_host ...', '-cert2 ...', 2397194206Ssimon '-key2 ...', '-servername_fatal' (subject to change). This allows 2398194206Ssimon testing the HostName extension for a specific single host name ('-cert' 2399194206Ssimon and '-key' remain fallbacks for handshakes without HostName 2400194206Ssimon negotiation). If the unrecogninzed_name alert has to be sent, this by 2401194206Ssimon default is a warning; it becomes fatal with the '-servername_fatal' 2402194206Ssimon option. 2403194206Ssimon 2404194206Ssimon [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson] 2405194206Ssimon 2406194206Ssimon *) Add AES and SSE2 assembly language support to VC++ build. 2407194206Ssimon [Steve Henson] 2408194206Ssimon 2409194206Ssimon *) Mitigate attack on final subtraction in Montgomery reduction. 2410194206Ssimon [Andy Polyakov] 2411194206Ssimon 2412194206Ssimon *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 2413194206Ssimon (which previously caused an internal error). 2414194206Ssimon [Bodo Moeller] 2415194206Ssimon 2416194206Ssimon *) Squeeze another 10% out of IGE mode when in != out. 2417194206Ssimon [Ben Laurie] 2418194206Ssimon 2419194206Ssimon *) AES IGE mode speedup. 2420194206Ssimon [Dean Gaudet (Google)] 2421194206Ssimon 2422194206Ssimon *) Add the Korean symmetric 128-bit cipher SEED (see 2423194206Ssimon http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and 2424194206Ssimon add SEED ciphersuites from RFC 4162: 2425194206Ssimon 2426194206Ssimon TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" 2427194206Ssimon TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" 2428194206Ssimon TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" 2429194206Ssimon TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" 2430194206Ssimon 2431194206Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 2432194206Ssimon series, SEED remains excluded from compilation unless OpenSSL 2433194206Ssimon is configured with 'enable-seed'. 2434194206Ssimon [KISA, Bodo Moeller] 2435194206Ssimon 2436194206Ssimon *) Mitigate branch prediction attacks, which can be practical if a 2437194206Ssimon single processor is shared, allowing a spy process to extract 2438194206Ssimon information. For detailed background information, see 2439194206Ssimon http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron, 2440194206Ssimon J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL 2441194206Ssimon and Necessary Software Countermeasures"). The core of the change 2442194206Ssimon are new versions BN_div_no_branch() and 2443194206Ssimon BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), 2444194206Ssimon respectively, which are slower, but avoid the security-relevant 2445194206Ssimon conditional branches. These are automatically called by BN_div() 2446194206Ssimon and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one 2447194206Ssimon of the input BIGNUMs. Also, BN_is_bit_set() has been changed to 2448194206Ssimon remove a conditional branch. 2449194206Ssimon 2450194206Ssimon BN_FLG_CONSTTIME is the new name for the previous 2451194206Ssimon BN_FLG_EXP_CONSTTIME flag, since it now affects more than just 2452194206Ssimon modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag 2453194206Ssimon in the exponent causes BN_mod_exp_mont() to use the alternative 2454194206Ssimon implementation in BN_mod_exp_mont_consttime().) The old name 2455194206Ssimon remains as a deprecated alias. 2456194206Ssimon 2457194206Ssimon Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general 2458194206Ssimon RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses 2459194206Ssimon constant-time implementations for more than just exponentiation. 2460194206Ssimon Here too the old name is kept as a deprecated alias. 2461194206Ssimon 2462194206Ssimon BN_BLINDING_new() will now use BN_dup() for the modulus so that 2463194206Ssimon the BN_BLINDING structure gets an independent copy of the 2464194206Ssimon modulus. This means that the previous "BIGNUM *m" argument to 2465194206Ssimon BN_BLINDING_new() and to BN_BLINDING_create_param() now 2466194206Ssimon essentially becomes "const BIGNUM *m", although we can't actually 2467194206Ssimon change this in the header file before 0.9.9. It allows 2468194206Ssimon RSA_setup_blinding() to use BN_with_flags() on the modulus to 2469194206Ssimon enable BN_FLG_CONSTTIME. 2470194206Ssimon 2471194206Ssimon [Matthew D Wood (Intel Corp)] 2472194206Ssimon 2473194206Ssimon *) In the SSL/TLS server implementation, be strict about session ID 2474194206Ssimon context matching (which matters if an application uses a single 2475194206Ssimon external cache for different purposes). Previously, 2476194206Ssimon out-of-context reuse was forbidden only if SSL_VERIFY_PEER was 2477194206Ssimon set. This did ensure strict client verification, but meant that, 2478194206Ssimon with applications using a single external cache for quite 2479194206Ssimon different requirements, clients could circumvent ciphersuite 2480194206Ssimon restrictions for a given session ID context by starting a session 2481194206Ssimon in a different context. 2482194206Ssimon [Bodo Moeller] 2483194206Ssimon 2484167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 2485167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 2486167612Ssimon authentication-only ciphersuites. 2487167612Ssimon [Bodo Moeller] 2488167612Ssimon 2489194206Ssimon *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was 2490194206Ssimon not complete and could lead to a possible single byte overflow 2491194206Ssimon (CVE-2007-5135) [Ben Laurie] 2492194206Ssimon 2493194206Ssimon Changes between 0.9.8d and 0.9.8e [23 Feb 2007] 2494194206Ssimon 2495167612Ssimon *) Since AES128 and AES256 (and similarly Camellia128 and 2496167612Ssimon Camellia256) share a single mask bit in the logic of 2497167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 2498167612Ssimon kludge to work properly if AES128 is available and AES256 isn't 2499167612Ssimon (or if Camellia128 is available and Camellia256 isn't). 2500167612Ssimon [Victor Duchovni] 2501167612Ssimon 2502167612Ssimon *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c 2503167612Ssimon (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): 2504167612Ssimon When a point or a seed is encoded in a BIT STRING, we need to 2505167612Ssimon prevent the removal of trailing zero bits to get the proper DER 2506167612Ssimon encoding. (By default, crypto/asn1/a_bitstr.c assumes the case 2507167612Ssimon of a NamedBitList, for which trailing 0 bits need to be removed.) 2508167612Ssimon [Bodo Moeller] 2509167612Ssimon 2510167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 2511167612Ssimon protocol version while receiving ClientHello even if the 2512167612Ssimon ClientHello is fragmented. (The server can't insist on the 2513167612Ssimon particular protocol version it has chosen before the ServerHello 2514167612Ssimon message has informed the client about his choice.) 2515167612Ssimon [Bodo Moeller] 2516167612Ssimon 2517167612Ssimon *) Add RFC 3779 support. 2518167612Ssimon [Rob Austein for ARIN, Ben Laurie] 2519167612Ssimon 2520167612Ssimon *) Load error codes if they are not already present instead of using a 2521167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 2522167612Ssimon Improve header file function name parsing. 2523167612Ssimon [Steve Henson] 2524167612Ssimon 2525167612Ssimon *) extend SMTP and IMAP protocol emulation in s_client to use EHLO 2526167612Ssimon or CAPABILITY handshake as required by RFCs. 2527167612Ssimon [Goetz Babin-Ebell] 2528167612Ssimon 2529162911Ssimon Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 2530162911Ssimon 2531162911Ssimon *) Introduce limits to prevent malicious keys being able to 2532162911Ssimon cause a denial of service. (CVE-2006-2940) 2533162911Ssimon [Steve Henson, Bodo Moeller] 2534162911Ssimon 2535162911Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 2536162911Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 2537162911Ssimon 2538162911Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 2539162911Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 2540162911Ssimon 2541162911Ssimon *) Fix SSL client code which could crash if connecting to a 2542162911Ssimon malicious SSLv2 server. (CVE-2006-4343) 2543162911Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 2544162911Ssimon 2545162911Ssimon *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 2546162911Ssimon match only those. Before that, "AES256-SHA" would be interpreted 2547162911Ssimon as a pattern and match "AES128-SHA" too (since AES128-SHA got 2548162911Ssimon the same strength classification in 0.9.7h) as we currently only 2549162911Ssimon have a single AES bit in the ciphersuite description bitmap. 2550162911Ssimon That change, however, also applied to ciphersuite strings such as 2551162911Ssimon "RC4-MD5" that intentionally matched multiple ciphersuites -- 2552162911Ssimon namely, SSL 2.0 ciphersuites in addition to the more common ones 2553162911Ssimon from SSL 3.0/TLS 1.0. 2554162911Ssimon 2555162911Ssimon So we change the selection algorithm again: Naming an explicit 2556162911Ssimon ciphersuite selects this one ciphersuite, and any other similar 2557162911Ssimon ciphersuite (same bitmap) from *other* protocol versions. 2558162911Ssimon Thus, "RC4-MD5" again will properly select both the SSL 2.0 2559162911Ssimon ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 2560162911Ssimon 2561162911Ssimon Since SSL 2.0 does not have any ciphersuites for which the 2562162911Ssimon 128/256 bit distinction would be relevant, this works for now. 2563162911Ssimon The proper fix will be to use different bits for AES128 and 2564162911Ssimon AES256, which would have avoided the problems from the beginning; 2565162911Ssimon however, bits are scarce, so we can only do this in a new release 2566162911Ssimon (not just a patchlevel) when we can change the SSL_CIPHER 2567162911Ssimon definition to split the single 'unsigned long mask' bitmap into 2568162911Ssimon multiple values to extend the available space. 2569162911Ssimon 2570162911Ssimon [Bodo Moeller] 2571162911Ssimon 2572162911Ssimon Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 2573162911Ssimon 2574162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 2575162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 2576162911Ssimon 2577162911Ssimon *) Add AES IGE and biIGE modes. 2578162911Ssimon [Ben Laurie] 2579162911Ssimon 2580162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 2581162911Ssimon possible instead of select(), since the latter has some 2582162911Ssimon undesirable limitations. 2583162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 2584162911Ssimon 2585162911Ssimon *) Disable "ECCdraft" ciphersuites more thoroughly. Now special 2586162911Ssimon treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 2587162911Ssimon cannot be implicitly activated as part of, e.g., the "AES" alias. 2588162911Ssimon However, please upgrade to OpenSSL 0.9.9[-dev] for 2589162911Ssimon non-experimental use of the ECC ciphersuites to get TLS extension 2590162911Ssimon support, which is required for curve and point format negotiation 2591162911Ssimon to avoid potential handshake problems. 2592162911Ssimon [Bodo Moeller] 2593162911Ssimon 2594162911Ssimon *) Disable rogue ciphersuites: 2595162911Ssimon 2596162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 2597162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 2598162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 2599162911Ssimon 2600162911Ssimon The latter two were purportedly from 2601162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 2602162911Ssimon appear there. 2603162911Ssimon 2604167612Ssimon Also deactivate the remaining ciphersuites from 2605162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 2606162911Ssimon unofficial, and the ID has long expired. 2607162911Ssimon [Bodo Moeller] 2608162911Ssimon 2609162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 2610162911Ssimon dual-core machines) and other potential thread-safety issues. 2611162911Ssimon [Bodo Moeller] 2612162911Ssimon 2613162911Ssimon *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 2614162911Ssimon versions), which is now available for royalty-free use 2615162911Ssimon (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). 2616162911Ssimon Also, add Camellia TLS ciphersuites from RFC 4132. 2617162911Ssimon 2618162911Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 2619162911Ssimon series, Camellia remains excluded from compilation unless OpenSSL 2620162911Ssimon is configured with 'enable-camellia'. 2621162911Ssimon [NTT] 2622162911Ssimon 2623162911Ssimon *) Disable the padding bug check when compression is in use. The padding 2624162911Ssimon bug check assumes the first packet is of even length, this is not 2625162911Ssimon necessarily true if compresssion is enabled and can result in false 2626162911Ssimon positives causing handshake failure. The actual bug test is ancient 2627162911Ssimon code so it is hoped that implementations will either have fixed it by 2628162911Ssimon now or any which still have the bug do not support compression. 2629162911Ssimon [Steve Henson] 2630162911Ssimon 2631160814Ssimon Changes between 0.9.8a and 0.9.8b [04 May 2006] 2632160814Ssimon 2633160814Ssimon *) When applying a cipher rule check to see if string match is an explicit 2634160814Ssimon cipher suite and only match that one cipher suite if it is. 2635160814Ssimon [Steve Henson] 2636160814Ssimon 2637160814Ssimon *) Link in manifests for VC++ if needed. 2638160814Ssimon [Austin Ziegler <halostatue@gmail.com>] 2639160814Ssimon 2640160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 2641160814Ssimon draft-ietf-tls-ecc-12.txt with proposed changes (but without 2642160814Ssimon TLS extensions, which are supported starting with the 0.9.9 2643160814Ssimon branch, not in the OpenSSL 0.9.8 branch). 2644160814Ssimon [Douglas Stebila] 2645160814Ssimon 2646160814Ssimon *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 2647160814Ssimon opaque EVP_CIPHER_CTX handling. 2648160814Ssimon [Steve Henson] 2649160814Ssimon 2650160814Ssimon *) Fixes and enhancements to zlib compression code. We now only use 2651160814Ssimon "zlib1.dll" and use the default __cdecl calling convention on Win32 2652160814Ssimon to conform with the standards mentioned here: 2653160814Ssimon http://www.zlib.net/DLL_FAQ.txt 2654160814Ssimon Static zlib linking now works on Windows and the new --with-zlib-include 2655160814Ssimon --with-zlib-lib options to Configure can be used to supply the location 2656160814Ssimon of the headers and library. Gracefully handle case where zlib library 2657160814Ssimon can't be loaded. 2658160814Ssimon [Steve Henson] 2659160814Ssimon 2660160814Ssimon *) Several fixes and enhancements to the OID generation code. The old code 2661160814Ssimon sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 2662160814Ssimon handle numbers larger than ULONG_MAX, truncated printing and had a 2663160814Ssimon non standard OBJ_obj2txt() behaviour. 2664160814Ssimon [Steve Henson] 2665160814Ssimon 2666160814Ssimon *) Add support for building of engines under engine/ as shared libraries 2667160814Ssimon under VC++ build system. 2668160814Ssimon [Steve Henson] 2669160814Ssimon 2670160814Ssimon *) Corrected the numerous bugs in the Win32 path splitter in DSO. 2671160814Ssimon Hopefully, we will not see any false combination of paths any more. 2672160814Ssimon [Richard Levitte] 2673160814Ssimon 2674160814Ssimon Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 2675160814Ssimon 2676160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 2677160814Ssimon (part of SSL_OP_ALL). This option used to disable the 2678160814Ssimon countermeasure against man-in-the-middle protocol-version 2679160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 2680160814Ssimon idea. (CVE-2005-2969) 2681160814Ssimon 2682160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 2683160814Ssimon for Information Security, National Institute of Advanced Industrial 2684160814Ssimon Science and Technology [AIST], Japan)] 2685160814Ssimon 2686160814Ssimon *) Add two function to clear and return the verify parameter flags. 2687160814Ssimon [Steve Henson] 2688160814Ssimon 2689160814Ssimon *) Keep cipherlists sorted in the source instead of sorting them at 2690160814Ssimon runtime, thus removing the need for a lock. 2691160814Ssimon [Nils Larsch] 2692160814Ssimon 2693160814Ssimon *) Avoid some small subgroup attacks in Diffie-Hellman. 2694160814Ssimon [Nick Mathewson and Ben Laurie] 2695160814Ssimon 2696160814Ssimon *) Add functions for well-known primes. 2697160814Ssimon [Nick Mathewson] 2698160814Ssimon 2699160814Ssimon *) Extended Windows CE support. 2700160814Ssimon [Satoshi Nakamura and Andy Polyakov] 2701160814Ssimon 2702160814Ssimon *) Initialize SSL_METHOD structures at compile time instead of during 2703160814Ssimon runtime, thus removing the need for a lock. 2704160814Ssimon [Steve Henson] 2705160814Ssimon 2706160814Ssimon *) Make PKCS7_decrypt() work even if no certificate is supplied by 2707160814Ssimon attempting to decrypt each encrypted key in turn. Add support to 2708160814Ssimon smime utility. 2709160814Ssimon [Steve Henson] 2710160814Ssimon 2711160814Ssimon Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 2712160814Ssimon 2713162911Ssimon [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 2714162911Ssimon OpenSSL 0.9.8.] 2715162911Ssimon 2716160814Ssimon *) Add libcrypto.pc and libssl.pc for those who feel they need them. 2717160814Ssimon [Richard Levitte] 2718160814Ssimon 2719160814Ssimon *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 2720160814Ssimon key into the same file any more. 2721160814Ssimon [Richard Levitte] 2722160814Ssimon 2723160814Ssimon *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 2724160814Ssimon [Andy Polyakov] 2725160814Ssimon 2726160814Ssimon *) Add -utf8 command line and config file option to 'ca'. 2727160814Ssimon [Stefan <stf@udoma.org] 2728160814Ssimon 2729160814Ssimon *) Removed the macro des_crypt(), as it seems to conflict with some 2730160814Ssimon libraries. Use DES_crypt(). 2731160814Ssimon [Richard Levitte] 2732160814Ssimon 2733160814Ssimon *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 2734160814Ssimon involves renaming the source and generated shared-libs for 2735160814Ssimon both. The engines will accept the corrected or legacy ids 2736160814Ssimon ('ncipher' and '4758_cca' respectively) when binding. NB, 2737160814Ssimon this only applies when building 'shared'. 2738160814Ssimon [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 2739160814Ssimon 2740160814Ssimon *) Add attribute functions to EVP_PKEY structure. Modify 2741160814Ssimon PKCS12_create() to recognize a CSP name attribute and 2742160814Ssimon use it. Make -CSP option work again in pkcs12 utility. 2743160814Ssimon [Steve Henson] 2744160814Ssimon 2745160814Ssimon *) Add new functionality to the bn blinding code: 2746160814Ssimon - automatic re-creation of the BN_BLINDING parameters after 2747160814Ssimon a fixed number of uses (currently 32) 2748160814Ssimon - add new function for parameter creation 2749160814Ssimon - introduce flags to control the update behaviour of the 2750160814Ssimon BN_BLINDING parameters 2751160814Ssimon - hide BN_BLINDING structure 2752160814Ssimon Add a second BN_BLINDING slot to the RSA structure to improve 2753160814Ssimon performance when a single RSA object is shared among several 2754160814Ssimon threads. 2755160814Ssimon [Nils Larsch] 2756160814Ssimon 2757160814Ssimon *) Add support for DTLS. 2758160814Ssimon [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 2759160814Ssimon 2760160814Ssimon *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 2761160814Ssimon to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 2762160814Ssimon [Walter Goulet] 2763160814Ssimon 2764160814Ssimon *) Remove buggy and incompletet DH cert support from 2765160814Ssimon ssl/ssl_rsa.c and ssl/s3_both.c 2766160814Ssimon [Nils Larsch] 2767160814Ssimon 2768160814Ssimon *) Use SHA-1 instead of MD5 as the default digest algorithm for 2769160814Ssimon the apps/openssl applications. 2770160814Ssimon [Nils Larsch] 2771160814Ssimon 2772160814Ssimon *) Compile clean with "-Wall -Wmissing-prototypes 2773160814Ssimon -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 2774160814Ssimon DEBUG_SAFESTACK must also be set. 2775160814Ssimon [Ben Laurie] 2776160814Ssimon 2777160814Ssimon *) Change ./Configure so that certain algorithms can be disabled by default. 2778160814Ssimon The new counterpiece to "no-xxx" is "enable-xxx". 2779160814Ssimon 2780160814Ssimon The patented RC5 and MDC2 algorithms will now be disabled unless 2781160814Ssimon "enable-rc5" and "enable-mdc2", respectively, are specified. 2782160814Ssimon 2783160814Ssimon (IDEA remains enabled despite being patented. This is because IDEA 2784160814Ssimon is frequently required for interoperability, and there is no license 2785160814Ssimon fee for non-commercial use. As before, "no-idea" can be used to 2786160814Ssimon avoid this algorithm.) 2787160814Ssimon 2788160814Ssimon [Bodo Moeller] 2789160814Ssimon 2790160814Ssimon *) Add processing of proxy certificates (see RFC 3820). This work was 2791160814Ssimon sponsored by KTH (The Royal Institute of Technology in Stockholm) and 2792160814Ssimon EGEE (Enabling Grids for E-science in Europe). 2793160814Ssimon [Richard Levitte] 2794160814Ssimon 2795160814Ssimon *) RC4 performance overhaul on modern architectures/implementations, such 2796160814Ssimon as Intel P4, IA-64 and AMD64. 2797160814Ssimon [Andy Polyakov] 2798160814Ssimon 2799160814Ssimon *) New utility extract-section.pl. This can be used specify an alternative 2800160814Ssimon section number in a pod file instead of having to treat each file as 2801160814Ssimon a separate case in Makefile. This can be done by adding two lines to the 2802160814Ssimon pod file: 2803160814Ssimon 2804160814Ssimon =for comment openssl_section:XXX 2805160814Ssimon 2806160814Ssimon The blank line is mandatory. 2807160814Ssimon 2808160814Ssimon [Steve Henson] 2809160814Ssimon 2810160814Ssimon *) New arguments -certform, -keyform and -pass for s_client and s_server 2811160814Ssimon to allow alternative format key and certificate files and passphrase 2812160814Ssimon sources. 2813160814Ssimon [Steve Henson] 2814160814Ssimon 2815160814Ssimon *) New structure X509_VERIFY_PARAM which combines current verify parameters, 2816160814Ssimon update associated structures and add various utility functions. 2817160814Ssimon 2818160814Ssimon Add new policy related verify parameters, include policy checking in 2819160814Ssimon standard verify code. Enhance 'smime' application with extra parameters 2820160814Ssimon to support policy checking and print out. 2821160814Ssimon [Steve Henson] 2822160814Ssimon 2823160814Ssimon *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 2824160814Ssimon Nehemiah processors. These extensions support AES encryption in hardware 2825160814Ssimon as well as RNG (though RNG support is currently disabled). 2826160814Ssimon [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 2827160814Ssimon 2828160814Ssimon *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 2829160814Ssimon [Geoff Thorpe] 2830160814Ssimon 2831160814Ssimon *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 2832160814Ssimon [Andy Polyakov and a number of other people] 2833160814Ssimon 2834160814Ssimon *) Improved PowerPC platform support. Most notably BIGNUM assembler 2835160814Ssimon implementation contributed by IBM. 2836160814Ssimon [Suresh Chari, Peter Waltenberg, Andy Polyakov] 2837160814Ssimon 2838160814Ssimon *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 2839160814Ssimon exponent rather than 'unsigned long'. There is a corresponding change to 2840160814Ssimon the new 'rsa_keygen' element of the RSA_METHOD structure. 2841160814Ssimon [Jelte Jansen, Geoff Thorpe] 2842160814Ssimon 2843160814Ssimon *) Functionality for creating the initial serial number file is now 2844160814Ssimon moved from CA.pl to the 'ca' utility with a new option -create_serial. 2845160814Ssimon 2846160814Ssimon (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 2847160814Ssimon number file to 1, which is bound to cause problems. To avoid 2848160814Ssimon the problems while respecting compatibility between different 0.9.7 2849160814Ssimon patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 2850160814Ssimon CA.pl for serial number initialization. With the new release 0.9.8, 2851160814Ssimon we can fix the problem directly in the 'ca' utility.) 2852160814Ssimon [Steve Henson] 2853160814Ssimon 2854160814Ssimon *) Reduced header interdepencies by declaring more opaque objects in 2855160814Ssimon ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 2856160814Ssimon give fewer recursive includes, which could break lazy source code - so 2857160814Ssimon this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 2858160814Ssimon developers should define this symbol when building and using openssl to 2859160814Ssimon ensure they track the recommended behaviour, interfaces, [etc], but 2860160814Ssimon backwards-compatible behaviour prevails when this isn't defined. 2861160814Ssimon [Geoff Thorpe] 2862160814Ssimon 2863160814Ssimon *) New function X509_POLICY_NODE_print() which prints out policy nodes. 2864160814Ssimon [Steve Henson] 2865160814Ssimon 2866160814Ssimon *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 2867160814Ssimon This will generate a random key of the appropriate length based on the 2868160814Ssimon cipher context. The EVP_CIPHER can provide its own random key generation 2869160814Ssimon routine to support keys of a specific form. This is used in the des and 2870160814Ssimon 3des routines to generate a key of the correct parity. Update S/MIME 2871160814Ssimon code to use new functions and hence generate correct parity DES keys. 2872160814Ssimon Add EVP_CHECK_DES_KEY #define to return an error if the key is not 2873160814Ssimon valid (weak or incorrect parity). 2874160814Ssimon [Steve Henson] 2875160814Ssimon 2876160814Ssimon *) Add a local set of CRLs that can be used by X509_verify_cert() as well 2877160814Ssimon as looking them up. This is useful when the verified structure may contain 2878160814Ssimon CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 2879160814Ssimon present unless the new PKCS7_NO_CRL flag is asserted. 2880160814Ssimon [Steve Henson] 2881160814Ssimon 2882160814Ssimon *) Extend ASN1 oid configuration module. It now additionally accepts the 2883160814Ssimon syntax: 2884160814Ssimon 2885160814Ssimon shortName = some long name, 1.2.3.4 2886160814Ssimon [Steve Henson] 2887160814Ssimon 2888160814Ssimon *) Reimplemented the BN_CTX implementation. There is now no more static 2889160814Ssimon limitation on the number of variables it can handle nor the depth of the 2890160814Ssimon "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 2891160814Ssimon information can now expand as required, and rather than having a single 2892160814Ssimon static array of bignums, BN_CTX now uses a linked-list of such arrays 2893160814Ssimon allowing it to expand on demand whilst maintaining the usefulness of 2894160814Ssimon BN_CTX's "bundling". 2895160814Ssimon [Geoff Thorpe] 2896160814Ssimon 2897160814Ssimon *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 2898160814Ssimon to allow all RSA operations to function using a single BN_CTX. 2899160814Ssimon [Geoff Thorpe] 2900160814Ssimon 2901160814Ssimon *) Preliminary support for certificate policy evaluation and checking. This 2902160814Ssimon is initially intended to pass the tests outlined in "Conformance Testing 2903160814Ssimon of Relying Party Client Certificate Path Processing Logic" v1.07. 2904160814Ssimon [Steve Henson] 2905160814Ssimon 2906160814Ssimon *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 2907160814Ssimon remained unused and not that useful. A variety of other little bignum 2908160814Ssimon tweaks and fixes have also been made continuing on from the audit (see 2909160814Ssimon below). 2910160814Ssimon [Geoff Thorpe] 2911160814Ssimon 2912160814Ssimon *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 2913160814Ssimon associated ASN1, EVP and SSL functions and old ASN1 macros. 2914160814Ssimon [Richard Levitte] 2915160814Ssimon 2916160814Ssimon *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 2917160814Ssimon and this should never fail. So the return value from the use of 2918160814Ssimon BN_set_word() (which can fail due to needless expansion) is now deprecated; 2919160814Ssimon if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 2920160814Ssimon [Geoff Thorpe] 2921160814Ssimon 2922160814Ssimon *) BN_CTX_get() should return zero-valued bignums, providing the same 2923160814Ssimon initialised value as BN_new(). 2924160814Ssimon [Geoff Thorpe, suggested by Ulf M�ller] 2925160814Ssimon 2926160814Ssimon *) Support for inhibitAnyPolicy certificate extension. 2927160814Ssimon [Steve Henson] 2928160814Ssimon 2929160814Ssimon *) An audit of the BIGNUM code is underway, for which debugging code is 2930160814Ssimon enabled when BN_DEBUG is defined. This makes stricter enforcements on what 2931160814Ssimon is considered valid when processing BIGNUMs, and causes execution to 2932160814Ssimon assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 2933160814Ssimon further steps are taken to deliberately pollute unused data in BIGNUM 2934160814Ssimon structures to try and expose faulty code further on. For now, openssl will 2935160814Ssimon (in its default mode of operation) continue to tolerate the inconsistent 2936160814Ssimon forms that it has tolerated in the past, but authors and packagers should 2937160814Ssimon consider trying openssl and their own applications when compiled with 2938160814Ssimon these debugging symbols defined. It will help highlight potential bugs in 2939160814Ssimon their own code, and will improve the test coverage for OpenSSL itself. At 2940160814Ssimon some point, these tighter rules will become openssl's default to improve 2941160814Ssimon maintainability, though the assert()s and other overheads will remain only 2942160814Ssimon in debugging configurations. See bn.h for more details. 2943160814Ssimon [Geoff Thorpe, Nils Larsch, Ulf M�ller] 2944160814Ssimon 2945160814Ssimon *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 2946160814Ssimon that can only be obtained through BN_CTX_new() (which implicitly 2947160814Ssimon initialises it). The presence of this function only made it possible 2948160814Ssimon to overwrite an existing structure (and cause memory leaks). 2949160814Ssimon [Geoff Thorpe] 2950160814Ssimon 2951160814Ssimon *) Because of the callback-based approach for implementing LHASH as a 2952160814Ssimon template type, lh_insert() adds opaque objects to hash-tables and 2953160814Ssimon lh_doall() or lh_doall_arg() are typically used with a destructor callback 2954160814Ssimon to clean up those corresponding objects before destroying the hash table 2955160814Ssimon (and losing the object pointers). So some over-zealous constifications in 2956160814Ssimon LHASH have been relaxed so that lh_insert() does not take (nor store) the 2957160814Ssimon objects as "const" and the lh_doall[_arg] callback wrappers are not 2958160814Ssimon prototyped to have "const" restrictions on the object pointers they are 2959160814Ssimon given (and so aren't required to cast them away any more). 2960160814Ssimon [Geoff Thorpe] 2961160814Ssimon 2962160814Ssimon *) The tmdiff.h API was so ugly and minimal that our own timing utility 2963160814Ssimon (speed) prefers to use its own implementation. The two implementations 2964160814Ssimon haven't been consolidated as yet (volunteers?) but the tmdiff API has had 2965160814Ssimon its object type properly exposed (MS_TM) instead of casting to/from "char 2966160814Ssimon *". This may still change yet if someone realises MS_TM and "ms_time_***" 2967160814Ssimon aren't necessarily the greatest nomenclatures - but this is what was used 2968160814Ssimon internally to the implementation so I've used that for now. 2969160814Ssimon [Geoff Thorpe] 2970160814Ssimon 2971160814Ssimon *) Ensure that deprecated functions do not get compiled when 2972160814Ssimon OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 2973160814Ssimon the self-tests were still using deprecated key-generation functions so 2974160814Ssimon these have been updated also. 2975160814Ssimon [Geoff Thorpe] 2976160814Ssimon 2977160814Ssimon *) Reorganise PKCS#7 code to separate the digest location functionality 2978160814Ssimon into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 2979160814Ssimon New function PKCS7_set_digest() to set the digest type for PKCS#7 2980160814Ssimon digestedData type. Add additional code to correctly generate the 2981160814Ssimon digestedData type and add support for this type in PKCS7 initialization 2982160814Ssimon functions. 2983160814Ssimon [Steve Henson] 2984160814Ssimon 2985160814Ssimon *) New function PKCS7_set0_type_other() this initializes a PKCS7 2986160814Ssimon structure of type "other". 2987160814Ssimon [Steve Henson] 2988160814Ssimon 2989160814Ssimon *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 2990160814Ssimon sure the loop does correctly stop and breaking ("division by zero") 2991160814Ssimon modulus operations are not performed. The (pre-generated) prime 2992160814Ssimon table crypto/bn/bn_prime.h was already correct, but it could not be 2993160814Ssimon re-generated on some platforms because of the "division by zero" 2994160814Ssimon situation in the script. 2995160814Ssimon [Ralf S. Engelschall] 2996160814Ssimon 2997160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 2998160814Ssimon draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 2999160814Ssimon SHA-1 now is only used for "small" curves (where the 3000160814Ssimon representation of a field element takes up to 24 bytes); for 3001160814Ssimon larger curves, the field element resulting from ECDH is directly 3002160814Ssimon used as premaster secret. 3003160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 3004160814Ssimon 3005160814Ssimon *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 3006160814Ssimon curve secp160r1 to the tests. 3007160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 3008160814Ssimon 3009160814Ssimon *) Add the possibility to load symbols globally with DSO. 3010160814Ssimon [G�tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 3011160814Ssimon 3012160814Ssimon *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 3013160814Ssimon control of the error stack. 3014160814Ssimon [Richard Levitte] 3015160814Ssimon 3016160814Ssimon *) Add support for STORE in ENGINE. 3017160814Ssimon [Richard Levitte] 3018160814Ssimon 3019160814Ssimon *) Add the STORE type. The intention is to provide a common interface 3020160814Ssimon to certificate and key stores, be they simple file-based stores, or 3021160814Ssimon HSM-type store, or LDAP stores, or... 3022160814Ssimon NOTE: The code is currently UNTESTED and isn't really used anywhere. 3023160814Ssimon [Richard Levitte] 3024160814Ssimon 3025160814Ssimon *) Add a generic structure called OPENSSL_ITEM. This can be used to 3026160814Ssimon pass a list of arguments to any function as well as provide a way 3027160814Ssimon for a function to pass data back to the caller. 3028160814Ssimon [Richard Levitte] 3029160814Ssimon 3030160814Ssimon *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 3031160814Ssimon works like BUF_strdup() but can be used to duplicate a portion of 3032160814Ssimon a string. The copy gets NUL-terminated. BUF_memdup() duplicates 3033160814Ssimon a memory area. 3034160814Ssimon [Richard Levitte] 3035160814Ssimon 3036160814Ssimon *) Add the function sk_find_ex() which works like sk_find(), but will 3037160814Ssimon return an index to an element even if an exact match couldn't be 3038160814Ssimon found. The index is guaranteed to point at the element where the 3039160814Ssimon searched-for key would be inserted to preserve sorting order. 3040160814Ssimon [Richard Levitte] 3041160814Ssimon 3042160814Ssimon *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 3043160814Ssimon takes an extra flags argument for optional functionality. Currently, 3044160814Ssimon the following flags are defined: 3045160814Ssimon 3046160814Ssimon OBJ_BSEARCH_VALUE_ON_NOMATCH 3047160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 3048160814Ssimon element where the comparing function returns a negative or zero 3049160814Ssimon number. 3050160814Ssimon 3051160814Ssimon OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 3052160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 3053160814Ssimon element where the comparing function returns zero. This is useful 3054160814Ssimon if there are more than one element where the comparing function 3055160814Ssimon returns zero. 3056160814Ssimon [Richard Levitte] 3057160814Ssimon 3058160814Ssimon *) Make it possible to create self-signed certificates with 'openssl ca' 3059160814Ssimon in such a way that the self-signed certificate becomes part of the 3060160814Ssimon CA database and uses the same mechanisms for serial number generation 3061160814Ssimon as all other certificate signing. The new flag '-selfsign' enables 3062160814Ssimon this functionality. Adapt CA.sh and CA.pl.in. 3063160814Ssimon [Richard Levitte] 3064160814Ssimon 3065160814Ssimon *) Add functionality to check the public key of a certificate request 3066160814Ssimon against a given private. This is useful to check that a certificate 3067160814Ssimon request can be signed by that key (self-signing). 3068160814Ssimon [Richard Levitte] 3069160814Ssimon 3070160814Ssimon *) Make it possible to have multiple active certificates with the same 3071160814Ssimon subject in the CA index file. This is done only if the keyword 3072160814Ssimon 'unique_subject' is set to 'no' in the main CA section (default 3073160814Ssimon if 'CA_default') of the configuration file. The value is saved 3074160814Ssimon with the database itself in a separate index attribute file, 3075160814Ssimon named like the index file with '.attr' appended to the name. 3076160814Ssimon [Richard Levitte] 3077160814Ssimon 3078160814Ssimon *) Generate muti valued AVAs using '+' notation in config files for 3079160814Ssimon req and dirName. 3080160814Ssimon [Steve Henson] 3081160814Ssimon 3082160814Ssimon *) Support for nameConstraints certificate extension. 3083160814Ssimon [Steve Henson] 3084160814Ssimon 3085160814Ssimon *) Support for policyConstraints certificate extension. 3086160814Ssimon [Steve Henson] 3087160814Ssimon 3088160814Ssimon *) Support for policyMappings certificate extension. 3089160814Ssimon [Steve Henson] 3090160814Ssimon 3091160814Ssimon *) Make sure the default DSA_METHOD implementation only uses its 3092160814Ssimon dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 3093160814Ssimon and change its own handlers to be NULL so as to remove unnecessary 3094160814Ssimon indirection. This lets alternative implementations fallback to the 3095160814Ssimon default implementation more easily. 3096160814Ssimon [Geoff Thorpe] 3097160814Ssimon 3098160814Ssimon *) Support for directoryName in GeneralName related extensions 3099160814Ssimon in config files. 3100160814Ssimon [Steve Henson] 3101160814Ssimon 3102160814Ssimon *) Make it possible to link applications using Makefile.shared. 3103160814Ssimon Make that possible even when linking against static libraries! 3104160814Ssimon [Richard Levitte] 3105160814Ssimon 3106160814Ssimon *) Support for single pass processing for S/MIME signing. This now 3107160814Ssimon means that S/MIME signing can be done from a pipe, in addition 3108160814Ssimon cleartext signing (multipart/signed type) is effectively streaming 3109160814Ssimon and the signed data does not need to be all held in memory. 3110160814Ssimon 3111160814Ssimon This is done with a new flag PKCS7_STREAM. When this flag is set 3112160814Ssimon PKCS7_sign() only initializes the PKCS7 structure and the actual signing 3113160814Ssimon is done after the data is output (and digests calculated) in 3114160814Ssimon SMIME_write_PKCS7(). 3115160814Ssimon [Steve Henson] 3116160814Ssimon 3117160814Ssimon *) Add full support for -rpath/-R, both in shared libraries and 3118160814Ssimon applications, at least on the platforms where it's known how 3119160814Ssimon to do it. 3120160814Ssimon [Richard Levitte] 3121160814Ssimon 3122160814Ssimon *) In crypto/ec/ec_mult.c, implement fast point multiplication with 3123160814Ssimon precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 3124160814Ssimon will now compute a table of multiples of the generator that 3125160814Ssimon makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 3126160814Ssimon faster (notably in the case of a single point multiplication, 3127160814Ssimon scalar * generator). 3128160814Ssimon [Nils Larsch, Bodo Moeller] 3129160814Ssimon 3130160814Ssimon *) IPv6 support for certificate extensions. The various extensions 3131160814Ssimon which use the IP:a.b.c.d can now take IPv6 addresses using the 3132160814Ssimon formats of RFC1884 2.2 . IPv6 addresses are now also displayed 3133160814Ssimon correctly. 3134160814Ssimon [Steve Henson] 3135160814Ssimon 3136160814Ssimon *) Added an ENGINE that implements RSA by performing private key 3137160814Ssimon exponentiations with the GMP library. The conversions to and from 3138160814Ssimon GMP's mpz_t format aren't optimised nor are any montgomery forms 3139160814Ssimon cached, and on x86 it appears OpenSSL's own performance has caught up. 3140160814Ssimon However there are likely to be other architectures where GMP could 3141160814Ssimon provide a boost. This ENGINE is not built in by default, but it can be 3142160814Ssimon specified at Configure time and should be accompanied by the necessary 3143160814Ssimon linker additions, eg; 3144160814Ssimon ./config -DOPENSSL_USE_GMP -lgmp 3145160814Ssimon [Geoff Thorpe] 3146160814Ssimon 3147160814Ssimon *) "openssl engine" will not display ENGINE/DSO load failure errors when 3148160814Ssimon testing availability of engines with "-t" - the old behaviour is 3149160814Ssimon produced by increasing the feature's verbosity with "-tt". 3150160814Ssimon [Geoff Thorpe] 3151160814Ssimon 3152160814Ssimon *) ECDSA routines: under certain error conditions uninitialized BN objects 3153160814Ssimon could be freed. Solution: make sure initialization is performed early 3154160814Ssimon enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 3155160814Ssimon via PR#459) 3156160814Ssimon [Lutz Jaenicke] 3157160814Ssimon 3158160814Ssimon *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 3159160814Ssimon and DH_METHOD (eg. by ENGINE implementations) to override the normal 3160160814Ssimon software implementations. For DSA and DH, parameter generation can 3161160814Ssimon also be overriden by providing the appropriate method callbacks. 3162160814Ssimon [Geoff Thorpe] 3163160814Ssimon 3164160814Ssimon *) Change the "progress" mechanism used in key-generation and 3165160814Ssimon primality testing to functions that take a new BN_GENCB pointer in 3166160814Ssimon place of callback/argument pairs. The new API functions have "_ex" 3167160814Ssimon postfixes and the older functions are reimplemented as wrappers for 3168160814Ssimon the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 3169160814Ssimon declarations of the old functions to help (graceful) attempts to 3170160814Ssimon migrate to the new functions. Also, the new key-generation API 3171160814Ssimon functions operate on a caller-supplied key-structure and return 3172160814Ssimon success/failure rather than returning a key or NULL - this is to 3173160814Ssimon help make "keygen" another member function of RSA_METHOD etc. 3174160814Ssimon 3175160814Ssimon Example for using the new callback interface: 3176160814Ssimon 3177160814Ssimon int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 3178160814Ssimon void *my_arg = ...; 3179160814Ssimon BN_GENCB my_cb; 3180160814Ssimon 3181160814Ssimon BN_GENCB_set(&my_cb, my_callback, my_arg); 3182160814Ssimon 3183160814Ssimon return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 3184160814Ssimon /* For the meaning of a, b in calls to my_callback(), see the 3185160814Ssimon * documentation of the function that calls the callback. 3186160814Ssimon * cb will point to my_cb; my_arg can be retrieved as cb->arg. 3187160814Ssimon * my_callback should return 1 if it wants BN_is_prime_ex() 3188160814Ssimon * to continue, or 0 to stop. 3189160814Ssimon */ 3190160814Ssimon 3191160814Ssimon [Geoff Thorpe] 3192160814Ssimon 3193160814Ssimon *) Change the ZLIB compression method to be stateful, and make it 3194160814Ssimon available to TLS with the number defined in 3195160814Ssimon draft-ietf-tls-compression-04.txt. 3196160814Ssimon [Richard Levitte] 3197160814Ssimon 3198160814Ssimon *) Add the ASN.1 structures and functions for CertificatePair, which 3199160814Ssimon is defined as follows (according to X.509_4thEditionDraftV6.pdf): 3200160814Ssimon 3201160814Ssimon CertificatePair ::= SEQUENCE { 3202160814Ssimon forward [0] Certificate OPTIONAL, 3203160814Ssimon reverse [1] Certificate OPTIONAL, 3204160814Ssimon -- at least one of the pair shall be present -- } 3205160814Ssimon 3206160814Ssimon Also implement the PEM functions to read and write certificate 3207160814Ssimon pairs, and defined the PEM tag as "CERTIFICATE PAIR". 3208160814Ssimon 3209160814Ssimon This needed to be defined, mostly for the sake of the LDAP 3210160814Ssimon attribute crossCertificatePair, but may prove useful elsewhere as 3211160814Ssimon well. 3212160814Ssimon [Richard Levitte] 3213160814Ssimon 3214160814Ssimon *) Make it possible to inhibit symlinking of shared libraries in 3215160814Ssimon Makefile.shared, for Cygwin's sake. 3216160814Ssimon [Richard Levitte] 3217160814Ssimon 3218160814Ssimon *) Extend the BIGNUM API by creating a function 3219160814Ssimon void BN_set_negative(BIGNUM *a, int neg); 3220160814Ssimon and a macro that behave like 3221160814Ssimon int BN_is_negative(const BIGNUM *a); 3222160814Ssimon 3223160814Ssimon to avoid the need to access 'a->neg' directly in applications. 3224160814Ssimon [Nils Larsch] 3225160814Ssimon 3226160814Ssimon *) Implement fast modular reduction for pseudo-Mersenne primes 3227160814Ssimon used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 3228160814Ssimon EC_GROUP_new_curve_GFp() will now automatically use this 3229160814Ssimon if applicable. 3230160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3231160814Ssimon 3232160814Ssimon *) Add new lock type (CRYPTO_LOCK_BN). 3233160814Ssimon [Bodo Moeller] 3234160814Ssimon 3235160814Ssimon *) Change the ENGINE framework to automatically load engines 3236160814Ssimon dynamically from specific directories unless they could be 3237160814Ssimon found to already be built in or loaded. Move all the 3238160814Ssimon current engines except for the cryptodev one to a new 3239160814Ssimon directory engines/. 3240160814Ssimon The engines in engines/ are built as shared libraries if 3241160814Ssimon the "shared" options was given to ./Configure or ./config. 3242160814Ssimon Otherwise, they are inserted in libcrypto.a. 3243160814Ssimon /usr/local/ssl/engines is the default directory for dynamic 3244160814Ssimon engines, but that can be overriden at configure time through 3245160814Ssimon the usual use of --prefix and/or --openssldir, and at run 3246160814Ssimon time with the environment variable OPENSSL_ENGINES. 3247160814Ssimon [Geoff Thorpe and Richard Levitte] 3248160814Ssimon 3249160814Ssimon *) Add Makefile.shared, a helper makefile to build shared 3250160814Ssimon libraries. Addapt Makefile.org. 3251160814Ssimon [Richard Levitte] 3252160814Ssimon 3253160814Ssimon *) Add version info to Win32 DLLs. 3254160814Ssimon [Peter 'Luna' Runestig" <peter@runestig.com>] 3255160814Ssimon 3256160814Ssimon *) Add new 'medium level' PKCS#12 API. Certificates and keys 3257160814Ssimon can be added using this API to created arbitrary PKCS#12 3258160814Ssimon files while avoiding the low level API. 3259160814Ssimon 3260160814Ssimon New options to PKCS12_create(), key or cert can be NULL and 3261160814Ssimon will then be omitted from the output file. The encryption 3262160814Ssimon algorithm NIDs can be set to -1 for no encryption, the mac 3263160814Ssimon iteration count can be set to 0 to omit the mac. 3264160814Ssimon 3265160814Ssimon Enhance pkcs12 utility by making the -nokeys and -nocerts 3266160814Ssimon options work when creating a PKCS#12 file. New option -nomac 3267160814Ssimon to omit the mac, NONE can be set for an encryption algorithm. 3268160814Ssimon New code is modified to use the enhanced PKCS12_create() 3269160814Ssimon instead of the low level API. 3270160814Ssimon [Steve Henson] 3271160814Ssimon 3272160814Ssimon *) Extend ASN1 encoder to support indefinite length constructed 3273160814Ssimon encoding. This can output sequences tags and octet strings in 3274160814Ssimon this form. Modify pk7_asn1.c to support indefinite length 3275160814Ssimon encoding. This is experimental and needs additional code to 3276160814Ssimon be useful, such as an ASN1 bio and some enhanced streaming 3277160814Ssimon PKCS#7 code. 3278160814Ssimon 3279160814Ssimon Extend template encode functionality so that tagging is passed 3280160814Ssimon down to the template encoder. 3281160814Ssimon [Steve Henson] 3282160814Ssimon 3283160814Ssimon *) Let 'openssl req' fail if an argument to '-newkey' is not 3284160814Ssimon recognized instead of using RSA as a default. 3285160814Ssimon [Bodo Moeller] 3286160814Ssimon 3287160814Ssimon *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 3288160814Ssimon As these are not official, they are not included in "ALL"; 3289160814Ssimon the "ECCdraft" ciphersuite group alias can be used to select them. 3290160814Ssimon [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 3291160814Ssimon 3292160814Ssimon *) Add ECDH engine support. 3293160814Ssimon [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 3294160814Ssimon 3295160814Ssimon *) Add ECDH in new directory crypto/ecdh/. 3296160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 3297160814Ssimon 3298160814Ssimon *) Let BN_rand_range() abort with an error after 100 iterations 3299160814Ssimon without success (which indicates a broken PRNG). 3300160814Ssimon [Bodo Moeller] 3301160814Ssimon 3302160814Ssimon *) Change BN_mod_sqrt() so that it verifies that the input value 3303160814Ssimon is really the square of the return value. (Previously, 3304160814Ssimon BN_mod_sqrt would show GIGO behaviour.) 3305160814Ssimon [Bodo Moeller] 3306160814Ssimon 3307160814Ssimon *) Add named elliptic curves over binary fields from X9.62, SECG, 3308160814Ssimon and WAP/WTLS; add OIDs that were still missing. 3309160814Ssimon 3310160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3311160814Ssimon (Sun Microsystems Laboratories)] 3312160814Ssimon 3313160814Ssimon *) Extend the EC library for elliptic curves over binary fields 3314160814Ssimon (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 3315160814Ssimon New EC_METHOD: 3316160814Ssimon 3317160814Ssimon EC_GF2m_simple_method 3318160814Ssimon 3319160814Ssimon New API functions: 3320160814Ssimon 3321160814Ssimon EC_GROUP_new_curve_GF2m 3322160814Ssimon EC_GROUP_set_curve_GF2m 3323160814Ssimon EC_GROUP_get_curve_GF2m 3324160814Ssimon EC_POINT_set_affine_coordinates_GF2m 3325160814Ssimon EC_POINT_get_affine_coordinates_GF2m 3326160814Ssimon EC_POINT_set_compressed_coordinates_GF2m 3327160814Ssimon 3328160814Ssimon Point compression for binary fields is disabled by default for 3329160814Ssimon patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 3330160814Ssimon enable it). 3331160814Ssimon 3332160814Ssimon As binary polynomials are represented as BIGNUMs, various members 3333160814Ssimon of the EC_GROUP and EC_POINT data structures can be shared 3334160814Ssimon between the implementations for prime fields and binary fields; 3335160814Ssimon the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 3336160814Ssimon are essentially identical to their ..._GFp counterparts. 3337160814Ssimon (For simplicity, the '..._GFp' prefix has been dropped from 3338160814Ssimon various internal method names.) 3339160814Ssimon 3340160814Ssimon An internal 'field_div' method (similar to 'field_mul' and 3341160814Ssimon 'field_sqr') has been added; this is used only for binary fields. 3342160814Ssimon 3343160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3344160814Ssimon (Sun Microsystems Laboratories)] 3345160814Ssimon 3346160814Ssimon *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 3347160814Ssimon through methods ('mul', 'precompute_mult'). 3348160814Ssimon 3349160814Ssimon The generic implementations (now internally called 'ec_wNAF_mul' 3350160814Ssimon and 'ec_wNAF_precomputed_mult') remain the default if these 3351160814Ssimon methods are undefined. 3352160814Ssimon 3353160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3354160814Ssimon (Sun Microsystems Laboratories)] 3355160814Ssimon 3356160814Ssimon *) New function EC_GROUP_get_degree, which is defined through 3357160814Ssimon EC_METHOD. For curves over prime fields, this returns the bit 3358160814Ssimon length of the modulus. 3359160814Ssimon 3360160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3361160814Ssimon (Sun Microsystems Laboratories)] 3362160814Ssimon 3363160814Ssimon *) New functions EC_GROUP_dup, EC_POINT_dup. 3364160814Ssimon (These simply call ..._new and ..._copy). 3365160814Ssimon 3366160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3367160814Ssimon (Sun Microsystems Laboratories)] 3368160814Ssimon 3369160814Ssimon *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 3370160814Ssimon Polynomials are represented as BIGNUMs (where the sign bit is not 3371160814Ssimon used) in the following functions [macros]: 3372160814Ssimon 3373160814Ssimon BN_GF2m_add 3374160814Ssimon BN_GF2m_sub [= BN_GF2m_add] 3375160814Ssimon BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 3376160814Ssimon BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 3377160814Ssimon BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 3378160814Ssimon BN_GF2m_mod_inv 3379160814Ssimon BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 3380160814Ssimon BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 3381160814Ssimon BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 3382160814Ssimon BN_GF2m_cmp [= BN_ucmp] 3383160814Ssimon 3384160814Ssimon (Note that only the 'mod' functions are actually for fields GF(2^m). 3385160814Ssimon BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 3386160814Ssimon 3387160814Ssimon For some functions, an the irreducible polynomial defining a 3388160814Ssimon field can be given as an 'unsigned int[]' with strictly 3389160814Ssimon decreasing elements giving the indices of those bits that are set; 3390160814Ssimon i.e., p[] represents the polynomial 3391160814Ssimon f(t) = t^p[0] + t^p[1] + ... + t^p[k] 3392160814Ssimon where 3393160814Ssimon p[0] > p[1] > ... > p[k] = 0. 3394160814Ssimon This applies to the following functions: 3395160814Ssimon 3396160814Ssimon BN_GF2m_mod_arr 3397160814Ssimon BN_GF2m_mod_mul_arr 3398160814Ssimon BN_GF2m_mod_sqr_arr 3399160814Ssimon BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 3400160814Ssimon BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 3401160814Ssimon BN_GF2m_mod_exp_arr 3402160814Ssimon BN_GF2m_mod_sqrt_arr 3403160814Ssimon BN_GF2m_mod_solve_quad_arr 3404160814Ssimon BN_GF2m_poly2arr 3405160814Ssimon BN_GF2m_arr2poly 3406160814Ssimon 3407160814Ssimon Conversion can be performed by the following functions: 3408160814Ssimon 3409160814Ssimon BN_GF2m_poly2arr 3410160814Ssimon BN_GF2m_arr2poly 3411160814Ssimon 3412160814Ssimon bntest.c has additional tests for binary polynomial arithmetic. 3413160814Ssimon 3414160814Ssimon Two implementations for BN_GF2m_mod_div() are available. 3415160814Ssimon The default algorithm simply uses BN_GF2m_mod_inv() and 3416160814Ssimon BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 3417160814Ssimon if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 3418160814Ssimon copyright notice in crypto/bn/bn_gf2m.c before enabling it). 3419160814Ssimon 3420160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3421160814Ssimon (Sun Microsystems Laboratories)] 3422160814Ssimon 3423160814Ssimon *) Add new error code 'ERR_R_DISABLED' that can be used when some 3424160814Ssimon functionality is disabled at compile-time. 3425160814Ssimon [Douglas Stebila <douglas.stebila@sun.com>] 3426160814Ssimon 3427160814Ssimon *) Change default behaviour of 'openssl asn1parse' so that more 3428160814Ssimon information is visible when viewing, e.g., a certificate: 3429160814Ssimon 3430160814Ssimon Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 3431160814Ssimon mode the content of non-printable OCTET STRINGs is output in a 3432160814Ssimon style similar to INTEGERs, but with '[HEX DUMP]' prepended to 3433160814Ssimon avoid the appearance of a printable string. 3434160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3435160814Ssimon 3436160814Ssimon *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 3437160814Ssimon functions 3438160814Ssimon EC_GROUP_set_asn1_flag() 3439160814Ssimon EC_GROUP_get_asn1_flag() 3440160814Ssimon EC_GROUP_set_point_conversion_form() 3441160814Ssimon EC_GROUP_get_point_conversion_form() 3442160814Ssimon These control ASN1 encoding details: 3443160814Ssimon - Curves (i.e., groups) are encoded explicitly unless asn1_flag 3444160814Ssimon has been set to OPENSSL_EC_NAMED_CURVE. 3445160814Ssimon - Points are encoded in uncompressed form by default; options for 3446160814Ssimon asn1_for are as for point2oct, namely 3447160814Ssimon POINT_CONVERSION_COMPRESSED 3448160814Ssimon POINT_CONVERSION_UNCOMPRESSED 3449160814Ssimon POINT_CONVERSION_HYBRID 3450160814Ssimon 3451160814Ssimon Also add 'seed' and 'seed_len' members to EC_GROUP with access 3452160814Ssimon functions 3453160814Ssimon EC_GROUP_set_seed() 3454160814Ssimon EC_GROUP_get0_seed() 3455160814Ssimon EC_GROUP_get_seed_len() 3456160814Ssimon This is used only for ASN1 purposes (so far). 3457160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3458160814Ssimon 3459160814Ssimon *) Add 'field_type' member to EC_METHOD, which holds the NID 3460160814Ssimon of the appropriate field type OID. The new function 3461160814Ssimon EC_METHOD_get_field_type() returns this value. 3462160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3463160814Ssimon 3464160814Ssimon *) Add functions 3465160814Ssimon EC_POINT_point2bn() 3466160814Ssimon EC_POINT_bn2point() 3467160814Ssimon EC_POINT_point2hex() 3468160814Ssimon EC_POINT_hex2point() 3469160814Ssimon providing useful interfaces to EC_POINT_point2oct() and 3470160814Ssimon EC_POINT_oct2point(). 3471160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3472160814Ssimon 3473160814Ssimon *) Change internals of the EC library so that the functions 3474160814Ssimon EC_GROUP_set_generator() 3475160814Ssimon EC_GROUP_get_generator() 3476160814Ssimon EC_GROUP_get_order() 3477160814Ssimon EC_GROUP_get_cofactor() 3478160814Ssimon are implemented directly in crypto/ec/ec_lib.c and not dispatched 3479160814Ssimon to methods, which would lead to unnecessary code duplication when 3480160814Ssimon adding different types of curves. 3481160814Ssimon [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 3482160814Ssimon 3483160814Ssimon *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 3484160814Ssimon arithmetic, and such that modified wNAFs are generated 3485160814Ssimon (which avoid length expansion in many cases). 3486160814Ssimon [Bodo Moeller] 3487160814Ssimon 3488160814Ssimon *) Add a function EC_GROUP_check_discriminant() (defined via 3489160814Ssimon EC_METHOD) that verifies that the curve discriminant is non-zero. 3490160814Ssimon 3491160814Ssimon Add a function EC_GROUP_check() that makes some sanity tests 3492160814Ssimon on a EC_GROUP, its generator and order. This includes 3493160814Ssimon EC_GROUP_check_discriminant(). 3494160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3495160814Ssimon 3496160814Ssimon *) Add ECDSA in new directory crypto/ecdsa/. 3497160814Ssimon 3498160814Ssimon Add applications 'openssl ecparam' and 'openssl ecdsa' 3499160814Ssimon (these are based on 'openssl dsaparam' and 'openssl dsa'). 3500160814Ssimon 3501160814Ssimon ECDSA support is also included in various other files across the 3502160814Ssimon library. Most notably, 3503160814Ssimon - 'openssl req' now has a '-newkey ecdsa:file' option; 3504160814Ssimon - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 3505160814Ssimon - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 3506160814Ssimon d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 3507160814Ssimon them suitable for ECDSA where domain parameters must be 3508160814Ssimon extracted before the specific public key; 3509160814Ssimon - ECDSA engine support has been added. 3510160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3511160814Ssimon 3512160814Ssimon *) Include some named elliptic curves, and add OIDs from X9.62, 3513160814Ssimon SECG, and WAP/WTLS. Each curve can be obtained from the new 3514160814Ssimon function 3515160814Ssimon EC_GROUP_new_by_curve_name(), 3516160814Ssimon and the list of available named curves can be obtained with 3517160814Ssimon EC_get_builtin_curves(). 3518160814Ssimon Also add a 'curve_name' member to EC_GROUP objects, which can be 3519160814Ssimon accessed via 3520160814Ssimon EC_GROUP_set_curve_name() 3521160814Ssimon EC_GROUP_get_curve_name() 3522160814Ssimon [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 3523160814Ssimon 3524160814Ssimon *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 3525160814Ssimon was actually never needed) and in BN_mul(). The removal in BN_mul() 3526160814Ssimon required a small change in bn_mul_part_recursive() and the addition 3527160814Ssimon of the functions bn_cmp_part_words(), bn_sub_part_words() and 3528160814Ssimon bn_add_part_words(), which do the same thing as bn_cmp_words(), 3529160814Ssimon bn_sub_words() and bn_add_words() except they take arrays with 3530160814Ssimon differing sizes. 3531160814Ssimon [Richard Levitte] 3532160814Ssimon 3533194206Ssimon Changes between 0.9.7l and 0.9.7m [23 Feb 2007] 3534194206Ssimon 3535167612Ssimon *) Cleanse PEM buffers before freeing them since they may contain 3536167612Ssimon sensitive data. 3537167612Ssimon [Benjamin Bennett <ben@psc.edu>] 3538167612Ssimon 3539167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 3540167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 3541167612Ssimon authentication-only ciphersuites. 3542167612Ssimon [Bodo Moeller] 3543167612Ssimon 3544167612Ssimon *) Since AES128 and AES256 share a single mask bit in the logic of 3545167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 3546167612Ssimon kludge to work properly if AES128 is available and AES256 isn't. 3547167612Ssimon [Victor Duchovni] 3548167612Ssimon 3549194206Ssimon *) Expand security boundary to match 1.1.1 module. 3550194206Ssimon [Steve Henson] 3551194206Ssimon 3552194206Ssimon *) Remove redundant features: hash file source, editing of test vectors 3553194206Ssimon modify fipsld to use external fips_premain.c signature. 3554194206Ssimon [Steve Henson] 3555194206Ssimon 3556194206Ssimon *) New perl script mkfipsscr.pl to create shell scripts or batch files to 3557194206Ssimon run algorithm test programs. 3558194206Ssimon [Steve Henson] 3559194206Ssimon 3560194206Ssimon *) Make algorithm test programs more tolerant of whitespace. 3561194206Ssimon [Steve Henson] 3562194206Ssimon 3563167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 3564167612Ssimon protocol version while receiving ClientHello even if the 3565167612Ssimon ClientHello is fragmented. (The server can't insist on the 3566167612Ssimon particular protocol version it has chosen before the ServerHello 3567167612Ssimon message has informed the client about his choice.) 3568167612Ssimon [Bodo Moeller] 3569167612Ssimon 3570167612Ssimon *) Load error codes if they are not already present instead of using a 3571167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 3572167612Ssimon [Steve Henson] 3573167612Ssimon 3574167612Ssimon Changes between 0.9.7k and 0.9.7l [28 Sep 2006] 3575167612Ssimon 3576167612Ssimon *) Introduce limits to prevent malicious keys being able to 3577167612Ssimon cause a denial of service. (CVE-2006-2940) 3578167612Ssimon [Steve Henson, Bodo Moeller] 3579167612Ssimon 3580167612Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 3581167612Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 3582167612Ssimon 3583167612Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 3584167612Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 3585167612Ssimon 3586167612Ssimon *) Fix SSL client code which could crash if connecting to a 3587167612Ssimon malicious SSLv2 server. (CVE-2006-4343) 3588167612Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 3589167612Ssimon 3590162911Ssimon *) Change ciphersuite string processing so that an explicit 3591162911Ssimon ciphersuite selects this one ciphersuite (so that "AES256-SHA" 3592162911Ssimon will no longer include "AES128-SHA"), and any other similar 3593162911Ssimon ciphersuite (same bitmap) from *other* protocol versions (so that 3594162911Ssimon "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 3595162911Ssimon SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 3596162911Ssimon changes from 0.9.8b and 0.9.8d. 3597162911Ssimon [Bodo Moeller] 3598162911Ssimon 3599162911Ssimon Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 3600162911Ssimon 3601162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 3602162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 3603162911Ssimon 3604162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 3605162911Ssimon possible instead of select(), since the latter has some 3606162911Ssimon undesirable limitations. 3607162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 3608162911Ssimon 3609162911Ssimon *) Disable rogue ciphersuites: 3610162911Ssimon 3611162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 3612162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 3613162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 3614162911Ssimon 3615162911Ssimon The latter two were purportedly from 3616162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 3617162911Ssimon appear there. 3618162911Ssimon 3619162911Ssimon Also deactive the remaining ciphersuites from 3620162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 3621162911Ssimon unofficial, and the ID has long expired. 3622162911Ssimon [Bodo Moeller] 3623162911Ssimon 3624162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 3625162911Ssimon dual-core machines) and other potential thread-safety issues. 3626162911Ssimon [Bodo Moeller] 3627162911Ssimon 3628162911Ssimon Changes between 0.9.7i and 0.9.7j [04 May 2006] 3629162911Ssimon 3630162911Ssimon *) Adapt fipsld and the build system to link against the validated FIPS 3631162911Ssimon module in FIPS mode. 3632162911Ssimon [Steve Henson] 3633162911Ssimon 3634162911Ssimon *) Fixes for VC++ 2005 build under Windows. 3635162911Ssimon [Steve Henson] 3636162911Ssimon 3637162911Ssimon *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 3638162911Ssimon from a Windows bash shell such as MSYS. It is autodetected from the 3639162911Ssimon "config" script when run from a VC++ environment. Modify standard VC++ 3640162911Ssimon build to use fipscanister.o from the GNU make build. 3641162911Ssimon [Steve Henson] 3642162911Ssimon 3643160814Ssimon Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 3644160814Ssimon 3645160814Ssimon *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 3646160814Ssimon The value now differs depending on if you build for FIPS or not. 3647160814Ssimon BEWARE! A program linked with a shared FIPSed libcrypto can't be 3648160814Ssimon safely run with a non-FIPSed libcrypto, as it may crash because of 3649160814Ssimon the difference induced by this change. 3650160814Ssimon [Andy Polyakov] 3651160814Ssimon 3652160814Ssimon Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 3653160814Ssimon 3654160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 3655160814Ssimon (part of SSL_OP_ALL). This option used to disable the 3656160814Ssimon countermeasure against man-in-the-middle protocol-version 3657160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 3658160814Ssimon idea. (CVE-2005-2969) 3659160814Ssimon 3660160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 3661160814Ssimon for Information Security, National Institute of Advanced Industrial 3662160814Ssimon Science and Technology [AIST], Japan)] 3663160814Ssimon 3664160814Ssimon *) Minimal support for X9.31 signatures and PSS padding modes. This is 3665160814Ssimon mainly for FIPS compliance and not fully integrated at this stage. 3666160814Ssimon [Steve Henson] 3667160814Ssimon 3668160814Ssimon *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 3669160814Ssimon the exponentiation using a fixed-length exponent. (Otherwise, 3670160814Ssimon the information leaked through timing could expose the secret key 3671160814Ssimon after many signatures; cf. Bleichenbacher's attack on DSA with 3672160814Ssimon biased k.) 3673160814Ssimon [Bodo Moeller] 3674160814Ssimon 3675160814Ssimon *) Make a new fixed-window mod_exp implementation the default for 3676160814Ssimon RSA, DSA, and DH private-key operations so that the sequence of 3677160814Ssimon squares and multiplies and the memory access pattern are 3678160814Ssimon independent of the particular secret key. This will mitigate 3679160814Ssimon cache-timing and potential related attacks. 3680160814Ssimon 3681160814Ssimon BN_mod_exp_mont_consttime() is the new exponentiation implementation, 3682160814Ssimon and this is automatically used by BN_mod_exp_mont() if the new flag 3683160814Ssimon BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 3684160814Ssimon will use this BN flag for private exponents unless the flag 3685160814Ssimon RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 3686160814Ssimon DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 3687160814Ssimon 3688160814Ssimon [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 3689160814Ssimon 3690160814Ssimon *) Change the client implementation for SSLv23_method() and 3691160814Ssimon SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 3692160814Ssimon Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 3693160814Ssimon (Previously, the SSL 2.0 backwards compatible Client Hello 3694160814Ssimon message format would be used even with SSL_OP_NO_SSLv2.) 3695160814Ssimon [Bodo Moeller] 3696160814Ssimon 3697160814Ssimon *) Add support for smime-type MIME parameter in S/MIME messages which some 3698160814Ssimon clients need. 3699160814Ssimon [Steve Henson] 3700160814Ssimon 3701160814Ssimon *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 3702160814Ssimon a threadsafe manner. Modify rsa code to use new function and add calls 3703160814Ssimon to dsa and dh code (which had race conditions before). 3704160814Ssimon [Steve Henson] 3705160814Ssimon 3706160814Ssimon *) Include the fixed error library code in the C error file definitions 3707160814Ssimon instead of fixing them up at runtime. This keeps the error code 3708160814Ssimon structures constant. 3709160814Ssimon [Steve Henson] 3710160814Ssimon 3711160814Ssimon Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 3712160814Ssimon 3713160814Ssimon [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 3714160814Ssimon OpenSSL 0.9.8.] 3715160814Ssimon 3716160814Ssimon *) Fixes for newer kerberos headers. NB: the casts are needed because 3717160814Ssimon the 'length' field is signed on one version and unsigned on another 3718160814Ssimon with no (?) obvious way to tell the difference, without these VC++ 3719160814Ssimon complains. Also the "definition" of FAR (blank) is no longer included 3720160814Ssimon nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 3721160814Ssimon some needed definitions. 3722160814Ssimon [Steve Henson] 3723160814Ssimon 3724160814Ssimon *) Undo Cygwin change. 3725160814Ssimon [Ulf M�ller] 3726160814Ssimon 3727160814Ssimon *) Added support for proxy certificates according to RFC 3820. 3728160814Ssimon Because they may be a security thread to unaware applications, 3729160814Ssimon they must be explicitely allowed in run-time. See 3730160814Ssimon docs/HOWTO/proxy_certificates.txt for further information. 3731160814Ssimon [Richard Levitte] 3732160814Ssimon 3733160814Ssimon Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 3734160814Ssimon 3735160814Ssimon *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 3736160814Ssimon server and client random values. Previously 3737160814Ssimon (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 3738160814Ssimon less random data when sizeof(time_t) > 4 (some 64 bit platforms). 3739160814Ssimon 3740160814Ssimon This change has negligible security impact because: 3741160814Ssimon 3742160814Ssimon 1. Server and client random values still have 24 bytes of pseudo random 3743160814Ssimon data. 3744160814Ssimon 3745160814Ssimon 2. Server and client random values are sent in the clear in the initial 3746160814Ssimon handshake. 3747160814Ssimon 3748160814Ssimon 3. The master secret is derived using the premaster secret (48 bytes in 3749160814Ssimon size for static RSA ciphersuites) as well as client server and random 3750160814Ssimon values. 3751160814Ssimon 3752160814Ssimon The OpenSSL team would like to thank the UK NISCC for bringing this issue 3753160814Ssimon to our attention. 3754160814Ssimon 3755160814Ssimon [Stephen Henson, reported by UK NISCC] 3756160814Ssimon 3757160814Ssimon *) Use Windows randomness collection on Cygwin. 3758160814Ssimon [Ulf M�ller] 3759160814Ssimon 3760160814Ssimon *) Fix hang in EGD/PRNGD query when communication socket is closed 3761160814Ssimon prematurely by EGD/PRNGD. 3762160814Ssimon [Darren Tucker <dtucker@zip.com.au> via Lutz J�nicke, resolves #1014] 3763160814Ssimon 3764160814Ssimon *) Prompt for pass phrases when appropriate for PKCS12 input format. 3765160814Ssimon [Steve Henson] 3766160814Ssimon 3767160814Ssimon *) Back-port of selected performance improvements from development 3768160814Ssimon branch, as well as improved support for PowerPC platforms. 3769160814Ssimon [Andy Polyakov] 3770160814Ssimon 3771160814Ssimon *) Add lots of checks for memory allocation failure, error codes to indicate 3772160814Ssimon failure and freeing up memory if a failure occurs. 3773160814Ssimon [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 3774160814Ssimon 3775160814Ssimon *) Add new -passin argument to dgst. 3776160814Ssimon [Steve Henson] 3777160814Ssimon 3778160814Ssimon *) Perform some character comparisons of different types in X509_NAME_cmp: 3779160814Ssimon this is needed for some certificates that reencode DNs into UTF8Strings 3780160814Ssimon (in violation of RFC3280) and can't or wont issue name rollover 3781160814Ssimon certificates. 3782160814Ssimon [Steve Henson] 3783160814Ssimon 3784160814Ssimon *) Make an explicit check during certificate validation to see that 3785160814Ssimon the CA setting in each certificate on the chain is correct. As a 3786160814Ssimon side effect always do the following basic checks on extensions, 3787160814Ssimon not just when there's an associated purpose to the check: 3788160814Ssimon 3789160814Ssimon - if there is an unhandled critical extension (unless the user 3790160814Ssimon has chosen to ignore this fault) 3791160814Ssimon - if the path length has been exceeded (if one is set at all) 3792160814Ssimon - that certain extensions fit the associated purpose (if one has 3793160814Ssimon been given) 3794160814Ssimon [Richard Levitte] 3795160814Ssimon 3796142425Snectar Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 3797142425Snectar 3798142425Snectar *) Avoid a race condition when CRLs are checked in a multi threaded 3799142425Snectar environment. This would happen due to the reordering of the revoked 3800142425Snectar entries during signature checking and serial number lookup. Now the 3801142425Snectar encoding is cached and the serial number sort performed under a lock. 3802142425Snectar Add new STACK function sk_is_sorted(). 3803142425Snectar [Steve Henson] 3804142425Snectar 3805142425Snectar *) Add Delta CRL to the extension code. 3806142425Snectar [Steve Henson] 3807142425Snectar 3808142425Snectar *) Various fixes to s3_pkt.c so alerts are sent properly. 3809142425Snectar [David Holmes <d.holmes@f5.com>] 3810142425Snectar 3811142425Snectar *) Reduce the chances of duplicate issuer name and serial numbers (in 3812142425Snectar violation of RFC3280) using the OpenSSL certificate creation utilities. 3813142425Snectar This is done by creating a random 64 bit value for the initial serial 3814142425Snectar number when a serial number file is created or when a self signed 3815142425Snectar certificate is created using 'openssl req -x509'. The initial serial 3816142425Snectar number file is created using 'openssl x509 -next_serial' in CA.pl 3817142425Snectar rather than being initialized to 1. 3818142425Snectar [Steve Henson] 3819142425Snectar 3820127128Snectar Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 3821127128Snectar 3822127128Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 3823160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 3824127128Snectar [Joe Orton, Steve Henson] 3825127128Snectar 3826127128Snectar *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 3827160814Ssimon (CVE-2004-0112) 3828127128Snectar [Joe Orton, Steve Henson] 3829127128Snectar 3830127128Snectar *) Make it possible to have multiple active certificates with the same 3831127128Snectar subject in the CA index file. This is done only if the keyword 3832127128Snectar 'unique_subject' is set to 'no' in the main CA section (default 3833127128Snectar if 'CA_default') of the configuration file. The value is saved 3834127128Snectar with the database itself in a separate index attribute file, 3835127128Snectar named like the index file with '.attr' appended to the name. 3836127128Snectar [Richard Levitte] 3837127128Snectar 3838127128Snectar *) X509 verify fixes. Disable broken certificate workarounds when 3839127128Snectar X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 3840127128Snectar keyUsage extension present. Don't accept CRLs with unhandled critical 3841127128Snectar extensions: since verify currently doesn't process CRL extensions this 3842127128Snectar rejects a CRL with *any* critical extensions. Add new verify error codes 3843127128Snectar for these cases. 3844127128Snectar [Steve Henson] 3845127128Snectar 3846127128Snectar *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 3847127128Snectar A clarification of RFC2560 will require the use of OCTET STRINGs and 3848127128Snectar some implementations cannot handle the current raw format. Since OpenSSL 3849127128Snectar copies and compares OCSP nonces as opaque blobs without any attempt at 3850127128Snectar parsing them this should not create any compatibility issues. 3851127128Snectar [Steve Henson] 3852127128Snectar 3853127128Snectar *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 3854127128Snectar calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 3855127128Snectar this HMAC (and other) operations are several times slower than OpenSSL 3856127128Snectar < 0.9.7. 3857127128Snectar [Steve Henson] 3858127128Snectar 3859127128Snectar *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 3860127128Snectar [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 3861127128Snectar 3862127128Snectar *) Use the correct content when signing type "other". 3863127128Snectar [Steve Henson] 3864127128Snectar 3865120631Snectar Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 3866120631Snectar 3867120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 3868120631Snectar 3869120631Snectar Stop out of bounds reads in the ASN1 code when presented with 3870160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 3871120631Snectar 3872160814Ssimon Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 3873120631Snectar 3874120631Snectar If verify callback ignores invalid public key errors don't try to check 3875120631Snectar certificate signature with the NULL public key. 3876120631Snectar 3877120631Snectar [Steve Henson] 3878120631Snectar 3879120631Snectar *) New -ignore_err option in ocsp application to stop the server 3880120631Snectar exiting on the first error in a request. 3881120631Snectar [Steve Henson] 3882120631Snectar 3883120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 3884120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 3885120631Snectar specifications. 3886120631Snectar [Steve Henson] 3887120631Snectar 3888120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 3889120631Snectar extra data after the compression methods not only for TLS 1.0 3890120631Snectar but also for SSL 3.0 (as required by the specification). 3891120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 3892120631Snectar 3893120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 3894120631Snectar when it's 512 *bits* long, not 512 bytes. 3895120631Snectar [Richard Levitte] 3896120631Snectar 3897120631Snectar *) Change AES_cbc_encrypt() so it outputs exact multiple of 3898120631Snectar blocks during encryption. 3899120631Snectar [Richard Levitte] 3900120631Snectar 3901120631Snectar *) Various fixes to base64 BIO and non blocking I/O. On write 3902120631Snectar flushes were not handled properly if the BIO retried. On read 3903120631Snectar data was not being buffered properly and had various logic bugs. 3904120631Snectar This also affects blocking I/O when the data being decoded is a 3905120631Snectar certain size. 3906120631Snectar [Steve Henson] 3907120631Snectar 3908120631Snectar *) Various S/MIME bugfixes and compatibility changes: 3909120631Snectar output correct application/pkcs7 MIME type if 3910120631Snectar PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 3911120631Snectar Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 3912120631Snectar of files as .eml work). Correctly handle very long lines in MIME 3913120631Snectar parser. 3914120631Snectar [Steve Henson] 3915120631Snectar 3916120631Snectar Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 3917120631Snectar 3918120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 3919120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 3920120631Snectar a protocol version number mismatch like a decryption error 3921120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 3922120631Snectar [Bodo Moeller] 3923120631Snectar 3924120631Snectar *) Turn on RSA blinding by default in the default implementation 3925120631Snectar to avoid a timing attack. Applications that don't want it can call 3926120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 3927120631Snectar They would be ill-advised to do so in most cases. 3928120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 3929120631Snectar 3930120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 3931120631Snectar seeded (in this case, the secret RSA exponent is abused as 3932120631Snectar an unpredictable seed -- if it is not unpredictable, there 3933120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 3934120631Snectar by remembering the creator's thread ID in rsa->blinding and 3935120631Snectar having all other threads use local one-time blinding factors 3936120631Snectar (this requires more computation than sharing rsa->blinding, but 3937120631Snectar avoids excessive locking; and if an RSA object is not shared 3938120631Snectar between threads, blinding will still be very fast). 3939120631Snectar [Bodo Moeller] 3940120631Snectar 3941120631Snectar *) Fixed a typo bug that would cause ENGINE_set_default() to set an 3942120631Snectar ENGINE as defaults for all supported algorithms irrespective of 3943120631Snectar the 'flags' parameter. 'flags' is now honoured, so applications 3944120631Snectar should make sure they are passing it correctly. 3945120631Snectar [Geoff Thorpe] 3946120631Snectar 3947120631Snectar *) Target "mingw" now allows native Windows code to be generated in 3948120631Snectar the Cygwin environment as well as with the MinGW compiler. 3949120631Snectar [Ulf Moeller] 3950120631Snectar 3951111147Snectar Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 3952111147Snectar 3953111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 3954111147Snectar via timing by performing a MAC computation even if incorrrect 3955111147Snectar block cipher padding has been found. This is a countermeasure 3956111147Snectar against active attacks where the attacker has to distinguish 3957160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 3958111147Snectar 3959111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 3960111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 3961111147Snectar Martin Vuagnoux (EPFL, Ilion)] 3962111147Snectar 3963111147Snectar *) Make the no-err option work as intended. The intention with no-err 3964111147Snectar is not to have the whole error stack handling routines removed from 3965111147Snectar libcrypto, it's only intended to remove all the function name and 3966111147Snectar reason texts, thereby removing some of the footprint that may not 3967111147Snectar be interesting if those errors aren't displayed anyway. 3968111147Snectar 3969111147Snectar NOTE: it's still possible for any application or module to have it's 3970111147Snectar own set of error texts inserted. The routines are there, just not 3971111147Snectar used by default when no-err is given. 3972111147Snectar [Richard Levitte] 3973111147Snectar 3974111147Snectar *) Add support for FreeBSD on IA64. 3975111147Snectar [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 3976111147Snectar 3977111147Snectar *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 3978111147Snectar Kerberos function mit_des_cbc_cksum(). Before this change, 3979111147Snectar the value returned by DES_cbc_cksum() was like the one from 3980111147Snectar mit_des_cbc_cksum(), except the bytes were swapped. 3981111147Snectar [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 3982111147Snectar 3983111147Snectar *) Allow an application to disable the automatic SSL chain building. 3984111147Snectar Before this a rather primitive chain build was always performed in 3985111147Snectar ssl3_output_cert_chain(): an application had no way to send the 3986111147Snectar correct chain if the automatic operation produced an incorrect result. 3987111147Snectar 3988111147Snectar Now the chain builder is disabled if either: 3989111147Snectar 3990111147Snectar 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 3991111147Snectar 3992111147Snectar 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 3993111147Snectar 3994111147Snectar The reasoning behind this is that an application would not want the 3995111147Snectar auto chain building to take place if extra chain certificates are 3996111147Snectar present and it might also want a means of sending no additional 3997111147Snectar certificates (for example the chain has two certificates and the 3998111147Snectar root is omitted). 3999111147Snectar [Steve Henson] 4000111147Snectar 4001111147Snectar *) Add the possibility to build without the ENGINE framework. 4002111147Snectar [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 4003111147Snectar 4004111147Snectar *) Under Win32 gmtime() can return NULL: check return value in 4005111147Snectar OPENSSL_gmtime(). Add error code for case where gmtime() fails. 4006111147Snectar [Steve Henson] 4007111147Snectar 4008111147Snectar *) DSA routines: under certain error conditions uninitialized BN objects 4009111147Snectar could be freed. Solution: make sure initialization is performed early 4010111147Snectar enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 4011111147Snectar Nils Larsch <nla@trustcenter.de> via PR#459) 4012111147Snectar [Lutz Jaenicke] 4013111147Snectar 4014111147Snectar *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 4015111147Snectar checked on reconnect on the client side, therefore session resumption 4016111147Snectar could still fail with a "ssl session id is different" error. This 4017111147Snectar behaviour is masked when SSL_OP_ALL is used due to 4018111147Snectar SSL_OP_MICROSOFT_SESS_ID_BUG being set. 4019111147Snectar Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4020111147Snectar followup to PR #377. 4021111147Snectar [Lutz Jaenicke] 4022111147Snectar 4023111147Snectar *) IA-32 assembler support enhancements: unified ELF targets, support 4024111147Snectar for SCO/Caldera platforms, fix for Cygwin shared build. 4025111147Snectar [Andy Polyakov] 4026111147Snectar 4027111147Snectar *) Add support for FreeBSD on sparc64. As a consequence, support for 4028111147Snectar FreeBSD on non-x86 processors is separate from x86 processors on 4029111147Snectar the config script, much like the NetBSD support. 4030111147Snectar [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 4031111147Snectar 4032109998Smarkm Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 4033109998Smarkm 4034120631Snectar [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 4035120631Snectar OpenSSL 0.9.7.] 4036120631Snectar 4037109998Smarkm *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 4038109998Smarkm code (06) was taken as the first octet of the session ID and the last 4039109998Smarkm octet was ignored consequently. As a result SSLv2 client side session 4040109998Smarkm caching could not have worked due to the session ID mismatch between 4041109998Smarkm client and server. 4042109998Smarkm Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4043109998Smarkm PR #377. 4044109998Smarkm [Lutz Jaenicke] 4045109998Smarkm 4046109998Smarkm *) Change the declaration of needed Kerberos libraries to use EX_LIBS 4047109998Smarkm instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 4048109998Smarkm removed entirely. 4049109998Smarkm [Richard Levitte] 4050109998Smarkm 4051109998Smarkm *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 4052109998Smarkm seems that in spite of existing for more than a year, many application 4053109998Smarkm author have done nothing to provide the necessary callbacks, which 4054109998Smarkm means that this particular engine will not work properly anywhere. 4055109998Smarkm This is a very unfortunate situation which forces us, in the name 4056109998Smarkm of usability, to give the hw_ncipher.c a static lock, which is part 4057109998Smarkm of libcrypto. 4058109998Smarkm NOTE: This is for the 0.9.7 series ONLY. This hack will never 4059109998Smarkm appear in 0.9.8 or later. We EXPECT application authors to have 4060109998Smarkm dealt properly with this when 0.9.8 is released (unless we actually 4061109998Smarkm make such changes in the libcrypto locking code that changes will 4062109998Smarkm have to be made anyway). 4063109998Smarkm [Richard Levitte] 4064109998Smarkm 4065109998Smarkm *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 4066109998Smarkm octets have been read, EOF or an error occurs. Without this change 4067109998Smarkm some truncated ASN1 structures will not produce an error. 4068109998Smarkm [Steve Henson] 4069109998Smarkm 4070109998Smarkm *) Disable Heimdal support, since it hasn't been fully implemented. 4071109998Smarkm Still give the possibility to force the use of Heimdal, but with 4072109998Smarkm warnings and a request that patches get sent to openssl-dev. 4073109998Smarkm [Richard Levitte] 4074109998Smarkm 4075109998Smarkm *) Add the VC-CE target, introduce the WINCE sysname, and add 4076109998Smarkm INSTALL.WCE and appropriate conditionals to make it build. 4077109998Smarkm [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 4078109998Smarkm 4079109998Smarkm *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 4080109998Smarkm cygssl-x.y.z.dll, where x, y and z are the major, minor and 4081109998Smarkm edit numbers of the version. 4082109998Smarkm [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 4083109998Smarkm 4084109998Smarkm *) Introduce safe string copy and catenation functions 4085109998Smarkm (BUF_strlcpy() and BUF_strlcat()). 4086109998Smarkm [Ben Laurie (CHATS) and Richard Levitte] 4087109998Smarkm 4088109998Smarkm *) Avoid using fixed-size buffers for one-line DNs. 4089109998Smarkm [Ben Laurie (CHATS)] 4090109998Smarkm 4091109998Smarkm *) Add BUF_MEM_grow_clean() to avoid information leakage when 4092109998Smarkm resizing buffers containing secrets, and use where appropriate. 4093109998Smarkm [Ben Laurie (CHATS)] 4094109998Smarkm 4095109998Smarkm *) Avoid using fixed size buffers for configuration file location. 4096109998Smarkm [Ben Laurie (CHATS)] 4097109998Smarkm 4098109998Smarkm *) Avoid filename truncation for various CA files. 4099109998Smarkm [Ben Laurie (CHATS)] 4100109998Smarkm 4101109998Smarkm *) Use sizeof in preference to magic numbers. 4102109998Smarkm [Ben Laurie (CHATS)] 4103109998Smarkm 4104109998Smarkm *) Avoid filename truncation in cert requests. 4105109998Smarkm [Ben Laurie (CHATS)] 4106109998Smarkm 4107109998Smarkm *) Add assertions to check for (supposedly impossible) buffer 4108109998Smarkm overflows. 4109109998Smarkm [Ben Laurie (CHATS)] 4110109998Smarkm 4111109998Smarkm *) Don't cache truncated DNS entries in the local cache (this could 4112109998Smarkm potentially lead to a spoofing attack). 4113109998Smarkm [Ben Laurie (CHATS)] 4114109998Smarkm 4115109998Smarkm *) Fix various buffers to be large enough for hex/decimal 4116109998Smarkm representations in a platform independent manner. 4117109998Smarkm [Ben Laurie (CHATS)] 4118109998Smarkm 4119109998Smarkm *) Add CRYPTO_realloc_clean() to avoid information leakage when 4120109998Smarkm resizing buffers containing secrets, and use where appropriate. 4121109998Smarkm [Ben Laurie (CHATS)] 4122109998Smarkm 4123109998Smarkm *) Add BIO_indent() to avoid much slightly worrying code to do 4124109998Smarkm indents. 4125109998Smarkm [Ben Laurie (CHATS)] 4126109998Smarkm 4127109998Smarkm *) Convert sprintf()/BIO_puts() to BIO_printf(). 4128109998Smarkm [Ben Laurie (CHATS)] 4129109998Smarkm 4130109998Smarkm *) buffer_gets() could terminate with the buffer only half 4131109998Smarkm full. Fixed. 4132109998Smarkm [Ben Laurie (CHATS)] 4133109998Smarkm 4134109998Smarkm *) Add assertions to prevent user-supplied crypto functions from 4135109998Smarkm overflowing internal buffers by having large block sizes, etc. 4136109998Smarkm [Ben Laurie (CHATS)] 4137109998Smarkm 4138109998Smarkm *) New OPENSSL_assert() macro (similar to assert(), but enabled 4139109998Smarkm unconditionally). 4140109998Smarkm [Ben Laurie (CHATS)] 4141109998Smarkm 4142109998Smarkm *) Eliminate unused copy of key in RC4. 4143109998Smarkm [Ben Laurie (CHATS)] 4144109998Smarkm 4145109998Smarkm *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 4146109998Smarkm [Ben Laurie (CHATS)] 4147109998Smarkm 4148109998Smarkm *) Fix off-by-one error in EGD path. 4149109998Smarkm [Ben Laurie (CHATS)] 4150109998Smarkm 4151109998Smarkm *) If RANDFILE path is too long, ignore instead of truncating. 4152109998Smarkm [Ben Laurie (CHATS)] 4153109998Smarkm 4154109998Smarkm *) Eliminate unused and incorrectly sized X.509 structure 4155109998Smarkm CBCParameter. 4156109998Smarkm [Ben Laurie (CHATS)] 4157109998Smarkm 4158109998Smarkm *) Eliminate unused and dangerous function knumber(). 4159109998Smarkm [Ben Laurie (CHATS)] 4160109998Smarkm 4161109998Smarkm *) Eliminate unused and dangerous structure, KSSL_ERR. 4162109998Smarkm [Ben Laurie (CHATS)] 4163109998Smarkm 4164109998Smarkm *) Protect against overlong session ID context length in an encoded 4165109998Smarkm session object. Since these are local, this does not appear to be 4166109998Smarkm exploitable. 4167109998Smarkm [Ben Laurie (CHATS)] 4168109998Smarkm 4169109998Smarkm *) Change from security patch (see 0.9.6e below) that did not affect 4170109998Smarkm the 0.9.6 release series: 4171109998Smarkm 4172109998Smarkm Remote buffer overflow in SSL3 protocol - an attacker could 4173109998Smarkm supply an oversized master key in Kerberos-enabled versions. 4174160814Ssimon (CVE-2002-0657) 4175109998Smarkm [Ben Laurie (CHATS)] 4176109998Smarkm 4177109998Smarkm *) Change the SSL kerb5 codes to match RFC 2712. 4178109998Smarkm [Richard Levitte] 4179109998Smarkm 4180109998Smarkm *) Make -nameopt work fully for req and add -reqopt switch. 4181109998Smarkm [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 4182109998Smarkm 4183109998Smarkm *) The "block size" for block ciphers in CFB and OFB mode should be 1. 4184109998Smarkm [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 4185109998Smarkm 4186109998Smarkm *) Make sure tests can be performed even if the corresponding algorithms 4187109998Smarkm have been removed entirely. This was also the last step to make 4188109998Smarkm OpenSSL compilable with DJGPP under all reasonable conditions. 4189109998Smarkm [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 4190109998Smarkm 4191109998Smarkm *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 4192109998Smarkm to allow version independent disabling of normally unselected ciphers, 4193109998Smarkm which may be activated as a side-effect of selecting a single cipher. 4194109998Smarkm 4195109998Smarkm (E.g., cipher list string "RSA" enables ciphersuites that are left 4196109998Smarkm out of "ALL" because they do not provide symmetric encryption. 4197109998Smarkm "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 4198109998Smarkm [Lutz Jaenicke, Bodo Moeller] 4199109998Smarkm 4200109998Smarkm *) Add appropriate support for separate platform-dependent build 4201109998Smarkm directories. The recommended way to make a platform-dependent 4202109998Smarkm build directory is the following (tested on Linux), maybe with 4203109998Smarkm some local tweaks: 4204109998Smarkm 4205109998Smarkm # Place yourself outside of the OpenSSL source tree. In 4206109998Smarkm # this example, the environment variable OPENSSL_SOURCE 4207109998Smarkm # is assumed to contain the absolute OpenSSL source directory. 4208109998Smarkm mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 4209109998Smarkm cd objtree/"`uname -s`-`uname -r`-`uname -m`" 4210111147Snectar (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 4211109998Smarkm mkdir -p `dirname $F` 4212109998Smarkm ln -s $OPENSSL_SOURCE/$F $F 4213109998Smarkm done 4214109998Smarkm 4215109998Smarkm To be absolutely sure not to disturb the source tree, a "make clean" 4216109998Smarkm is a good thing. If it isn't successfull, don't worry about it, 4217109998Smarkm it probably means the source directory is very clean. 4218109998Smarkm [Richard Levitte] 4219109998Smarkm 4220109998Smarkm *) Make sure any ENGINE control commands make local copies of string 4221109998Smarkm pointers passed to them whenever necessary. Otherwise it is possible 4222109998Smarkm the caller may have overwritten (or deallocated) the original string 4223109998Smarkm data when a later ENGINE operation tries to use the stored values. 4224109998Smarkm [G�tz Babin-Ebell <babinebell@trustcenter.de>] 4225109998Smarkm 4226109998Smarkm *) Improve diagnostics in file reading and command-line digests. 4227109998Smarkm [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 4228109998Smarkm 4229109998Smarkm *) Add AES modes CFB and OFB to the object database. Correct an 4230109998Smarkm error in AES-CFB decryption. 4231109998Smarkm [Richard Levitte] 4232109998Smarkm 4233109998Smarkm *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 4234109998Smarkm allows existing EVP_CIPHER_CTX structures to be reused after 4235109998Smarkm calling EVP_*Final(). This behaviour is used by encryption 4236109998Smarkm BIOs and some applications. This has the side effect that 4237109998Smarkm applications must explicitly clean up cipher contexts with 4238109998Smarkm EVP_CIPHER_CTX_cleanup() or they will leak memory. 4239109998Smarkm [Steve Henson] 4240109998Smarkm 4241109998Smarkm *) Check the values of dna and dnb in bn_mul_recursive before calling 4242109998Smarkm bn_mul_comba (a non zero value means the a or b arrays do not contain 4243109998Smarkm n2 elements) and fallback to bn_mul_normal if either is not zero. 4244109998Smarkm [Steve Henson] 4245109998Smarkm 4246109998Smarkm *) Fix escaping of non-ASCII characters when using the -subj option 4247109998Smarkm of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 4248109998Smarkm [Lutz Jaenicke] 4249109998Smarkm 4250109998Smarkm *) Make object definitions compliant to LDAP (RFC2256): SN is the short 4251109998Smarkm form for "surname", serialNumber has no short form. 4252109998Smarkm Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 4253109998Smarkm therefore remove "mail" short name for "internet 7". 4254109998Smarkm The OID for unique identifiers in X509 certificates is 4255109998Smarkm x500UniqueIdentifier, not uniqueIdentifier. 4256109998Smarkm Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 4257109998Smarkm [Lutz Jaenicke] 4258109998Smarkm 4259109998Smarkm *) Add an "init" command to the ENGINE config module and auto initialize 4260109998Smarkm ENGINEs. Without any "init" command the ENGINE will be initialized 4261109998Smarkm after all ctrl commands have been executed on it. If init=1 the 4262109998Smarkm ENGINE is initailized at that point (ctrls before that point are run 4263109998Smarkm on the uninitialized ENGINE and after on the initialized one). If 4264109998Smarkm init=0 then the ENGINE will not be iniatialized at all. 4265109998Smarkm [Steve Henson] 4266109998Smarkm 4267109998Smarkm *) Fix the 'app_verify_callback' interface so that the user-defined 4268109998Smarkm argument is actually passed to the callback: In the 4269109998Smarkm SSL_CTX_set_cert_verify_callback() prototype, the callback 4270109998Smarkm declaration has been changed from 4271109998Smarkm int (*cb)() 4272109998Smarkm into 4273109998Smarkm int (*cb)(X509_STORE_CTX *,void *); 4274109998Smarkm in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 4275109998Smarkm i=s->ctx->app_verify_callback(&ctx) 4276109998Smarkm has been changed into 4277109998Smarkm i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 4278109998Smarkm 4279109998Smarkm To update applications using SSL_CTX_set_cert_verify_callback(), 4280109998Smarkm a dummy argument can be added to their callback functions. 4281109998Smarkm [D. K. Smetters <smetters@parc.xerox.com>] 4282109998Smarkm 4283109998Smarkm *) Added the '4758cca' ENGINE to support IBM 4758 cards. 4284109998Smarkm [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 4285109998Smarkm 4286109998Smarkm *) Add and OPENSSL_LOAD_CONF define which will cause 4287109998Smarkm OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 4288109998Smarkm This allows older applications to transparently support certain 4289109998Smarkm OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 4290109998Smarkm Two new functions OPENSSL_add_all_algorithms_noconf() which will never 4291109998Smarkm load the config file and OPENSSL_add_all_algorithms_conf() which will 4292109998Smarkm always load it have also been added. 4293109998Smarkm [Steve Henson] 4294109998Smarkm 4295109998Smarkm *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 4296109998Smarkm Adjust NIDs and EVP layer. 4297109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4298109998Smarkm 4299109998Smarkm *) Config modules support in openssl utility. 4300109998Smarkm 4301109998Smarkm Most commands now load modules from the config file, 4302109998Smarkm though in a few (such as version) this isn't done 4303109998Smarkm because it couldn't be used for anything. 4304109998Smarkm 4305109998Smarkm In the case of ca and req the config file used is 4306109998Smarkm the same as the utility itself: that is the -config 4307109998Smarkm command line option can be used to specify an 4308109998Smarkm alternative file. 4309109998Smarkm [Steve Henson] 4310109998Smarkm 4311109998Smarkm *) Move default behaviour from OPENSSL_config(). If appname is NULL 4312109998Smarkm use "openssl_conf" if filename is NULL use default openssl config file. 4313109998Smarkm [Steve Henson] 4314109998Smarkm 4315109998Smarkm *) Add an argument to OPENSSL_config() to allow the use of an alternative 4316109998Smarkm config section name. Add a new flag to tolerate a missing config file 4317109998Smarkm and move code to CONF_modules_load_file(). 4318109998Smarkm [Steve Henson] 4319109998Smarkm 4320109998Smarkm *) Support for crypto accelerator cards from Accelerated Encryption 4321109998Smarkm Processing, www.aep.ie. (Use engine 'aep') 4322109998Smarkm The support was copied from 0.9.6c [engine] and adapted/corrected 4323109998Smarkm to work with the new engine framework. 4324109998Smarkm [AEP Inc. and Richard Levitte] 4325109998Smarkm 4326109998Smarkm *) Support for SureWare crypto accelerator cards from Baltimore 4327109998Smarkm Technologies. (Use engine 'sureware') 4328109998Smarkm The support was copied from 0.9.6c [engine] and adapted 4329109998Smarkm to work with the new engine framework. 4330109998Smarkm [Richard Levitte] 4331109998Smarkm 4332109998Smarkm *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 4333109998Smarkm make the newer ENGINE framework commands for the CHIL engine work. 4334109998Smarkm [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 4335109998Smarkm 4336109998Smarkm *) Make it possible to produce shared libraries on ReliantUNIX. 4337109998Smarkm [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 4338109998Smarkm 4339109998Smarkm *) Add the configuration target debug-linux-ppro. 4340109998Smarkm Make 'openssl rsa' use the general key loading routines 4341109998Smarkm implemented in apps.c, and make those routines able to 4342109998Smarkm handle the key format FORMAT_NETSCAPE and the variant 4343109998Smarkm FORMAT_IISSGC. 4344109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4345109998Smarkm 4346109998Smarkm *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 4347109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4348109998Smarkm 4349109998Smarkm *) Add -keyform to rsautl, and document -engine. 4350109998Smarkm [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 4351109998Smarkm 4352109998Smarkm *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 4353109998Smarkm BIO_R_NO_SUCH_FILE error code rather than the generic 4354109998Smarkm ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 4355109998Smarkm [Ben Laurie] 4356109998Smarkm 4357109998Smarkm *) Add new functions 4358109998Smarkm ERR_peek_last_error 4359109998Smarkm ERR_peek_last_error_line 4360109998Smarkm ERR_peek_last_error_line_data. 4361109998Smarkm These are similar to 4362109998Smarkm ERR_peek_error 4363109998Smarkm ERR_peek_error_line 4364109998Smarkm ERR_peek_error_line_data, 4365109998Smarkm but report on the latest error recorded rather than the first one 4366109998Smarkm still in the error queue. 4367109998Smarkm [Ben Laurie, Bodo Moeller] 4368109998Smarkm 4369109998Smarkm *) default_algorithms option in ENGINE config module. This allows things 4370109998Smarkm like: 4371109998Smarkm default_algorithms = ALL 4372109998Smarkm default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 4373109998Smarkm [Steve Henson] 4374109998Smarkm 4375109998Smarkm *) Prelminary ENGINE config module. 4376109998Smarkm [Steve Henson] 4377109998Smarkm 4378109998Smarkm *) New experimental application configuration code. 4379109998Smarkm [Steve Henson] 4380109998Smarkm 4381109998Smarkm *) Change the AES code to follow the same name structure as all other 4382109998Smarkm symmetric ciphers, and behave the same way. Move everything to 4383109998Smarkm the directory crypto/aes, thereby obsoleting crypto/rijndael. 4384109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4385109998Smarkm 4386109998Smarkm *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 4387109998Smarkm [Ben Laurie and Theo de Raadt] 4388109998Smarkm 4389109998Smarkm *) Add option to output public keys in req command. 4390109998Smarkm [Massimiliano Pala madwolf@openca.org] 4391109998Smarkm 4392109998Smarkm *) Use wNAFs in EC_POINTs_mul() for improved efficiency 4393109998Smarkm (up to about 10% better than before for P-192 and P-224). 4394109998Smarkm [Bodo Moeller] 4395109998Smarkm 4396109998Smarkm *) New functions/macros 4397109998Smarkm 4398109998Smarkm SSL_CTX_set_msg_callback(ctx, cb) 4399109998Smarkm SSL_CTX_set_msg_callback_arg(ctx, arg) 4400109998Smarkm SSL_set_msg_callback(ssl, cb) 4401109998Smarkm SSL_set_msg_callback_arg(ssl, arg) 4402109998Smarkm 4403109998Smarkm to request calling a callback function 4404109998Smarkm 4405109998Smarkm void cb(int write_p, int version, int content_type, 4406109998Smarkm const void *buf, size_t len, SSL *ssl, void *arg) 4407109998Smarkm 4408109998Smarkm whenever a protocol message has been completely received 4409109998Smarkm (write_p == 0) or sent (write_p == 1). Here 'version' is the 4410109998Smarkm protocol version according to which the SSL library interprets 4411109998Smarkm the current protocol message (SSL2_VERSION, SSL3_VERSION, or 4412109998Smarkm TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 4413109998Smarkm the content type as defined in the SSL 3.0/TLS 1.0 protocol 4414109998Smarkm specification (change_cipher_spec(20), alert(21), handshake(22)). 4415109998Smarkm 'buf' and 'len' point to the actual message, 'ssl' to the 4416109998Smarkm SSL object, and 'arg' is the application-defined value set by 4417109998Smarkm SSL[_CTX]_set_msg_callback_arg(). 4418109998Smarkm 4419109998Smarkm 'openssl s_client' and 'openssl s_server' have new '-msg' options 4420109998Smarkm to enable a callback that displays all protocol messages. 4421109998Smarkm [Bodo Moeller] 4422109998Smarkm 4423109998Smarkm *) Change the shared library support so shared libraries are built as 4424109998Smarkm soon as the corresponding static library is finished, and thereby get 4425109998Smarkm openssl and the test programs linked against the shared library. 4426109998Smarkm This still only happens when the keyword "shard" has been given to 4427109998Smarkm the configuration scripts. 4428109998Smarkm 4429109998Smarkm NOTE: shared library support is still an experimental thing, and 4430109998Smarkm backward binary compatibility is still not guaranteed. 4431109998Smarkm ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 4432109998Smarkm 4433109998Smarkm *) Add support for Subject Information Access extension. 4434109998Smarkm [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 4435109998Smarkm 4436109998Smarkm *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 4437109998Smarkm additional bytes when new memory had to be allocated, not just 4438109998Smarkm when reusing an existing buffer. 4439109998Smarkm [Bodo Moeller] 4440109998Smarkm 4441109998Smarkm *) New command line and configuration option 'utf8' for the req command. 4442109998Smarkm This allows field values to be specified as UTF8 strings. 4443109998Smarkm [Steve Henson] 4444109998Smarkm 4445109998Smarkm *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 4446109998Smarkm runs for the former and machine-readable output for the latter. 4447109998Smarkm [Ben Laurie] 4448109998Smarkm 4449109998Smarkm *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 4450109998Smarkm of the e-mail address in the DN (i.e., it will go into a certificate 4451109998Smarkm extension only). The new configuration file option 'email_in_dn = no' 4452109998Smarkm has the same effect. 4453109998Smarkm [Massimiliano Pala madwolf@openca.org] 4454109998Smarkm 4455109998Smarkm *) Change all functions with names starting with des_ to be starting 4456109998Smarkm with DES_ instead. Add wrappers that are compatible with libdes, 4457109998Smarkm but are named _ossl_old_des_*. Finally, add macros that map the 4458109998Smarkm des_* symbols to the corresponding _ossl_old_des_* if libdes 4459109998Smarkm compatibility is desired. If OpenSSL 0.9.6c compatibility is 4460109998Smarkm desired, the des_* symbols will be mapped to DES_*, with one 4461109998Smarkm exception. 4462109998Smarkm 4463109998Smarkm Since we provide two compatibility mappings, the user needs to 4464109998Smarkm define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 4465109998Smarkm compatibility is desired. The default (i.e., when that macro 4466109998Smarkm isn't defined) is OpenSSL 0.9.6c compatibility. 4467109998Smarkm 4468109998Smarkm There are also macros that enable and disable the support of old 4469109998Smarkm des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 4470109998Smarkm and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 4471109998Smarkm are defined, the default will apply: to support the old des routines. 4472109998Smarkm 4473109998Smarkm In either case, one must include openssl/des.h to get the correct 4474109998Smarkm definitions. Do not try to just include openssl/des_old.h, that 4475109998Smarkm won't work. 4476109998Smarkm 4477109998Smarkm NOTE: This is a major break of an old API into a new one. Software 4478109998Smarkm authors are encouraged to switch to the DES_ style functions. Some 4479109998Smarkm time in the future, des_old.h and the libdes compatibility functions 4480109998Smarkm will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 4481109998Smarkm default), and then completely removed. 4482109998Smarkm [Richard Levitte] 4483109998Smarkm 4484109998Smarkm *) Test for certificates which contain unsupported critical extensions. 4485109998Smarkm If such a certificate is found during a verify operation it is 4486109998Smarkm rejected by default: this behaviour can be overridden by either 4487109998Smarkm handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 4488109998Smarkm by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 4489109998Smarkm X509_supported_extension() has also been added which returns 1 if a 4490109998Smarkm particular extension is supported. 4491109998Smarkm [Steve Henson] 4492109998Smarkm 4493109998Smarkm *) Modify the behaviour of EVP cipher functions in similar way to digests 4494109998Smarkm to retain compatibility with existing code. 4495109998Smarkm [Steve Henson] 4496109998Smarkm 4497109998Smarkm *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 4498109998Smarkm compatibility with existing code. In particular the 'ctx' parameter does 4499109998Smarkm not have to be to be initialized before the call to EVP_DigestInit() and 4500109998Smarkm it is tidied up after a call to EVP_DigestFinal(). New function 4501109998Smarkm EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 4502109998Smarkm EVP_MD_CTX_copy() changed to not require the destination to be 4503109998Smarkm initialized valid and new function EVP_MD_CTX_copy_ex() added which 4504109998Smarkm requires the destination to be valid. 4505109998Smarkm 4506109998Smarkm Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 4507109998Smarkm EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 4508109998Smarkm [Steve Henson] 4509109998Smarkm 4510109998Smarkm *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 4511109998Smarkm so that complete 'Handshake' protocol structures are kept in memory 4512109998Smarkm instead of overwriting 'msg_type' and 'length' with 'body' data. 4513109998Smarkm [Bodo Moeller] 4514109998Smarkm 4515109998Smarkm *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 4516109998Smarkm [Massimo Santin via Richard Levitte] 4517109998Smarkm 4518109998Smarkm *) Major restructuring to the underlying ENGINE code. This includes 4519109998Smarkm reduction of linker bloat, separation of pure "ENGINE" manipulation 4520109998Smarkm (initialisation, etc) from functionality dealing with implementations 4521109998Smarkm of specific crypto iterfaces. This change also introduces integrated 4522109998Smarkm support for symmetric ciphers and digest implementations - so ENGINEs 4523109998Smarkm can now accelerate these by providing EVP_CIPHER and EVP_MD 4524109998Smarkm implementations of their own. This is detailed in crypto/engine/README 4525109998Smarkm as it couldn't be adequately described here. However, there are a few 4526109998Smarkm API changes worth noting - some RSA, DSA, DH, and RAND functions that 4527109998Smarkm were changed in the original introduction of ENGINE code have now 4528109998Smarkm reverted back - the hooking from this code to ENGINE is now a good 4529109998Smarkm deal more passive and at run-time, operations deal directly with 4530109998Smarkm RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 4531109998Smarkm dereferencing through an ENGINE pointer any more. Also, the ENGINE 4532109998Smarkm functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 4533109998Smarkm they were not being used by the framework as there is no concept of a 4534109998Smarkm BIGNUM_METHOD and they could not be generalised to the new 4535109998Smarkm 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 4536109998Smarkm ENGINE_cpy() has been removed as it cannot be consistently defined in 4537109998Smarkm the new code. 4538109998Smarkm [Geoff Thorpe] 4539109998Smarkm 4540109998Smarkm *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 4541109998Smarkm [Steve Henson] 4542109998Smarkm 4543109998Smarkm *) Change mkdef.pl to sort symbols that get the same entry number, 4544109998Smarkm and make sure the automatically generated functions ERR_load_* 4545109998Smarkm become part of libeay.num as well. 4546109998Smarkm [Richard Levitte] 4547109998Smarkm 4548109998Smarkm *) New function SSL_renegotiate_pending(). This returns true once 4549109998Smarkm renegotiation has been requested (either SSL_renegotiate() call 4550109998Smarkm or HelloRequest/ClientHello receveived from the peer) and becomes 4551109998Smarkm false once a handshake has been completed. 4552109998Smarkm (For servers, SSL_renegotiate() followed by SSL_do_handshake() 4553109998Smarkm sends a HelloRequest, but does not ensure that a handshake takes 4554109998Smarkm place. SSL_renegotiate_pending() is useful for checking if the 4555109998Smarkm client has followed the request.) 4556109998Smarkm [Bodo Moeller] 4557109998Smarkm 4558109998Smarkm *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 4559109998Smarkm By default, clients may request session resumption even during 4560109998Smarkm renegotiation (if session ID contexts permit); with this option, 4561109998Smarkm session resumption is possible only in the first handshake. 4562109998Smarkm 4563109998Smarkm SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 4564109998Smarkm more bits available for options that should not be part of 4565109998Smarkm SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 4566109998Smarkm [Bodo Moeller] 4567109998Smarkm 4568109998Smarkm *) Add some demos for certificate and certificate request creation. 4569109998Smarkm [Steve Henson] 4570109998Smarkm 4571109998Smarkm *) Make maximum certificate chain size accepted from the peer application 4572109998Smarkm settable (SSL*_get/set_max_cert_list()), as proposed by 4573109998Smarkm "Douglas E. Engert" <deengert@anl.gov>. 4574109998Smarkm [Lutz Jaenicke] 4575109998Smarkm 4576109998Smarkm *) Add support for shared libraries for Unixware-7 4577109998Smarkm (Boyd Lynn Gerber <gerberb@zenez.com>). 4578109998Smarkm [Lutz Jaenicke] 4579109998Smarkm 4580109998Smarkm *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 4581109998Smarkm be done prior to destruction. Use this to unload error strings from 4582109998Smarkm ENGINEs that load their own error strings. NB: This adds two new API 4583109998Smarkm functions to "get" and "set" this destroy handler in an ENGINE. 4584109998Smarkm [Geoff Thorpe] 4585109998Smarkm 4586109998Smarkm *) Alter all existing ENGINE implementations (except "openssl" and 4587109998Smarkm "openbsd") to dynamically instantiate their own error strings. This 4588109998Smarkm makes them more flexible to be built both as statically-linked ENGINEs 4589109998Smarkm and self-contained shared-libraries loadable via the "dynamic" ENGINE. 4590109998Smarkm Also, add stub code to each that makes building them as self-contained 4591109998Smarkm shared-libraries easier (see README.ENGINE). 4592109998Smarkm [Geoff Thorpe] 4593109998Smarkm 4594109998Smarkm *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 4595109998Smarkm implementations into applications that are completely implemented in 4596109998Smarkm self-contained shared-libraries. The "dynamic" ENGINE exposes control 4597109998Smarkm commands that can be used to configure what shared-library to load and 4598109998Smarkm to control aspects of the way it is handled. Also, made an update to 4599109998Smarkm the README.ENGINE file that brings its information up-to-date and 4600109998Smarkm provides some information and instructions on the "dynamic" ENGINE 4601109998Smarkm (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 4602109998Smarkm [Geoff Thorpe] 4603109998Smarkm 4604109998Smarkm *) Make it possible to unload ranges of ERR strings with a new 4605109998Smarkm "ERR_unload_strings" function. 4606109998Smarkm [Geoff Thorpe] 4607109998Smarkm 4608109998Smarkm *) Add a copy() function to EVP_MD. 4609109998Smarkm [Ben Laurie] 4610109998Smarkm 4611109998Smarkm *) Make EVP_MD routines take a context pointer instead of just the 4612109998Smarkm md_data void pointer. 4613109998Smarkm [Ben Laurie] 4614109998Smarkm 4615109998Smarkm *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 4616109998Smarkm that the digest can only process a single chunk of data 4617109998Smarkm (typically because it is provided by a piece of 4618109998Smarkm hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 4619109998Smarkm is only going to provide a single chunk of data, and hence the 4620109998Smarkm framework needn't accumulate the data for oneshot drivers. 4621109998Smarkm [Ben Laurie] 4622109998Smarkm 4623109998Smarkm *) As with "ERR", make it possible to replace the underlying "ex_data" 4624109998Smarkm functions. This change also alters the storage and management of global 4625109998Smarkm ex_data state - it's now all inside ex_data.c and all "class" code (eg. 4626109998Smarkm RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 4627109998Smarkm index counters. The API functions that use this state have been changed 4628109998Smarkm to take a "class_index" rather than pointers to the class's local STACK 4629109998Smarkm and counter, and there is now an API function to dynamically create new 4630109998Smarkm classes. This centralisation allows us to (a) plug a lot of the 4631109998Smarkm thread-safety problems that existed, and (b) makes it possible to clean 4632109998Smarkm up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 4633109998Smarkm such data would previously have always leaked in application code and 4634109998Smarkm workarounds were in place to make the memory debugging turn a blind eye 4635109998Smarkm to it. Application code that doesn't use this new function will still 4636109998Smarkm leak as before, but their memory debugging output will announce it now 4637109998Smarkm rather than letting it slide. 4638109998Smarkm 4639109998Smarkm Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 4640109998Smarkm induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 4641109998Smarkm has a return value to indicate success or failure. 4642109998Smarkm [Geoff Thorpe] 4643109998Smarkm 4644109998Smarkm *) Make it possible to replace the underlying "ERR" functions such that the 4645109998Smarkm global state (2 LHASH tables and 2 locks) is only used by the "default" 4646109998Smarkm implementation. This change also adds two functions to "get" and "set" 4647109998Smarkm the implementation prior to it being automatically set the first time 4648109998Smarkm any other ERR function takes place. Ie. an application can call "get", 4649109998Smarkm pass the return value to a module it has just loaded, and that module 4650109998Smarkm can call its own "set" function using that value. This means the 4651109998Smarkm module's "ERR" operations will use (and modify) the error state in the 4652109998Smarkm application and not in its own statically linked copy of OpenSSL code. 4653109998Smarkm [Geoff Thorpe] 4654109998Smarkm 4655109998Smarkm *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 4656109998Smarkm reference counts. This performs normal REF_PRINT/REF_CHECK macros on 4657109998Smarkm the operation, and provides a more encapsulated way for external code 4658109998Smarkm (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 4659109998Smarkm to use these functions rather than manually incrementing the counts. 4660109998Smarkm 4661109998Smarkm Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 4662109998Smarkm [Geoff Thorpe] 4663109998Smarkm 4664109998Smarkm *) Add EVP test program. 4665109998Smarkm [Ben Laurie] 4666109998Smarkm 4667109998Smarkm *) Add symmetric cipher support to ENGINE. Expect the API to change! 4668109998Smarkm [Ben Laurie] 4669109998Smarkm 4670109998Smarkm *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 4671109998Smarkm X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 4672109998Smarkm X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 4673109998Smarkm These allow a CRL to be built without having to access X509_CRL fields 4674109998Smarkm directly. Modify 'ca' application to use new functions. 4675109998Smarkm [Steve Henson] 4676109998Smarkm 4677109998Smarkm *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 4678109998Smarkm bug workarounds. Rollback attack detection is a security feature. 4679109998Smarkm The problem will only arise on OpenSSL servers when TLSv1 is not 4680109998Smarkm available (sslv3_server_method() or SSL_OP_NO_TLSv1). 4681109998Smarkm Software authors not wanting to support TLSv1 will have special reasons 4682109998Smarkm for their choice and can explicitly enable this option. 4683109998Smarkm [Bodo Moeller, Lutz Jaenicke] 4684109998Smarkm 4685109998Smarkm *) Rationalise EVP so it can be extended: don't include a union of 4686109998Smarkm cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 4687109998Smarkm (similar to those existing for EVP_CIPHER_CTX). 4688109998Smarkm Usage example: 4689109998Smarkm 4690109998Smarkm EVP_MD_CTX md; 4691109998Smarkm 4692109998Smarkm EVP_MD_CTX_init(&md); /* new function call */ 4693109998Smarkm EVP_DigestInit(&md, EVP_sha1()); 4694109998Smarkm EVP_DigestUpdate(&md, in, len); 4695109998Smarkm EVP_DigestFinal(&md, out, NULL); 4696109998Smarkm EVP_MD_CTX_cleanup(&md); /* new function call */ 4697109998Smarkm 4698109998Smarkm [Ben Laurie] 4699109998Smarkm 4700109998Smarkm *) Make DES key schedule conform to the usual scheme, as well as 4701109998Smarkm correcting its structure. This means that calls to DES functions 4702109998Smarkm now have to pass a pointer to a des_key_schedule instead of a 4703109998Smarkm plain des_key_schedule (which was actually always a pointer 4704109998Smarkm anyway): E.g., 4705109998Smarkm 4706109998Smarkm des_key_schedule ks; 4707109998Smarkm 4708109998Smarkm des_set_key_checked(..., &ks); 4709109998Smarkm des_ncbc_encrypt(..., &ks, ...); 4710109998Smarkm 4711109998Smarkm (Note that a later change renames 'des_...' into 'DES_...'.) 4712109998Smarkm [Ben Laurie] 4713109998Smarkm 4714109998Smarkm *) Initial reduction of linker bloat: the use of some functions, such as 4715109998Smarkm PEM causes large amounts of unused functions to be linked in due to 4716109998Smarkm poor organisation. For example pem_all.c contains every PEM function 4717109998Smarkm which has a knock on effect of linking in large amounts of (unused) 4718109998Smarkm ASN1 code. Grouping together similar functions and splitting unrelated 4719109998Smarkm functions prevents this. 4720109998Smarkm [Steve Henson] 4721109998Smarkm 4722109998Smarkm *) Cleanup of EVP macros. 4723109998Smarkm [Ben Laurie] 4724109998Smarkm 4725109998Smarkm *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 4726109998Smarkm correct _ecb suffix. 4727109998Smarkm [Ben Laurie] 4728109998Smarkm 4729109998Smarkm *) Add initial OCSP responder support to ocsp application. The 4730109998Smarkm revocation information is handled using the text based index 4731109998Smarkm use by the ca application. The responder can either handle 4732109998Smarkm requests generated internally, supplied in files (for example 4733109998Smarkm via a CGI script) or using an internal minimal server. 4734109998Smarkm [Steve Henson] 4735109998Smarkm 4736109998Smarkm *) Add configuration choices to get zlib compression for TLS. 4737109998Smarkm [Richard Levitte] 4738109998Smarkm 4739109998Smarkm *) Changes to Kerberos SSL for RFC 2712 compliance: 4740109998Smarkm 1. Implemented real KerberosWrapper, instead of just using 4741109998Smarkm KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 4742109998Smarkm 2. Implemented optional authenticator field of KerberosWrapper. 4743109998Smarkm 4744109998Smarkm Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 4745109998Smarkm and authenticator structs; see crypto/krb5/. 4746109998Smarkm 4747109998Smarkm Generalized Kerberos calls to support multiple Kerberos libraries. 4748109998Smarkm [Vern Staats <staatsvr@asc.hpc.mil>, 4749109998Smarkm Jeffrey Altman <jaltman@columbia.edu> 4750109998Smarkm via Richard Levitte] 4751109998Smarkm 4752109998Smarkm *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 4753109998Smarkm already does with RSA. testdsa.h now has 'priv_key/pub_key' 4754109998Smarkm values for each of the key sizes rather than having just 4755109998Smarkm parameters (and 'speed' generating keys each time). 4756109998Smarkm [Geoff Thorpe] 4757109998Smarkm 4758109998Smarkm *) Speed up EVP routines. 4759109998Smarkm Before: 4760109998Smarkmencrypt 4761109998Smarkmtype 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 4762109998Smarkmdes-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 4763109998Smarkmdes-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 4764109998Smarkmdes-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 4765109998Smarkmdecrypt 4766109998Smarkmdes-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 4767109998Smarkmdes-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 4768109998Smarkmdes-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 4769109998Smarkm After: 4770109998Smarkmencrypt 4771109998Smarkmdes-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 4772109998Smarkmdecrypt 4773109998Smarkmdes-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 4774109998Smarkm [Ben Laurie] 4775109998Smarkm 4776109998Smarkm *) Added the OS2-EMX target. 4777109998Smarkm ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 4778109998Smarkm 4779109998Smarkm *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 4780109998Smarkm to support NCONF routines in extension code. New function CONF_set_nconf() 4781109998Smarkm to allow functions which take an NCONF to also handle the old LHASH 4782109998Smarkm structure: this means that the old CONF compatible routines can be 4783109998Smarkm retained (in particular wrt extensions) without having to duplicate the 4784109998Smarkm code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 4785109998Smarkm [Steve Henson] 4786109998Smarkm 4787109998Smarkm *) Enhance the general user interface with mechanisms for inner control 4788109998Smarkm and with possibilities to have yes/no kind of prompts. 4789109998Smarkm [Richard Levitte] 4790109998Smarkm 4791109998Smarkm *) Change all calls to low level digest routines in the library and 4792109998Smarkm applications to use EVP. Add missing calls to HMAC_cleanup() and 4793109998Smarkm don't assume HMAC_CTX can be copied using memcpy(). 4794109998Smarkm [Verdon Walker <VWalker@novell.com>, Steve Henson] 4795109998Smarkm 4796109998Smarkm *) Add the possibility to control engines through control names but with 4797109998Smarkm arbitrary arguments instead of just a string. 4798109998Smarkm Change the key loaders to take a UI_METHOD instead of a callback 4799109998Smarkm function pointer. NOTE: this breaks binary compatibility with earlier 4800109998Smarkm versions of OpenSSL [engine]. 4801109998Smarkm Adapt the nCipher code for these new conditions and add a card insertion 4802109998Smarkm callback. 4803109998Smarkm [Richard Levitte] 4804109998Smarkm 4805109998Smarkm *) Enhance the general user interface with mechanisms to better support 4806109998Smarkm dialog box interfaces, application-defined prompts, the possibility 4807109998Smarkm to use defaults (for example default passwords from somewhere else) 4808109998Smarkm and interrupts/cancellations. 4809109998Smarkm [Richard Levitte] 4810109998Smarkm 4811109998Smarkm *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 4812109998Smarkm attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 4813109998Smarkm [Steve Henson] 4814109998Smarkm 4815109998Smarkm *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 4816109998Smarkm tidy up some unnecessarily weird code in 'sk_new()'). 4817109998Smarkm [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 4818109998Smarkm 4819109998Smarkm *) Change the key loading routines for ENGINEs to use the same kind 4820109998Smarkm callback (pem_password_cb) as all other routines that need this 4821109998Smarkm kind of callback. 4822109998Smarkm [Richard Levitte] 4823109998Smarkm 4824109998Smarkm *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 4825109998Smarkm 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 4826109998Smarkm than this minimum value is recommended. 4827109998Smarkm [Lutz Jaenicke] 4828109998Smarkm 4829109998Smarkm *) New random seeder for OpenVMS, using the system process statistics 4830109998Smarkm that are easily reachable. 4831109998Smarkm [Richard Levitte] 4832109998Smarkm 4833109998Smarkm *) Windows apparently can't transparently handle global 4834109998Smarkm variables defined in DLLs. Initialisations such as: 4835109998Smarkm 4836109998Smarkm const ASN1_ITEM *it = &ASN1_INTEGER_it; 4837109998Smarkm 4838109998Smarkm wont compile. This is used by the any applications that need to 4839109998Smarkm declare their own ASN1 modules. This was fixed by adding the option 4840109998Smarkm EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 4841109998Smarkm needed for static libraries under Win32. 4842109998Smarkm [Steve Henson] 4843109998Smarkm 4844109998Smarkm *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 4845109998Smarkm setting of purpose and trust fields. New X509_STORE trust and 4846109998Smarkm purpose functions and tidy up setting in other SSL functions. 4847109998Smarkm [Steve Henson] 4848109998Smarkm 4849109998Smarkm *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 4850109998Smarkm structure. These are inherited by X509_STORE_CTX when it is 4851109998Smarkm initialised. This allows various defaults to be set in the 4852109998Smarkm X509_STORE structure (such as flags for CRL checking and custom 4853109998Smarkm purpose or trust settings) for functions which only use X509_STORE_CTX 4854109998Smarkm internally such as S/MIME. 4855109998Smarkm 4856109998Smarkm Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 4857109998Smarkm trust settings if they are not set in X509_STORE. This allows X509_STORE 4858109998Smarkm purposes and trust (in S/MIME for example) to override any set by default. 4859109998Smarkm 4860109998Smarkm Add command line options for CRL checking to smime, s_client and s_server 4861109998Smarkm applications. 4862109998Smarkm [Steve Henson] 4863109998Smarkm 4864109998Smarkm *) Initial CRL based revocation checking. If the CRL checking flag(s) 4865109998Smarkm are set then the CRL is looked up in the X509_STORE structure and 4866109998Smarkm its validity and signature checked, then if the certificate is found 4867109998Smarkm in the CRL the verify fails with a revoked error. 4868109998Smarkm 4869109998Smarkm Various new CRL related callbacks added to X509_STORE_CTX structure. 4870109998Smarkm 4871109998Smarkm Command line options added to 'verify' application to support this. 4872109998Smarkm 4873109998Smarkm This needs some additional work, such as being able to handle multiple 4874109998Smarkm CRLs with different times, extension based lookup (rather than just 4875109998Smarkm by subject name) and ultimately more complete V2 CRL extension 4876109998Smarkm handling. 4877109998Smarkm [Steve Henson] 4878109998Smarkm 4879109998Smarkm *) Add a general user interface API (crypto/ui/). This is designed 4880109998Smarkm to replace things like des_read_password and friends (backward 4881109998Smarkm compatibility functions using this new API are provided). 4882109998Smarkm The purpose is to remove prompting functions from the DES code 4883109998Smarkm section as well as provide for prompting through dialog boxes in 4884109998Smarkm a window system and the like. 4885109998Smarkm [Richard Levitte] 4886109998Smarkm 4887109998Smarkm *) Add "ex_data" support to ENGINE so implementations can add state at a 4888109998Smarkm per-structure level rather than having to store it globally. 4889109998Smarkm [Geoff] 4890109998Smarkm 4891109998Smarkm *) Make it possible for ENGINE structures to be copied when retrieved by 4892109998Smarkm ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 4893109998Smarkm This causes the "original" ENGINE structure to act like a template, 4894109998Smarkm analogous to the RSA vs. RSA_METHOD type of separation. Because of this 4895109998Smarkm operational state can be localised to each ENGINE structure, despite the 4896109998Smarkm fact they all share the same "methods". New ENGINE structures returned in 4897109998Smarkm this case have no functional references and the return value is the single 4898109998Smarkm structural reference. This matches the single structural reference returned 4899109998Smarkm by ENGINE_by_id() normally, when it is incremented on the pre-existing 4900109998Smarkm ENGINE structure. 4901109998Smarkm [Geoff] 4902109998Smarkm 4903109998Smarkm *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 4904109998Smarkm needs to match any other type at all we need to manually clear the 4905109998Smarkm tag cache. 4906109998Smarkm [Steve Henson] 4907109998Smarkm 4908109998Smarkm *) Changes to the "openssl engine" utility to include; 4909109998Smarkm - verbosity levels ('-v', '-vv', and '-vvv') that provide information 4910109998Smarkm about an ENGINE's available control commands. 4911109998Smarkm - executing control commands from command line arguments using the 4912109998Smarkm '-pre' and '-post' switches. '-post' is only used if '-t' is 4913109998Smarkm specified and the ENGINE is successfully initialised. The syntax for 4914109998Smarkm the individual commands are colon-separated, for example; 4915109998Smarkm openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 4916109998Smarkm [Geoff] 4917109998Smarkm 4918109998Smarkm *) New dynamic control command support for ENGINEs. ENGINEs can now 4919109998Smarkm declare their own commands (numbers), names (strings), descriptions, 4920109998Smarkm and input types for run-time discovery by calling applications. A 4921109998Smarkm subset of these commands are implicitly classed as "executable" 4922109998Smarkm depending on their input type, and only these can be invoked through 4923109998Smarkm the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 4924109998Smarkm can be based on user input, config files, etc). The distinction is 4925109998Smarkm that "executable" commands cannot return anything other than a boolean 4926109998Smarkm result and can only support numeric or string input, whereas some 4927109998Smarkm discoverable commands may only be for direct use through 4928109998Smarkm ENGINE_ctrl(), eg. supporting the exchange of binary data, function 4929109998Smarkm pointers, or other custom uses. The "executable" commands are to 4930109998Smarkm support parameterisations of ENGINE behaviour that can be 4931109998Smarkm unambiguously defined by ENGINEs and used consistently across any 4932109998Smarkm OpenSSL-based application. Commands have been added to all the 4933109998Smarkm existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 4934109998Smarkm control over shared-library paths without source code alterations. 4935109998Smarkm [Geoff] 4936109998Smarkm 4937109998Smarkm *) Changed all ENGINE implementations to dynamically allocate their 4938109998Smarkm ENGINEs rather than declaring them statically. Apart from this being 4939109998Smarkm necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 4940109998Smarkm this also allows the implementations to compile without using the 4941109998Smarkm internal engine_int.h header. 4942109998Smarkm [Geoff] 4943109998Smarkm 4944109998Smarkm *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 4945109998Smarkm 'const' value. Any code that should be able to modify a RAND_METHOD 4946109998Smarkm should already have non-const pointers to it (ie. they should only 4947109998Smarkm modify their own ones). 4948109998Smarkm [Geoff] 4949109998Smarkm 4950109998Smarkm *) Made a variety of little tweaks to the ENGINE code. 4951109998Smarkm - "atalla" and "ubsec" string definitions were moved from header files 4952109998Smarkm to C code. "nuron" string definitions were placed in variables 4953109998Smarkm rather than hard-coded - allowing parameterisation of these values 4954109998Smarkm later on via ctrl() commands. 4955109998Smarkm - Removed unused "#if 0"'d code. 4956109998Smarkm - Fixed engine list iteration code so it uses ENGINE_free() to release 4957109998Smarkm structural references. 4958109998Smarkm - Constified the RAND_METHOD element of ENGINE structures. 4959109998Smarkm - Constified various get/set functions as appropriate and added 4960109998Smarkm missing functions (including a catch-all ENGINE_cpy that duplicates 4961109998Smarkm all ENGINE values onto a new ENGINE except reference counts/state). 4962109998Smarkm - Removed NULL parameter checks in get/set functions. Setting a method 4963109998Smarkm or function to NULL is a way of cancelling out a previously set 4964109998Smarkm value. Passing a NULL ENGINE parameter is just plain stupid anyway 4965109998Smarkm and doesn't justify the extra error symbols and code. 4966109998Smarkm - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 4967109998Smarkm flags from engine_int.h to engine.h. 4968109998Smarkm - Changed prototypes for ENGINE handler functions (init(), finish(), 4969109998Smarkm ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 4970109998Smarkm [Geoff] 4971109998Smarkm 4972109998Smarkm *) Implement binary inversion algorithm for BN_mod_inverse in addition 4973109998Smarkm to the algorithm using long division. The binary algorithm can be 4974109998Smarkm used only if the modulus is odd. On 32-bit systems, it is faster 4975109998Smarkm only for relatively small moduli (roughly 20-30% for 128-bit moduli, 4976109998Smarkm roughly 5-15% for 256-bit moduli), so we use it only for moduli 4977109998Smarkm up to 450 bits. In 64-bit environments, the binary algorithm 4978109998Smarkm appears to be advantageous for much longer moduli; here we use it 4979109998Smarkm for moduli up to 2048 bits. 4980109998Smarkm [Bodo Moeller] 4981109998Smarkm 4982109998Smarkm *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 4983109998Smarkm could not support the combine flag in choice fields. 4984109998Smarkm [Steve Henson] 4985109998Smarkm 4986109998Smarkm *) Add a 'copy_extensions' option to the 'ca' utility. This copies 4987109998Smarkm extensions from a certificate request to the certificate. 4988109998Smarkm [Steve Henson] 4989109998Smarkm 4990109998Smarkm *) Allow multiple 'certopt' and 'nameopt' options to be separated 4991109998Smarkm by commas. Add 'namopt' and 'certopt' options to the 'ca' config 4992109998Smarkm file: this allows the display of the certificate about to be 4993109998Smarkm signed to be customised, to allow certain fields to be included 4994109998Smarkm or excluded and extension details. The old system didn't display 4995109998Smarkm multicharacter strings properly, omitted fields not in the policy 4996109998Smarkm and couldn't display additional details such as extensions. 4997109998Smarkm [Steve Henson] 4998109998Smarkm 4999109998Smarkm *) Function EC_POINTs_mul for multiple scalar multiplication 5000109998Smarkm of an arbitrary number of elliptic curve points 5001109998Smarkm \sum scalars[i]*points[i], 5002109998Smarkm optionally including the generator defined for the EC_GROUP: 5003109998Smarkm scalar*generator + \sum scalars[i]*points[i]. 5004109998Smarkm 5005109998Smarkm EC_POINT_mul is a simple wrapper function for the typical case 5006109998Smarkm that the point list has just one item (besides the optional 5007109998Smarkm generator). 5008109998Smarkm [Bodo Moeller] 5009109998Smarkm 5010109998Smarkm *) First EC_METHODs for curves over GF(p): 5011109998Smarkm 5012109998Smarkm EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 5013109998Smarkm operations and provides various method functions that can also 5014109998Smarkm operate with faster implementations of modular arithmetic. 5015109998Smarkm 5016109998Smarkm EC_GFp_mont_method() reuses most functions that are part of 5017109998Smarkm EC_GFp_simple_method, but uses Montgomery arithmetic. 5018109998Smarkm 5019109998Smarkm [Bodo Moeller; point addition and point doubling 5020109998Smarkm implementation directly derived from source code provided by 5021109998Smarkm Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 5022109998Smarkm 5023109998Smarkm *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 5024109998Smarkm crypto/ec/ec_lib.c): 5025109998Smarkm 5026109998Smarkm Curves are EC_GROUP objects (with an optional group generator) 5027109998Smarkm based on EC_METHODs that are built into the library. 5028109998Smarkm 5029109998Smarkm Points are EC_POINT objects based on EC_GROUP objects. 5030109998Smarkm 5031109998Smarkm Most of the framework would be able to handle curves over arbitrary 5032109998Smarkm finite fields, but as there are no obvious types for fields other 5033109998Smarkm than GF(p), some functions are limited to that for now. 5034109998Smarkm [Bodo Moeller] 5035109998Smarkm 5036109998Smarkm *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 5037109998Smarkm that the file contains a complete HTTP response. 5038109998Smarkm [Richard Levitte] 5039109998Smarkm 5040109998Smarkm *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 5041109998Smarkm change the def and num file printf format specifier from "%-40sXXX" 5042109998Smarkm to "%-39s XXX". The latter will always guarantee a space after the 5043109998Smarkm field while the former will cause them to run together if the field 5044109998Smarkm is 40 of more characters long. 5045109998Smarkm [Steve Henson] 5046109998Smarkm 5047109998Smarkm *) Constify the cipher and digest 'method' functions and structures 5048109998Smarkm and modify related functions to take constant EVP_MD and EVP_CIPHER 5049109998Smarkm pointers. 5050109998Smarkm [Steve Henson] 5051109998Smarkm 5052109998Smarkm *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 5053109998Smarkm in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 5054109998Smarkm [Bodo Moeller] 5055109998Smarkm 5056109998Smarkm *) Modify EVP_Digest*() routines so they now return values. Although the 5057109998Smarkm internal software routines can never fail additional hardware versions 5058109998Smarkm might. 5059109998Smarkm [Steve Henson] 5060109998Smarkm 5061109998Smarkm *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 5062109998Smarkm 5063109998Smarkm Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 5064109998Smarkm (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 5065109998Smarkm 5066109998Smarkm ASN1 error codes 5067109998Smarkm ERR_R_NESTED_ASN1_ERROR 5068109998Smarkm ... 5069109998Smarkm ERR_R_MISSING_ASN1_EOS 5070109998Smarkm were 4 .. 9, conflicting with 5071109998Smarkm ERR_LIB_RSA (= ERR_R_RSA_LIB) 5072109998Smarkm ... 5073109998Smarkm ERR_LIB_PEM (= ERR_R_PEM_LIB). 5074109998Smarkm They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 5075109998Smarkm 5076109998Smarkm Add new error code 'ERR_R_INTERNAL_ERROR'. 5077109998Smarkm [Bodo Moeller] 5078109998Smarkm 5079109998Smarkm *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 5080109998Smarkm suffices. 5081109998Smarkm [Bodo Moeller] 5082109998Smarkm 5083109998Smarkm *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 5084109998Smarkm sets the subject name for a new request or supersedes the 5085109998Smarkm subject name in a given request. Formats that can be parsed are 5086109998Smarkm 'CN=Some Name, OU=myOU, C=IT' 5087109998Smarkm and 5088109998Smarkm 'CN=Some Name/OU=myOU/C=IT'. 5089109998Smarkm 5090109998Smarkm Add options '-batch' and '-verbose' to 'openssl req'. 5091109998Smarkm [Massimiliano Pala <madwolf@hackmasters.net>] 5092109998Smarkm 5093109998Smarkm *) Introduce the possibility to access global variables through 5094109998Smarkm functions on platform were that's the best way to handle exporting 5095109998Smarkm global variables in shared libraries. To enable this functionality, 5096109998Smarkm one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 5097109998Smarkm "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 5098109998Smarkm is normally done by Configure or something similar). 5099109998Smarkm 5100109998Smarkm To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 5101109998Smarkm in the source file (foo.c) like this: 5102109998Smarkm 5103109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 5104109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(double,bar); 5105109998Smarkm 5106109998Smarkm To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 5107109998Smarkm and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 5108109998Smarkm 5109109998Smarkm OPENSSL_DECLARE_GLOBAL(int,foo); 5110109998Smarkm #define foo OPENSSL_GLOBAL_REF(foo) 5111109998Smarkm OPENSSL_DECLARE_GLOBAL(double,bar); 5112109998Smarkm #define bar OPENSSL_GLOBAL_REF(bar) 5113109998Smarkm 5114109998Smarkm The #defines are very important, and therefore so is including the 5115109998Smarkm header file everywhere where the defined globals are used. 5116109998Smarkm 5117109998Smarkm The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 5118109998Smarkm of ASN.1 items, but that structure is a bit different. 5119109998Smarkm 5120109998Smarkm The largest change is in util/mkdef.pl which has been enhanced with 5121109998Smarkm better and easier to understand logic to choose which symbols should 5122109998Smarkm go into the Windows .def files as well as a number of fixes and code 5123109998Smarkm cleanup (among others, algorithm keywords are now sorted 5124109998Smarkm lexicographically to avoid constant rewrites). 5125109998Smarkm [Richard Levitte] 5126109998Smarkm 5127109998Smarkm *) In BN_div() keep a copy of the sign of 'num' before writing the 5128109998Smarkm result to 'rm' because if rm==num the value will be overwritten 5129109998Smarkm and produce the wrong result if 'num' is negative: this caused 5130109998Smarkm problems with BN_mod() and BN_nnmod(). 5131109998Smarkm [Steve Henson] 5132109998Smarkm 5133109998Smarkm *) Function OCSP_request_verify(). This checks the signature on an 5134109998Smarkm OCSP request and verifies the signer certificate. The signer 5135109998Smarkm certificate is just checked for a generic purpose and OCSP request 5136109998Smarkm trust settings. 5137109998Smarkm [Steve Henson] 5138109998Smarkm 5139109998Smarkm *) Add OCSP_check_validity() function to check the validity of OCSP 5140109998Smarkm responses. OCSP responses are prepared in real time and may only 5141109998Smarkm be a few seconds old. Simply checking that the current time lies 5142109998Smarkm between thisUpdate and nextUpdate max reject otherwise valid responses 5143109998Smarkm caused by either OCSP responder or client clock inaccuracy. Instead 5144109998Smarkm we allow thisUpdate and nextUpdate to fall within a certain period of 5145109998Smarkm the current time. The age of the response can also optionally be 5146109998Smarkm checked. Two new options -validity_period and -status_age added to 5147109998Smarkm ocsp utility. 5148109998Smarkm [Steve Henson] 5149109998Smarkm 5150109998Smarkm *) If signature or public key algorithm is unrecognized print out its 5151109998Smarkm OID rather that just UNKNOWN. 5152109998Smarkm [Steve Henson] 5153109998Smarkm 5154109998Smarkm *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 5155109998Smarkm OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 5156109998Smarkm ID to be generated from the issuer certificate alone which can then be 5157109998Smarkm passed to OCSP_id_issuer_cmp(). 5158109998Smarkm [Steve Henson] 5159109998Smarkm 5160109998Smarkm *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 5161109998Smarkm ASN1 modules to export functions returning ASN1_ITEM pointers 5162109998Smarkm instead of the ASN1_ITEM structures themselves. This adds several 5163109998Smarkm new macros which allow the underlying ASN1 function/structure to 5164109998Smarkm be accessed transparently. As a result code should not use ASN1_ITEM 5165109998Smarkm references directly (such as &X509_it) but instead use the relevant 5166109998Smarkm macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 5167109998Smarkm use of the new ASN1 code on platforms where exporting structures 5168109998Smarkm is problematical (for example in shared libraries) but exporting 5169109998Smarkm functions returning pointers to structures is not. 5170109998Smarkm [Steve Henson] 5171109998Smarkm 5172109998Smarkm *) Add support for overriding the generation of SSL/TLS session IDs. 5173109998Smarkm These callbacks can be registered either in an SSL_CTX or per SSL. 5174109998Smarkm The purpose of this is to allow applications to control, if they wish, 5175109998Smarkm the arbitrary values chosen for use as session IDs, particularly as it 5176109998Smarkm can be useful for session caching in multiple-server environments. A 5177109998Smarkm command-line switch for testing this (and any client code that wishes 5178109998Smarkm to use such a feature) has been added to "s_server". 5179109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 5180109998Smarkm 5181109998Smarkm *) Modify mkdef.pl to recognise and parse preprocessor conditionals 5182109998Smarkm of the form '#if defined(...) || defined(...) || ...' and 5183109998Smarkm '#if !defined(...) && !defined(...) && ...'. This also avoids 5184109998Smarkm the growing number of special cases it was previously handling. 5185109998Smarkm [Richard Levitte] 5186109998Smarkm 5187109998Smarkm *) Make all configuration macros available for application by making 5188109998Smarkm sure they are available in opensslconf.h, by giving them names starting 5189109998Smarkm with "OPENSSL_" to avoid conflicts with other packages and by making 5190109998Smarkm sure e_os2.h will cover all platform-specific cases together with 5191109998Smarkm opensslconf.h. 5192109998Smarkm Additionally, it is now possible to define configuration/platform- 5193109998Smarkm specific names (called "system identities"). In the C code, these 5194109998Smarkm are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 5195109998Smarkm macro with the name beginning with "OPENSSL_SYS_", which is determined 5196109998Smarkm from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 5197109998Smarkm what is available. 5198109998Smarkm [Richard Levitte] 5199109998Smarkm 5200109998Smarkm *) New option -set_serial to 'req' and 'x509' this allows the serial 5201109998Smarkm number to use to be specified on the command line. Previously self 5202109998Smarkm signed certificates were hard coded with serial number 0 and the 5203109998Smarkm CA options of 'x509' had to use a serial number in a file which was 5204109998Smarkm auto incremented. 5205109998Smarkm [Steve Henson] 5206109998Smarkm 5207109998Smarkm *) New options to 'ca' utility to support V2 CRL entry extensions. 5208109998Smarkm Currently CRL reason, invalidity date and hold instruction are 5209109998Smarkm supported. Add new CRL extensions to V3 code and some new objects. 5210109998Smarkm [Steve Henson] 5211109998Smarkm 5212109998Smarkm *) New function EVP_CIPHER_CTX_set_padding() this is used to 5213109998Smarkm disable standard block padding (aka PKCS#5 padding) in the EVP 5214109998Smarkm API, which was previously mandatory. This means that the data is 5215109998Smarkm not padded in any way and so the total length much be a multiple 5216109998Smarkm of the block size, otherwise an error occurs. 5217109998Smarkm [Steve Henson] 5218109998Smarkm 5219109998Smarkm *) Initial (incomplete) OCSP SSL support. 5220109998Smarkm [Steve Henson] 5221109998Smarkm 5222109998Smarkm *) New function OCSP_parse_url(). This splits up a URL into its host, 5223109998Smarkm port and path components: primarily to parse OCSP URLs. New -url 5224109998Smarkm option to ocsp utility. 5225109998Smarkm [Steve Henson] 5226109998Smarkm 5227109998Smarkm *) New nonce behavior. The return value of OCSP_check_nonce() now 5228109998Smarkm reflects the various checks performed. Applications can decide 5229109998Smarkm whether to tolerate certain situations such as an absent nonce 5230109998Smarkm in a response when one was present in a request: the ocsp application 5231109998Smarkm just prints out a warning. New function OCSP_add1_basic_nonce() 5232109998Smarkm this is to allow responders to include a nonce in a response even if 5233109998Smarkm the request is nonce-less. 5234109998Smarkm [Steve Henson] 5235109998Smarkm 5236109998Smarkm *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 5237109998Smarkm skipped when using openssl x509 multiple times on a single input file, 5238109998Smarkm e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 5239109998Smarkm [Bodo Moeller] 5240109998Smarkm 5241109998Smarkm *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 5242109998Smarkm set string type: to handle setting ASN1_TIME structures. Fix ca 5243109998Smarkm utility to correctly initialize revocation date of CRLs. 5244109998Smarkm [Steve Henson] 5245109998Smarkm 5246109998Smarkm *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 5247109998Smarkm the clients preferred ciphersuites and rather use its own preferences. 5248109998Smarkm Should help to work around M$ SGC (Server Gated Cryptography) bug in 5249109998Smarkm Internet Explorer by ensuring unchanged hash method during stepup. 5250109998Smarkm (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 5251109998Smarkm [Lutz Jaenicke] 5252109998Smarkm 5253109998Smarkm *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 5254109998Smarkm to aes and add a new 'exist' option to print out symbols that don't 5255109998Smarkm appear to exist. 5256109998Smarkm [Steve Henson] 5257109998Smarkm 5258109998Smarkm *) Additional options to ocsp utility to allow flags to be set and 5259109998Smarkm additional certificates supplied. 5260109998Smarkm [Steve Henson] 5261109998Smarkm 5262109998Smarkm *) Add the option -VAfile to 'openssl ocsp', so the user can give the 5263109998Smarkm OCSP client a number of certificate to only verify the response 5264109998Smarkm signature against. 5265109998Smarkm [Richard Levitte] 5266109998Smarkm 5267109998Smarkm *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 5268109998Smarkm handle the new API. Currently only ECB, CBC modes supported. Add new 5269109998Smarkm AES OIDs. 5270109998Smarkm 5271109998Smarkm Add TLS AES ciphersuites as described in RFC3268, "Advanced 5272109998Smarkm Encryption Standard (AES) Ciphersuites for Transport Layer 5273109998Smarkm Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 5274109998Smarkm not enabled by default and were not part of the "ALL" ciphersuite 5275109998Smarkm alias because they were not yet official; they could be 5276109998Smarkm explicitly requested by specifying the "AESdraft" ciphersuite 5277109998Smarkm group alias. In the final release of OpenSSL 0.9.7, the group 5278109998Smarkm alias is called "AES" and is part of "ALL".) 5279109998Smarkm [Ben Laurie, Steve Henson, Bodo Moeller] 5280109998Smarkm 5281109998Smarkm *) New function OCSP_copy_nonce() to copy nonce value (if present) from 5282109998Smarkm request to response. 5283109998Smarkm [Steve Henson] 5284109998Smarkm 5285109998Smarkm *) Functions for OCSP responders. OCSP_request_onereq_count(), 5286109998Smarkm OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 5287109998Smarkm extract information from a certificate request. OCSP_response_create() 5288109998Smarkm creates a response and optionally adds a basic response structure. 5289109998Smarkm OCSP_basic_add1_status() adds a complete single response to a basic 5290109998Smarkm response and returns the OCSP_SINGLERESP structure just added (to allow 5291109998Smarkm extensions to be included for example). OCSP_basic_add1_cert() adds a 5292109998Smarkm certificate to a basic response and OCSP_basic_sign() signs a basic 5293109998Smarkm response with various flags. New helper functions ASN1_TIME_check() 5294109998Smarkm (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 5295109998Smarkm (converts ASN1_TIME to GeneralizedTime). 5296109998Smarkm [Steve Henson] 5297109998Smarkm 5298109998Smarkm *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 5299109998Smarkm in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 5300109998Smarkm structure from a certificate. X509_pubkey_digest() digests the public_key 5301109998Smarkm contents: this is used in various key identifiers. 5302109998Smarkm [Steve Henson] 5303109998Smarkm 5304109998Smarkm *) Make sk_sort() tolerate a NULL argument. 5305109998Smarkm [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 5306109998Smarkm 5307109998Smarkm *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 5308109998Smarkm passed by the function are trusted implicitly. If any of them signed the 5309109998Smarkm response then it is assumed to be valid and is not verified. 5310109998Smarkm [Steve Henson] 5311109998Smarkm 5312109998Smarkm *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 5313109998Smarkm to data. This was previously part of the PKCS7 ASN1 code. This 5314109998Smarkm was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 5315109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 5316109998Smarkm <support@securenetterm.com>] 5317109998Smarkm 5318109998Smarkm *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 5319109998Smarkm routines: without these tracing memory leaks is very painful. 5320109998Smarkm Fix leaks in PKCS12 and PKCS7 routines. 5321109998Smarkm [Steve Henson] 5322109998Smarkm 5323109998Smarkm *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 5324109998Smarkm Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 5325109998Smarkm effectively meant GeneralizedTime would never be used. Now it 5326109998Smarkm is initialised to -1 but X509_time_adj() now has to check the value 5327109998Smarkm and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 5328109998Smarkm V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 5329109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 5330109998Smarkm <support@securenetterm.com>] 5331109998Smarkm 5332109998Smarkm *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 5333109998Smarkm result in a zero length in the ASN1_INTEGER structure which was 5334109998Smarkm not consistent with the structure when d2i_ASN1_INTEGER() was used 5335109998Smarkm and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 5336109998Smarkm to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 5337109998Smarkm where it did not print out a minus for negative ASN1_INTEGER. 5338109998Smarkm [Steve Henson] 5339109998Smarkm 5340109998Smarkm *) Add summary printout to ocsp utility. The various functions which 5341109998Smarkm convert status values to strings have been renamed to: 5342109998Smarkm OCSP_response_status_str(), OCSP_cert_status_str() and 5343109998Smarkm OCSP_crl_reason_str() and are no longer static. New options 5344109998Smarkm to verify nonce values and to disable verification. OCSP response 5345109998Smarkm printout format cleaned up. 5346109998Smarkm [Steve Henson] 5347109998Smarkm 5348109998Smarkm *) Add additional OCSP certificate checks. These are those specified 5349109998Smarkm in RFC2560. This consists of two separate checks: the CA of the 5350109998Smarkm certificate being checked must either be the OCSP signer certificate 5351109998Smarkm or the issuer of the OCSP signer certificate. In the latter case the 5352109998Smarkm OCSP signer certificate must contain the OCSP signing extended key 5353109998Smarkm usage. This check is performed by attempting to match the OCSP 5354109998Smarkm signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 5355109998Smarkm in the OCSP_CERTID structures of the response. 5356109998Smarkm [Steve Henson] 5357109998Smarkm 5358109998Smarkm *) Initial OCSP certificate verification added to OCSP_basic_verify() 5359109998Smarkm and related routines. This uses the standard OpenSSL certificate 5360109998Smarkm verify routines to perform initial checks (just CA validity) and 5361109998Smarkm to obtain the certificate chain. Then additional checks will be 5362109998Smarkm performed on the chain. Currently the root CA is checked to see 5363109998Smarkm if it is explicitly trusted for OCSP signing. This is used to set 5364109998Smarkm a root CA as a global signing root: that is any certificate that 5365109998Smarkm chains to that CA is an acceptable OCSP signing certificate. 5366109998Smarkm [Steve Henson] 5367109998Smarkm 5368109998Smarkm *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 5369109998Smarkm extensions from a separate configuration file. 5370109998Smarkm As when reading extensions from the main configuration file, 5371109998Smarkm the '-extensions ...' option may be used for specifying the 5372109998Smarkm section to use. 5373109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 5374109998Smarkm 5375109998Smarkm *) New OCSP utility. Allows OCSP requests to be generated or 5376109998Smarkm read. The request can be sent to a responder and the output 5377109998Smarkm parsed, outputed or printed in text form. Not complete yet: 5378109998Smarkm still needs to check the OCSP response validity. 5379109998Smarkm [Steve Henson] 5380109998Smarkm 5381109998Smarkm *) New subcommands for 'openssl ca': 5382109998Smarkm 'openssl ca -status <serial>' prints the status of the cert with 5383109998Smarkm the given serial number (according to the index file). 5384109998Smarkm 'openssl ca -updatedb' updates the expiry status of certificates 5385109998Smarkm in the index file. 5386109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 5387109998Smarkm 5388109998Smarkm *) New '-newreq-nodes' command option to CA.pl. This is like 5389109998Smarkm '-newreq', but calls 'openssl req' with the '-nodes' option 5390109998Smarkm so that the resulting key is not encrypted. 5391109998Smarkm [Damien Miller <djm@mindrot.org>] 5392109998Smarkm 5393109998Smarkm *) New configuration for the GNU Hurd. 5394109998Smarkm [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 5395109998Smarkm 5396109998Smarkm *) Initial code to implement OCSP basic response verify. This 5397109998Smarkm is currently incomplete. Currently just finds the signer's 5398109998Smarkm certificate and verifies the signature on the response. 5399109998Smarkm [Steve Henson] 5400109998Smarkm 5401109998Smarkm *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 5402109998Smarkm value of OPENSSLDIR. This is available via the new '-d' option 5403109998Smarkm to 'openssl version', and is also included in 'openssl version -a'. 5404109998Smarkm [Bodo Moeller] 5405109998Smarkm 5406109998Smarkm *) Allowing defining memory allocation callbacks that will be given 5407109998Smarkm file name and line number information in additional arguments 5408109998Smarkm (a const char* and an int). The basic functionality remains, as 5409109998Smarkm well as the original possibility to just replace malloc(), 5410109998Smarkm realloc() and free() by functions that do not know about these 5411109998Smarkm additional arguments. To register and find out the current 5412109998Smarkm settings for extended allocation functions, the following 5413109998Smarkm functions are provided: 5414109998Smarkm 5415109998Smarkm CRYPTO_set_mem_ex_functions 5416109998Smarkm CRYPTO_set_locked_mem_ex_functions 5417109998Smarkm CRYPTO_get_mem_ex_functions 5418109998Smarkm CRYPTO_get_locked_mem_ex_functions 5419109998Smarkm 5420109998Smarkm These work the same way as CRYPTO_set_mem_functions and friends. 5421109998Smarkm CRYPTO_get_[locked_]mem_functions now writes 0 where such an 5422109998Smarkm extended allocation function is enabled. 5423109998Smarkm Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 5424109998Smarkm a conventional allocation function is enabled. 5425109998Smarkm [Richard Levitte, Bodo Moeller] 5426109998Smarkm 5427109998Smarkm *) Finish off removing the remaining LHASH function pointer casts. 5428109998Smarkm There should no longer be any prototype-casting required when using 5429109998Smarkm the LHASH abstraction, and any casts that remain are "bugs". See 5430109998Smarkm the callback types and macros at the head of lhash.h for details 5431109998Smarkm (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 5432109998Smarkm [Geoff Thorpe] 5433109998Smarkm 5434109998Smarkm *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 5435109998Smarkm If /dev/[u]random devices are not available or do not return enough 5436109998Smarkm entropy, EGD style sockets (served by EGD or PRNGD) will automatically 5437109998Smarkm be queried. 5438109998Smarkm The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 5439109998Smarkm /etc/entropy will be queried once each in this sequence, quering stops 5440109998Smarkm when enough entropy was collected without querying more sockets. 5441109998Smarkm [Lutz Jaenicke] 5442109998Smarkm 5443109998Smarkm *) Change the Unix RAND_poll() variant to be able to poll several 5444109998Smarkm random devices, as specified by DEVRANDOM, until a sufficient amount 5445109998Smarkm of data has been collected. We spend at most 10 ms on each file 5446109998Smarkm (select timeout) and read in non-blocking mode. DEVRANDOM now 5447109998Smarkm defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 5448109998Smarkm (previously it was just the string "/dev/urandom"), so on typical 5449109998Smarkm platforms the 10 ms delay will never occur. 5450109998Smarkm Also separate out the Unix variant to its own file, rand_unix.c. 5451109998Smarkm For VMS, there's a currently-empty rand_vms.c. 5452109998Smarkm [Richard Levitte] 5453109998Smarkm 5454109998Smarkm *) Move OCSP client related routines to ocsp_cl.c. These 5455109998Smarkm provide utility functions which an application needing 5456109998Smarkm to issue a request to an OCSP responder and analyse the 5457109998Smarkm response will typically need: as opposed to those which an 5458109998Smarkm OCSP responder itself would need which will be added later. 5459109998Smarkm 5460109998Smarkm OCSP_request_sign() signs an OCSP request with an API similar 5461109998Smarkm to PKCS7_sign(). OCSP_response_status() returns status of OCSP 5462109998Smarkm response. OCSP_response_get1_basic() extracts basic response 5463109998Smarkm from response. OCSP_resp_find_status(): finds and extracts status 5464109998Smarkm information from an OCSP_CERTID structure (which will be created 5465109998Smarkm when the request structure is built). These are built from lower 5466109998Smarkm level functions which work on OCSP_SINGLERESP structures but 5467109998Smarkm wont normally be used unless the application wishes to examine 5468109998Smarkm extensions in the OCSP response for example. 5469109998Smarkm 5470109998Smarkm Replace nonce routines with a pair of functions. 5471109998Smarkm OCSP_request_add1_nonce() adds a nonce value and optionally 5472109998Smarkm generates a random value. OCSP_check_nonce() checks the 5473109998Smarkm validity of the nonce in an OCSP response. 5474109998Smarkm [Steve Henson] 5475109998Smarkm 5476109998Smarkm *) Change function OCSP_request_add() to OCSP_request_add0_id(). 5477109998Smarkm This doesn't copy the supplied OCSP_CERTID and avoids the 5478109998Smarkm need to free up the newly created id. Change return type 5479109998Smarkm to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 5480109998Smarkm This can then be used to add extensions to the request. 5481109998Smarkm Deleted OCSP_request_new(), since most of its functionality 5482109998Smarkm is now in OCSP_REQUEST_new() (and the case insensitive name 5483109998Smarkm clash) apart from the ability to set the request name which 5484109998Smarkm will be added elsewhere. 5485109998Smarkm [Steve Henson] 5486109998Smarkm 5487109998Smarkm *) Update OCSP API. Remove obsolete extensions argument from 5488109998Smarkm various functions. Extensions are now handled using the new 5489109998Smarkm OCSP extension code. New simple OCSP HTTP function which 5490109998Smarkm can be used to send requests and parse the response. 5491109998Smarkm [Steve Henson] 5492109998Smarkm 5493109998Smarkm *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 5494109998Smarkm ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 5495109998Smarkm uses the special reorder version of SET OF to sort the attributes 5496109998Smarkm and reorder them to match the encoded order. This resolves a long 5497109998Smarkm standing problem: a verify on a PKCS7 structure just after signing 5498109998Smarkm it used to fail because the attribute order did not match the 5499109998Smarkm encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 5500109998Smarkm it uses the received order. This is necessary to tolerate some broken 5501109998Smarkm software that does not order SET OF. This is handled by encoding 5502109998Smarkm as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 5503109998Smarkm to produce the required SET OF. 5504109998Smarkm [Steve Henson] 5505109998Smarkm 5506109998Smarkm *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 5507109998Smarkm OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 5508109998Smarkm files to get correct declarations of the ASN.1 item variables. 5509109998Smarkm [Richard Levitte] 5510109998Smarkm 5511109998Smarkm *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 5512109998Smarkm PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 5513109998Smarkm asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 5514109998Smarkm NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 5515109998Smarkm New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 5516109998Smarkm ASN1_ITEM and no wrapper functions. 5517109998Smarkm [Steve Henson] 5518109998Smarkm 5519109998Smarkm *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 5520109998Smarkm replace the old function pointer based I/O routines. Change most of 5521109998Smarkm the *_d2i_bio() and *_d2i_fp() functions to use these. 5522109998Smarkm [Steve Henson] 5523109998Smarkm 5524109998Smarkm *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 5525109998Smarkm lines, recognice more "algorithms" that can be deselected, and make 5526109998Smarkm it complain about algorithm deselection that isn't recognised. 5527109998Smarkm [Richard Levitte] 5528109998Smarkm 5529109998Smarkm *) New ASN1 functions to handle dup, sign, verify, digest, pack and 5530109998Smarkm unpack operations in terms of ASN1_ITEM. Modify existing wrappers 5531109998Smarkm to use new functions. Add NO_ASN1_OLD which can be set to remove 5532109998Smarkm some old style ASN1 functions: this can be used to determine if old 5533109998Smarkm code will still work when these eventually go away. 5534109998Smarkm [Steve Henson] 5535109998Smarkm 5536109998Smarkm *) New extension functions for OCSP structures, these follow the 5537109998Smarkm same conventions as certificates and CRLs. 5538109998Smarkm [Steve Henson] 5539109998Smarkm 5540109998Smarkm *) New function X509V3_add1_i2d(). This automatically encodes and 5541109998Smarkm adds an extension. Its behaviour can be customised with various 5542109998Smarkm flags to append, replace or delete. Various wrappers added for 5543109998Smarkm certifcates and CRLs. 5544109998Smarkm [Steve Henson] 5545109998Smarkm 5546109998Smarkm *) Fix to avoid calling the underlying ASN1 print routine when 5547109998Smarkm an extension cannot be parsed. Correct a typo in the 5548109998Smarkm OCSP_SERVICELOC extension. Tidy up print OCSP format. 5549109998Smarkm [Steve Henson] 5550109998Smarkm 5551109998Smarkm *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 5552109998Smarkm entries for variables. 5553109998Smarkm [Steve Henson] 5554109998Smarkm 5555109998Smarkm *) Add functionality to apps/openssl.c for detecting locking 5556109998Smarkm problems: As the program is single-threaded, all we have 5557109998Smarkm to do is register a locking callback using an array for 5558109998Smarkm storing which locks are currently held by the program. 5559109998Smarkm [Bodo Moeller] 5560109998Smarkm 5561109998Smarkm *) Use a lock around the call to CRYPTO_get_ex_new_index() in 5562109998Smarkm SSL_get_ex_data_X509_STORE_idx(), which is used in 5563109998Smarkm ssl_verify_cert_chain() and thus can be called at any time 5564109998Smarkm during TLS/SSL handshakes so that thread-safety is essential. 5565109998Smarkm Unfortunately, the ex_data design is not at all suited 5566109998Smarkm for multi-threaded use, so it probably should be abolished. 5567109998Smarkm [Bodo Moeller] 5568109998Smarkm 5569109998Smarkm *) Added Broadcom "ubsec" ENGINE to OpenSSL. 5570109998Smarkm [Broadcom, tweaked and integrated by Geoff Thorpe] 5571109998Smarkm 5572109998Smarkm *) Move common extension printing code to new function 5573109998Smarkm X509V3_print_extensions(). Reorganise OCSP print routines and 5574109998Smarkm implement some needed OCSP ASN1 functions. Add OCSP extensions. 5575109998Smarkm [Steve Henson] 5576109998Smarkm 5577109998Smarkm *) New function X509_signature_print() to remove duplication in some 5578109998Smarkm print routines. 5579109998Smarkm [Steve Henson] 5580109998Smarkm 5581109998Smarkm *) Add a special meaning when SET OF and SEQUENCE OF flags are both 5582109998Smarkm set (this was treated exactly the same as SET OF previously). This 5583109998Smarkm is used to reorder the STACK representing the structure to match the 5584109998Smarkm encoding. This will be used to get round a problem where a PKCS7 5585109998Smarkm structure which was signed could not be verified because the STACK 5586109998Smarkm order did not reflect the encoded order. 5587109998Smarkm [Steve Henson] 5588109998Smarkm 5589109998Smarkm *) Reimplement the OCSP ASN1 module using the new code. 5590109998Smarkm [Steve Henson] 5591109998Smarkm 5592109998Smarkm *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 5593109998Smarkm for its ASN1 operations. The old style function pointers still exist 5594109998Smarkm for now but they will eventually go away. 5595109998Smarkm [Steve Henson] 5596109998Smarkm 5597109998Smarkm *) Merge in replacement ASN1 code from the ASN1 branch. This almost 5598109998Smarkm completely replaces the old ASN1 functionality with a table driven 5599109998Smarkm encoder and decoder which interprets an ASN1_ITEM structure describing 5600109998Smarkm the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 5601109998Smarkm largely maintained. Almost all of the old asn1_mac.h macro based ASN1 5602109998Smarkm has also been converted to the new form. 5603109998Smarkm [Steve Henson] 5604109998Smarkm 5605109998Smarkm *) Change BN_mod_exp_recp so that negative moduli are tolerated 5606109998Smarkm (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 5607109998Smarkm so that BN_mod_exp_mont and BN_mod_exp_mont_word work 5608109998Smarkm for negative moduli. 5609109998Smarkm [Bodo Moeller] 5610109998Smarkm 5611109998Smarkm *) Fix BN_uadd and BN_usub: Always return non-negative results instead 5612109998Smarkm of not touching the result's sign bit. 5613109998Smarkm [Bodo Moeller] 5614109998Smarkm 5615109998Smarkm *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 5616109998Smarkm set. 5617109998Smarkm [Bodo Moeller] 5618109998Smarkm 5619109998Smarkm *) Changed the LHASH code to use prototypes for callbacks, and created 5620109998Smarkm macros to declare and implement thin (optionally static) functions 5621109998Smarkm that provide type-safety and avoid function pointer casting for the 5622109998Smarkm type-specific callbacks. 5623109998Smarkm [Geoff Thorpe] 5624109998Smarkm 5625109998Smarkm *) Added Kerberos Cipher Suites to be used with TLS, as written in 5626109998Smarkm RFC 2712. 5627109998Smarkm [Veers Staats <staatsvr@asc.hpc.mil>, 5628109998Smarkm Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 5629109998Smarkm 5630109998Smarkm *) Reformat the FAQ so the different questions and answers can be divided 5631109998Smarkm in sections depending on the subject. 5632109998Smarkm [Richard Levitte] 5633109998Smarkm 5634109998Smarkm *) Have the zlib compression code load ZLIB.DLL dynamically under 5635109998Smarkm Windows. 5636109998Smarkm [Richard Levitte] 5637109998Smarkm 5638109998Smarkm *) New function BN_mod_sqrt for computing square roots modulo a prime 5639109998Smarkm (using the probabilistic Tonelli-Shanks algorithm unless 5640109998Smarkm p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 5641109998Smarkm be handled deterministically). 5642109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5643109998Smarkm 5644109998Smarkm *) Make BN_mod_inverse faster by explicitly handling small quotients 5645109998Smarkm in the Euclid loop. (Speed gain about 20% for small moduli [256 or 5646109998Smarkm 512 bits], about 30% for larger ones [1024 or 2048 bits].) 5647109998Smarkm [Bodo Moeller] 5648109998Smarkm 5649109998Smarkm *) New function BN_kronecker. 5650109998Smarkm [Bodo Moeller] 5651109998Smarkm 5652109998Smarkm *) Fix BN_gcd so that it works on negative inputs; the result is 5653109998Smarkm positive unless both parameters are zero. 5654109998Smarkm Previously something reasonably close to an infinite loop was 5655109998Smarkm possible because numbers could be growing instead of shrinking 5656109998Smarkm in the implementation of Euclid's algorithm. 5657109998Smarkm [Bodo Moeller] 5658109998Smarkm 5659109998Smarkm *) Fix BN_is_word() and BN_is_one() macros to take into account the 5660109998Smarkm sign of the number in question. 5661109998Smarkm 5662109998Smarkm Fix BN_is_word(a,w) to work correctly for w == 0. 5663109998Smarkm 5664109998Smarkm The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 5665109998Smarkm because its test if the absolute value of 'a' equals 'w'. 5666109998Smarkm Note that BN_abs_is_word does *not* handle w == 0 reliably; 5667109998Smarkm it exists mostly for use in the implementations of BN_is_zero(), 5668109998Smarkm BN_is_one(), and BN_is_word(). 5669109998Smarkm [Bodo Moeller] 5670109998Smarkm 5671109998Smarkm *) New function BN_swap. 5672109998Smarkm [Bodo Moeller] 5673109998Smarkm 5674109998Smarkm *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 5675109998Smarkm the exponentiation functions are more likely to produce reasonable 5676109998Smarkm results on negative inputs. 5677109998Smarkm [Bodo Moeller] 5678109998Smarkm 5679109998Smarkm *) Change BN_mod_mul so that the result is always non-negative. 5680109998Smarkm Previously, it could be negative if one of the factors was negative; 5681109998Smarkm I don't think anyone really wanted that behaviour. 5682109998Smarkm [Bodo Moeller] 5683109998Smarkm 5684109998Smarkm *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 5685109998Smarkm (except for exponentiation, which stays in crypto/bn/bn_exp.c, 5686109998Smarkm and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 5687109998Smarkm and add new functions: 5688109998Smarkm 5689109998Smarkm BN_nnmod 5690109998Smarkm BN_mod_sqr 5691109998Smarkm BN_mod_add 5692109998Smarkm BN_mod_add_quick 5693109998Smarkm BN_mod_sub 5694109998Smarkm BN_mod_sub_quick 5695109998Smarkm BN_mod_lshift1 5696109998Smarkm BN_mod_lshift1_quick 5697109998Smarkm BN_mod_lshift 5698109998Smarkm BN_mod_lshift_quick 5699109998Smarkm 5700109998Smarkm These functions always generate non-negative results. 5701109998Smarkm 5702109998Smarkm BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 5703109998Smarkm such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 5704109998Smarkm 5705109998Smarkm BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 5706109998Smarkm BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 5707109998Smarkm be reduced modulo m. 5708109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5709109998Smarkm 5710111147Snectar#if 0 5711111147Snectar The following entry accidentily appeared in the CHANGES file 5712111147Snectar distributed with OpenSSL 0.9.7. The modifications described in 5713111147Snectar it do *not* apply to OpenSSL 0.9.7. 5714111147Snectar 5715109998Smarkm *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 5716109998Smarkm was actually never needed) and in BN_mul(). The removal in BN_mul() 5717109998Smarkm required a small change in bn_mul_part_recursive() and the addition 5718109998Smarkm of the functions bn_cmp_part_words(), bn_sub_part_words() and 5719109998Smarkm bn_add_part_words(), which do the same thing as bn_cmp_words(), 5720109998Smarkm bn_sub_words() and bn_add_words() except they take arrays with 5721109998Smarkm differing sizes. 5722109998Smarkm [Richard Levitte] 5723111147Snectar#endif 5724109998Smarkm 5725109998Smarkm *) In 'openssl passwd', verify passwords read from the terminal 5726109998Smarkm unless the '-salt' option is used (which usually means that 5727109998Smarkm verification would just waste user's time since the resulting 5728109998Smarkm hash is going to be compared with some given password hash) 5729109998Smarkm or the new '-noverify' option is used. 5730109998Smarkm 5731109998Smarkm This is an incompatible change, but it does not affect 5732109998Smarkm non-interactive use of 'openssl passwd' (passwords on the command 5733109998Smarkm line, '-stdin' option, '-in ...' option) and thus should not 5734109998Smarkm cause any problems. 5735109998Smarkm [Bodo Moeller] 5736109998Smarkm 5737109998Smarkm *) Remove all references to RSAref, since there's no more need for it. 5738109998Smarkm [Richard Levitte] 5739109998Smarkm 5740109998Smarkm *) Make DSO load along a path given through an environment variable 5741109998Smarkm (SHLIB_PATH) with shl_load(). 5742109998Smarkm [Richard Levitte] 5743109998Smarkm 5744109998Smarkm *) Constify the ENGINE code as a result of BIGNUM constification. 5745109998Smarkm Also constify the RSA code and most things related to it. In a 5746109998Smarkm few places, most notable in the depth of the ASN.1 code, ugly 5747109998Smarkm casts back to non-const were required (to be solved at a later 5748109998Smarkm time) 5749109998Smarkm [Richard Levitte] 5750109998Smarkm 5751109998Smarkm *) Make it so the openssl application has all engines loaded by default. 5752109998Smarkm [Richard Levitte] 5753109998Smarkm 5754109998Smarkm *) Constify the BIGNUM routines a little more. 5755109998Smarkm [Richard Levitte] 5756109998Smarkm 5757109998Smarkm *) Add the following functions: 5758109998Smarkm 5759109998Smarkm ENGINE_load_cswift() 5760109998Smarkm ENGINE_load_chil() 5761109998Smarkm ENGINE_load_atalla() 5762109998Smarkm ENGINE_load_nuron() 5763109998Smarkm ENGINE_load_builtin_engines() 5764109998Smarkm 5765109998Smarkm That way, an application can itself choose if external engines that 5766109998Smarkm are built-in in OpenSSL shall ever be used or not. The benefit is 5767109998Smarkm that applications won't have to be linked with libdl or other dso 5768109998Smarkm libraries unless it's really needed. 5769109998Smarkm 5770109998Smarkm Changed 'openssl engine' to load all engines on demand. 5771109998Smarkm Changed the engine header files to avoid the duplication of some 5772109998Smarkm declarations (they differed!). 5773109998Smarkm [Richard Levitte] 5774109998Smarkm 5775109998Smarkm *) 'openssl engine' can now list capabilities. 5776109998Smarkm [Richard Levitte] 5777109998Smarkm 5778109998Smarkm *) Better error reporting in 'openssl engine'. 5779109998Smarkm [Richard Levitte] 5780109998Smarkm 5781109998Smarkm *) Never call load_dh_param(NULL) in s_server. 5782109998Smarkm [Bodo Moeller] 5783109998Smarkm 5784109998Smarkm *) Add engine application. It can currently list engines by name and 5785109998Smarkm identity, and test if they are actually available. 5786109998Smarkm [Richard Levitte] 5787109998Smarkm 5788109998Smarkm *) Improve RPM specification file by forcing symbolic linking and making 5789109998Smarkm sure the installed documentation is also owned by root.root. 5790109998Smarkm [Damien Miller <djm@mindrot.org>] 5791109998Smarkm 5792109998Smarkm *) Give the OpenSSL applications more possibilities to make use of 5793109998Smarkm keys (public as well as private) handled by engines. 5794109998Smarkm [Richard Levitte] 5795109998Smarkm 5796109998Smarkm *) Add OCSP code that comes from CertCo. 5797109998Smarkm [Richard Levitte] 5798109998Smarkm 5799109998Smarkm *) Add VMS support for the Rijndael code. 5800109998Smarkm [Richard Levitte] 5801109998Smarkm 5802109998Smarkm *) Added untested support for Nuron crypto accelerator. 5803109998Smarkm [Ben Laurie] 5804109998Smarkm 5805109998Smarkm *) Add support for external cryptographic devices. This code was 5806109998Smarkm previously distributed separately as the "engine" branch. 5807109998Smarkm [Geoff Thorpe, Richard Levitte] 5808109998Smarkm 5809109998Smarkm *) Rework the filename-translation in the DSO code. It is now possible to 5810109998Smarkm have far greater control over how a "name" is turned into a filename 5811109998Smarkm depending on the operating environment and any oddities about the 5812109998Smarkm different shared library filenames on each system. 5813109998Smarkm [Geoff Thorpe] 5814109998Smarkm 5815109998Smarkm *) Support threads on FreeBSD-elf in Configure. 5816109998Smarkm [Richard Levitte] 5817109998Smarkm 5818109998Smarkm *) Fix for SHA1 assembly problem with MASM: it produces 5819109998Smarkm warnings about corrupt line number information when assembling 5820109998Smarkm with debugging information. This is caused by the overlapping 5821109998Smarkm of two sections. 5822109998Smarkm [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 5823109998Smarkm 5824109998Smarkm *) NCONF changes. 5825109998Smarkm NCONF_get_number() has no error checking at all. As a replacement, 5826109998Smarkm NCONF_get_number_e() is defined (_e for "error checking") and is 5827109998Smarkm promoted strongly. The old NCONF_get_number is kept around for 5828109998Smarkm binary backward compatibility. 5829109998Smarkm Make it possible for methods to load from something other than a BIO, 5830109998Smarkm by providing a function pointer that is given a name instead of a BIO. 5831109998Smarkm For example, this could be used to load configuration data from an 5832109998Smarkm LDAP server. 5833109998Smarkm [Richard Levitte] 5834109998Smarkm 5835109998Smarkm *) Fix for non blocking accept BIOs. Added new I/O special reason 5836109998Smarkm BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 5837109998Smarkm with non blocking I/O was not possible because no retry code was 5838109998Smarkm implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 5839109998Smarkm this case. 5840109998Smarkm [Steve Henson] 5841109998Smarkm 5842109998Smarkm *) Added the beginnings of Rijndael support. 5843109998Smarkm [Ben Laurie] 5844109998Smarkm 5845109998Smarkm *) Fix for bug in DirectoryString mask setting. Add support for 5846109998Smarkm X509_NAME_print_ex() in 'req' and X509_print_ex() function 5847109998Smarkm to allow certificate printing to more controllable, additional 5848109998Smarkm 'certopt' option to 'x509' to allow new printing options to be 5849109998Smarkm set. 5850109998Smarkm [Steve Henson] 5851109998Smarkm 5852109998Smarkm *) Clean old EAY MD5 hack from e_os.h. 5853109998Smarkm [Richard Levitte] 5854109998Smarkm 5855142425Snectar Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 5856142425Snectar 5857142425Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 5858160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 5859142425Snectar [Joe Orton, Steve Henson] 5860142425Snectar 5861142425Snectar Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 5862142425Snectar 5863142425Snectar *) Fix additional bug revealed by the NISCC test suite: 5864142425Snectar 5865142425Snectar Stop bug triggering large recursion when presented with 5866160814Ssimon certain ASN.1 tags (CVE-2003-0851) 5867142425Snectar [Steve Henson] 5868142425Snectar 5869120631Snectar Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 5870120631Snectar 5871120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 5872120631Snectar 5873120631Snectar Stop out of bounds reads in the ASN1 code when presented with 5874160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 5875120631Snectar 5876120631Snectar If verify callback ignores invalid public key errors don't try to check 5877120631Snectar certificate signature with the NULL public key. 5878120631Snectar 5879120631Snectar [Steve Henson] 5880120631Snectar 5881120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 5882120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 5883120631Snectar specifications. 5884120631Snectar [Steve Henson] 5885120631Snectar 5886120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 5887120631Snectar extra data after the compression methods not only for TLS 1.0 5888120631Snectar but also for SSL 3.0 (as required by the specification). 5889120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 5890120631Snectar 5891120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 5892120631Snectar when it's 512 *bits* long, not 512 bytes. 5893120631Snectar [Richard Levitte] 5894120631Snectar 5895120631Snectar Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 5896120631Snectar 5897120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 5898120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 5899120631Snectar a protocol version number mismatch like a decryption error 5900120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 5901120631Snectar [Bodo Moeller] 5902120631Snectar 5903120631Snectar *) Turn on RSA blinding by default in the default implementation 5904120631Snectar to avoid a timing attack. Applications that don't want it can call 5905120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 5906120631Snectar They would be ill-advised to do so in most cases. 5907120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 5908120631Snectar 5909120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 5910120631Snectar seeded (in this case, the secret RSA exponent is abused as 5911120631Snectar an unpredictable seed -- if it is not unpredictable, there 5912120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 5913120631Snectar by remembering the creator's thread ID in rsa->blinding and 5914120631Snectar having all other threads use local one-time blinding factors 5915120631Snectar (this requires more computation than sharing rsa->blinding, but 5916120631Snectar avoids excessive locking; and if an RSA object is not shared 5917120631Snectar between threads, blinding will still be very fast). 5918120631Snectar [Bodo Moeller] 5919120631Snectar 5920111147Snectar Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 5921111147Snectar 5922111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 5923111147Snectar via timing by performing a MAC computation even if incorrrect 5924111147Snectar block cipher padding has been found. This is a countermeasure 5925111147Snectar against active attacks where the attacker has to distinguish 5926160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 5927111147Snectar 5928111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 5929111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 5930111147Snectar Martin Vuagnoux (EPFL, Ilion)] 5931111147Snectar 5932109998Smarkm Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 5933109998Smarkm 5934109998Smarkm *) New function OPENSSL_cleanse(), which is used to cleanse a section of 5935109998Smarkm memory from it's contents. This is done with a counter that will 5936109998Smarkm place alternating values in each byte. This can be used to solve 5937109998Smarkm two issues: 1) the removal of calls to memset() by highly optimizing 5938109998Smarkm compilers, and 2) cleansing with other values than 0, since those can 5939109998Smarkm be read through on certain media, for example a swap space on disk. 5940109998Smarkm [Geoff Thorpe] 5941109998Smarkm 5942109998Smarkm *) Bugfix: client side session caching did not work with external caching, 5943109998Smarkm because the session->cipher setting was not restored when reloading 5944109998Smarkm from the external cache. This problem was masked, when 5945109998Smarkm SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 5946109998Smarkm (Found by Steve Haslam <steve@araqnid.ddts.net>.) 5947109998Smarkm [Lutz Jaenicke] 5948109998Smarkm 5949109998Smarkm *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 5950109998Smarkm length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 5951109998Smarkm [Zeev Lieber <zeev-l@yahoo.com>] 5952109998Smarkm 5953109998Smarkm *) Undo an undocumented change introduced in 0.9.6e which caused 5954109998Smarkm repeated calls to OpenSSL_add_all_ciphers() and 5955109998Smarkm OpenSSL_add_all_digests() to be ignored, even after calling 5956109998Smarkm EVP_cleanup(). 5957109998Smarkm [Richard Levitte] 5958109998Smarkm 5959109998Smarkm *) Change the default configuration reader to deal with last line not 5960109998Smarkm being properly terminated. 5961109998Smarkm [Richard Levitte] 5962109998Smarkm 5963109998Smarkm *) Change X509_NAME_cmp() so it applies the special rules on handling 5964109998Smarkm DN values that are of type PrintableString, as well as RDNs of type 5965109998Smarkm emailAddress where the value has the type ia5String. 5966109998Smarkm [stefank@valicert.com via Richard Levitte] 5967109998Smarkm 5968109998Smarkm *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 5969109998Smarkm the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 5970109998Smarkm doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 5971109998Smarkm the bitwise-OR of the two for use by the majority of applications 5972109998Smarkm wanting this behaviour, and update the docs. The documented 5973109998Smarkm behaviour and actual behaviour were inconsistent and had been 5974109998Smarkm changing anyway, so this is more a bug-fix than a behavioural 5975109998Smarkm change. 5976109998Smarkm [Geoff Thorpe, diagnosed by Nadav Har'El] 5977109998Smarkm 5978109998Smarkm *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 5979109998Smarkm (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 5980109998Smarkm [Bodo Moeller] 5981109998Smarkm 5982109998Smarkm *) Fix initialization code race conditions in 5983109998Smarkm SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 5984109998Smarkm SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 5985109998Smarkm SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 5986109998Smarkm TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 5987109998Smarkm ssl2_get_cipher_by_char(), 5988109998Smarkm ssl3_get_cipher_by_char(). 5989109998Smarkm [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 5990109998Smarkm 5991109998Smarkm *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 5992109998Smarkm the cached sessions are flushed, as the remove_cb() might use ex_data 5993109998Smarkm contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 5994109998Smarkm (see [openssl.org #212]). 5995109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 5996109998Smarkm 5997109998Smarkm *) Fix typo in OBJ_txt2obj which incorrectly passed the content 5998109998Smarkm length, instead of the encoding length to d2i_ASN1_OBJECT. 5999109998Smarkm [Steve Henson] 6000109998Smarkm 6001101618Snectar Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 6002101618Snectar 6003101618Snectar *) [In 0.9.6g-engine release:] 6004101618Snectar Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 6005101618Snectar [Lynn Gazis <lgazis@rainbow.com>] 6006101618Snectar 6007101613Snectar Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 6008101613Snectar 6009101613Snectar *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 6010101613Snectar and get fix the header length calculation. 6011101613Snectar [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 6012101613Snectar Alon Kantor <alonk@checkpoint.com> (and others), 6013101613Snectar Steve Henson] 6014101613Snectar 6015101613Snectar *) Use proper error handling instead of 'assertions' in buffer 6016101613Snectar overflow checks added in 0.9.6e. This prevents DoS (the 6017101613Snectar assertions could call abort()). 6018101613Snectar [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 6019101613Snectar 6020100936Snectar Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 6021100936Snectar 6022109998Smarkm *) Add various sanity checks to asn1_get_length() to reject 6023109998Smarkm the ASN1 length bytes if they exceed sizeof(long), will appear 6024109998Smarkm negative or the content length exceeds the length of the 6025109998Smarkm supplied buffer. 6026109998Smarkm [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6027109998Smarkm 6028100936Snectar *) Fix cipher selection routines: ciphers without encryption had no flags 6029100936Snectar for the cipher strength set and where therefore not handled correctly 6030100936Snectar by the selection routines (PR #130). 6031100936Snectar [Lutz Jaenicke] 6032100936Snectar 6033100936Snectar *) Fix EVP_dsa_sha macro. 6034100936Snectar [Nils Larsch] 6035100936Snectar 6036100936Snectar *) New option 6037100936Snectar SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 6038100936Snectar for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 6039100936Snectar that was added in OpenSSL 0.9.6d. 6040100936Snectar 6041100936Snectar As the countermeasure turned out to be incompatible with some 6042100936Snectar broken SSL implementations, the new option is part of SSL_OP_ALL. 6043100936Snectar SSL_OP_ALL is usually employed when compatibility with weird SSL 6044100936Snectar implementations is desired (e.g. '-bugs' option to 's_client' and 6045100936Snectar 's_server'), so the new option is automatically set in many 6046100936Snectar applications. 6047100936Snectar [Bodo Moeller] 6048100936Snectar 6049100936Snectar *) Changes in security patch: 6050100936Snectar 6051100936Snectar Changes marked "(CHATS)" were sponsored by the Defense Advanced 6052100936Snectar Research Projects Agency (DARPA) and Air Force Research Laboratory, 6053100936Snectar Air Force Materiel Command, USAF, under agreement number 6054100936Snectar F30602-01-2-0537. 6055100936Snectar 6056100936Snectar *) Add various sanity checks to asn1_get_length() to reject 6057100936Snectar the ASN1 length bytes if they exceed sizeof(long), will appear 6058100936Snectar negative or the content length exceeds the length of the 6059160814Ssimon supplied buffer. (CVE-2002-0659) 6060100936Snectar [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6061100936Snectar 6062100936Snectar *) Assertions for various potential buffer overflows, not known to 6063100936Snectar happen in practice. 6064100936Snectar [Ben Laurie (CHATS)] 6065100936Snectar 6066100936Snectar *) Various temporary buffers to hold ASCII versions of integers were 6067160814Ssimon too small for 64 bit platforms. (CVE-2002-0655) 6068100936Snectar [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 6069100936Snectar 6070100936Snectar *) Remote buffer overflow in SSL3 protocol - an attacker could 6071160814Ssimon supply an oversized session ID to a client. (CVE-2002-0656) 6072100936Snectar [Ben Laurie (CHATS)] 6073100936Snectar 6074100936Snectar *) Remote buffer overflow in SSL2 protocol - an attacker could 6075160814Ssimon supply an oversized client master key. (CVE-2002-0656) 6076100936Snectar [Ben Laurie (CHATS)] 6077100936Snectar 6078100928Snectar Changes between 0.9.6c and 0.9.6d [9 May 2002] 6079100928Snectar 6080100928Snectar *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 6081100928Snectar encoded as NULL) with id-dsa-with-sha1. 6082100928Snectar [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 6083100928Snectar 6084100928Snectar *) Check various X509_...() return values in apps/req.c. 6085100928Snectar [Nils Larsch <nla@trustcenter.de>] 6086100928Snectar 6087100928Snectar *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 6088100928Snectar an end-of-file condition would erronously be flagged, when the CRLF 6089100928Snectar was just at the end of a processed block. The bug was discovered when 6090100928Snectar processing data through a buffering memory BIO handing the data to a 6091100928Snectar BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 6092100928Snectar <ptsekov@syntrex.com> and Nedelcho Stanev. 6093100928Snectar [Lutz Jaenicke] 6094100928Snectar 6095100928Snectar *) Implement a countermeasure against a vulnerability recently found 6096100928Snectar in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 6097100928Snectar before application data chunks to avoid the use of known IVs 6098100928Snectar with data potentially chosen by the attacker. 6099100928Snectar [Bodo Moeller] 6100100928Snectar 6101100928Snectar *) Fix length checks in ssl3_get_client_hello(). 6102100928Snectar [Bodo Moeller] 6103100928Snectar 6104100928Snectar *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 6105100928Snectar to prevent ssl3_read_internal() from incorrectly assuming that 6106100928Snectar ssl3_read_bytes() found application data while handshake 6107100928Snectar processing was enabled when in fact s->s3->in_read_app_data was 6108100928Snectar merely automatically cleared during the initial handshake. 6109100928Snectar [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 6110100928Snectar 6111100928Snectar *) Fix object definitions for Private and Enterprise: they were not 6112100928Snectar recognized in their shortname (=lowercase) representation. Extend 6113100928Snectar obj_dat.pl to issue an error when using undefined keywords instead 6114100928Snectar of silently ignoring the problem (Svenning Sorensen 6115100928Snectar <sss@sss.dnsalias.net>). 6116100928Snectar [Lutz Jaenicke] 6117100928Snectar 6118100928Snectar *) Fix DH_generate_parameters() so that it works for 'non-standard' 6119100928Snectar generators, i.e. generators other than 2 and 5. (Previously, the 6120100928Snectar code did not properly initialise the 'add' and 'rem' values to 6121100928Snectar BN_generate_prime().) 6122100928Snectar 6123100928Snectar In the new general case, we do not insist that 'generator' is 6124100928Snectar actually a primitive root: This requirement is rather pointless; 6125100928Snectar a generator of the order-q subgroup is just as good, if not 6126100928Snectar better. 6127100928Snectar [Bodo Moeller] 6128100928Snectar 6129100928Snectar *) Map new X509 verification errors to alerts. Discovered and submitted by 6130100928Snectar Tom Wu <tom@arcot.com>. 6131100928Snectar [Lutz Jaenicke] 6132100928Snectar 6133100928Snectar *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 6134100928Snectar returning non-zero before the data has been completely received 6135100928Snectar when using non-blocking I/O. 6136100928Snectar [Bodo Moeller; problem pointed out by John Hughes] 6137100928Snectar 6138100928Snectar *) Some of the ciphers missed the strength entry (SSL_LOW etc). 6139100928Snectar [Ben Laurie, Lutz Jaenicke] 6140100928Snectar 6141100928Snectar *) Fix bug in SSL_clear(): bad sessions were not removed (found by 6142100928Snectar Yoram Zahavi <YoramZ@gilian.com>). 6143100928Snectar [Lutz Jaenicke] 6144100928Snectar 6145100928Snectar *) Add information about CygWin 1.3 and on, and preserve proper 6146100928Snectar configuration for the versions before that. 6147100928Snectar [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 6148100928Snectar 6149100928Snectar *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 6150100928Snectar check whether we deal with a copy of a session and do not delete from 6151100928Snectar the cache in this case. Problem reported by "Izhar Shoshani Levi" 6152100928Snectar <izhar@checkpoint.com>. 6153100928Snectar [Lutz Jaenicke] 6154100928Snectar 6155100928Snectar *) Do not store session data into the internal session cache, if it 6156100928Snectar is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 6157100928Snectar flag is set). Proposed by Aslam <aslam@funk.com>. 6158100928Snectar [Lutz Jaenicke] 6159100928Snectar 6160100928Snectar *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 6161100928Snectar value is 0. 6162100928Snectar [Richard Levitte] 6163100928Snectar 6164109998Smarkm *) [In 0.9.6d-engine release:] 6165109998Smarkm Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 6166100928Snectar [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 6167100928Snectar 6168100928Snectar *) Add the configuration target linux-s390x. 6169100928Snectar [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 6170100928Snectar 6171100928Snectar *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 6172100928Snectar ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 6173100928Snectar variable as an indication that a ClientHello message has been 6174100928Snectar received. As the flag value will be lost between multiple 6175100928Snectar invocations of ssl3_accept when using non-blocking I/O, the 6176100928Snectar function may not be aware that a handshake has actually taken 6177100928Snectar place, thus preventing a new session from being added to the 6178100928Snectar session cache. 6179100928Snectar 6180100928Snectar To avoid this problem, we now set s->new_session to 2 instead of 6181100928Snectar using a local variable. 6182100928Snectar [Lutz Jaenicke, Bodo Moeller] 6183100928Snectar 6184100928Snectar *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 6185100928Snectar if the SSL_R_LENGTH_MISMATCH error is detected. 6186100928Snectar [Geoff Thorpe, Bodo Moeller] 6187100928Snectar 6188100928Snectar *) New 'shared_ldflag' column in Configure platform table. 6189100928Snectar [Richard Levitte] 6190100928Snectar 6191100928Snectar *) Fix EVP_CIPHER_mode macro. 6192100928Snectar ["Dan S. Camper" <dan@bti.net>] 6193100928Snectar 6194100928Snectar *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 6195100928Snectar type, we must throw them away by setting rr->length to 0. 6196100928Snectar [D P Chang <dpc@qualys.com>] 6197100928Snectar 619889837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 619989837Skris 620089837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 620189837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 620289837Skris worked incorrectly for those cases where range = 10..._2 and 620389837Skris 3*range is two bits longer than range.) 620489837Skris [Bodo Moeller] 620589837Skris 620689837Skris *) Only add signing time to PKCS7 structures if it is not already 620789837Skris present. 620889837Skris [Steve Henson] 620989837Skris 621089837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 621189837Skris OBJ_ld_ce should be OBJ_id_ce. 621289837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 621389837Skris incorrect (cf. RFC 3039). 621489837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 621589837Skris 621689837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 621789837Skris returns early because it has nothing to do. 621889837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 621989837Skris 622089837Skris *) [In 0.9.6c-engine release:] 622189837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 622289837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 622389837Skris 622489837Skris *) [In 0.9.6c-engine release:] 622589837Skris Add support for Cryptographic Appliance's keyserver technology. 622689837Skris (Use engine 'keyclient') 622789837Skris [Cryptographic Appliances and Geoff Thorpe] 622889837Skris 622989837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 623089837Skris is called via tools/c89.sh because arguments have to be 623189837Skris rearranged (all '-L' options must appear before the first object 623289837Skris modules). 623389837Skris [Richard Shapiro <rshapiro@abinitio.com>] 623489837Skris 623589837Skris *) [In 0.9.6c-engine release:] 623689837Skris Add support for Broadcom crypto accelerator cards, backported 623789837Skris from 0.9.7. 623889837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 623989837Skris 624089837Skris *) [In 0.9.6c-engine release:] 624189837Skris Add support for SureWare crypto accelerator cards from 624289837Skris Baltimore Technologies. (Use engine 'sureware') 624389837Skris [Baltimore Technologies and Mark Cox] 624489837Skris 624589837Skris *) [In 0.9.6c-engine release:] 624689837Skris Add support for crypto accelerator cards from Accelerated 624789837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 624889837Skris [AEP Inc. and Mark Cox] 624989837Skris 625089837Skris *) Add a configuration entry for gcc on UnixWare. 625189837Skris [Gary Benson <gbenson@redhat.com>] 625289837Skris 625389837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 625489837Skris messages are stored in a single piece (fixed-length part and 625589837Skris variable-length part combined) and fix various bugs found on the way. 625689837Skris [Bodo Moeller] 625789837Skris 625889837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 625989837Skris instead. BIO_gethostbyname() does not know what timeouts are 626089837Skris appropriate, so entries would stay in cache even when they have 626189837Skris become invalid. 626289837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 626389837Skris 626489837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 626589837Skris faced with a pathologically small ClientHello fragment that does 626689837Skris not contain client_version: Instead of aborting with an error, 626789837Skris simply choose the highest available protocol version (i.e., 626889837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 626989837Skris messages are never sent like this, but this change gives us 627089837Skris strictly correct behaviour at least for TLS. 627189837Skris [Bodo Moeller] 627289837Skris 627389837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 627489837Skris never resets s->method to s->ctx->method when called from within 627589837Skris one of the SSL handshake functions. 627689837Skris [Bodo Moeller; problem pointed out by Niko Baric] 627789837Skris 627889837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 627989837Skris (sent using the client's version number) if client_version is 628089837Skris smaller than the protocol version in use. Also change 628189837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 628289837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 628389837Skris the client will at least see that alert. 628489837Skris [Bodo Moeller] 628589837Skris 628689837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 628789837Skris correctly. 628889837Skris [Bodo Moeller] 628989837Skris 629089837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 629189837Skris client receives HelloRequest while in a handshake. 629289837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 629389837Skris 629489837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 629589837Skris should end in 'break', not 'goto end' which circuments various 629689837Skris cleanups done in state SSL_ST_OK. But session related stuff 629789837Skris must be disabled for SSL_ST_OK in the case that we just sent a 629889837Skris HelloRequest. 629989837Skris 630089837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 630189837Skris before just sending a HelloRequest. 630289837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 630389837Skris 630489837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 630589837Skris reveal whether illegal block cipher padding was found or a MAC 630689837Skris verification error occured. (Neither SSLerr() codes nor alerts 630789837Skris are directly visible to potential attackers, but the information 630889837Skris may leak via logfiles.) 630989837Skris 631089837Skris Similar changes are not required for the SSL 2.0 implementation 631189837Skris because the number of padding bytes is sent in clear for SSL 2.0, 631289837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 631389837Skris failed to verify that the purported number of padding bytes is in 631489837Skris the legal range. 631589837Skris [Bodo Moeller] 631689837Skris 631789837Skris *) Add OpenUNIX-8 support including shared libraries 631889837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 631989837Skris [Lutz Jaenicke] 632089837Skris 632189837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 632289837Skris 'wristwatch attack' using huge encoding parameters (cf. 632389837Skris James H. Manger's CRYPTO 2001 paper). Note that the 632489837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 632589837Skris encoding parameters and hence was not vulnerable. 632689837Skris [Bodo Moeller] 632789837Skris 632889837Skris *) BN_sqr() bug fix. 632989837Skris [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 633089837Skris 633189837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 633289837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 633389837Skris followed by modular reduction. 633489837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 633589837Skris 633689837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 633789837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 633889837Skris [Bodo Moeller] 633989837Skris 634089837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 634189837Skris This function was broken, as the check for a new client hello message 634289837Skris to handle SGC did not allow these large messages. 634389837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 634489837Skris [Lutz Jaenicke] 634589837Skris 634689837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 634789837Skris [Lutz Jaenicke] 634889837Skris 634989837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 635089837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 635189837Skris [Lutz Jaenicke] 635289837Skris 635389837Skris *) Rework the configuration and shared library support for Tru64 Unix. 635489837Skris The configuration part makes use of modern compiler features and 635589837Skris still retains old compiler behavior for those that run older versions 635689837Skris of the OS. The shared library support part includes a variant that 635789837Skris uses the RPATH feature, and is available through the special 635889837Skris configuration target "alpha-cc-rpath", which will never be selected 635989837Skris automatically. 636089837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 636189837Skris 636289837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 636389837Skris with the same message size as in ssl3_get_certificate_request(). 636489837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 636589837Skris messages might inadvertently be reject as too long. 636689837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 636789837Skris 636889837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 636989837Skris [Andy Polyakov] 637089837Skris 637189837Skris *) Modified SSL library such that the verify_callback that has been set 637289837Skris specificly for an SSL object with SSL_set_verify() is actually being 637389837Skris used. Before the change, a verify_callback set with this function was 637489837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 637589837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 637689837Skris to allow the necessary settings. 637789837Skris [Lutz Jaenicke] 637889837Skris 637989837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 638089837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 638189837Skris done automatically (in contradiction to the requirements of the C 638289837Skris standard). This made problems when used from OpenSSH. 638389837Skris [Lutz Jaenicke] 638489837Skris 638589837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 638689837Skris dh->length and always used 638789837Skris 638889837Skris BN_rand_range(priv_key, dh->p). 638989837Skris 639089837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 639189837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 639289837Skris dh->length (recommended exponent length) is much smaller than the 639389837Skris length of dh->p. We could use BN_rand_range() if the order of 639489837Skris the subgroup was stored in the DH structure, but we only have 639589837Skris dh->length. 639689837Skris 639789837Skris So switch back to 639889837Skris 639989837Skris BN_rand(priv_key, l, ...) 640089837Skris 640189837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 640289837Skris otherwise. 640389837Skris [Bodo Moeller] 640489837Skris 640589837Skris *) In 640689837Skris 640789837Skris RSA_eay_public_encrypt 640889837Skris RSA_eay_private_decrypt 640989837Skris RSA_eay_private_encrypt (signing) 641089837Skris RSA_eay_public_decrypt (signature verification) 641189837Skris 641289837Skris (default implementations for RSA_public_encrypt, 641389837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 641489837Skris always reject numbers >= n. 641589837Skris [Bodo Moeller] 641689837Skris 641789837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 641889837Skris to synchronize access to 'locking_thread'. This is necessary on 641989837Skris systems where access to 'locking_thread' (an 'unsigned long' 642089837Skris variable) is not atomic. 642189837Skris [Bodo Moeller] 642289837Skris 642389837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 642489837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 642589837Skris a race condition if 0 is a valid thread ID. 642689837Skris [Travis Vitek <vitek@roguewave.com>] 642789837Skris 642889837Skris *) Add support for shared libraries under Irix. 642989837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 643089837Skris 643189837Skris *) Add configuration option to build on Linux on both big-endian and 643289837Skris little-endian MIPS. 643389837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 643489837Skris 643589837Skris *) Add the possibility to create shared libraries on HP-UX. 643689837Skris [Richard Levitte] 643789837Skris 643879998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 643979998Skris 644079998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 644179998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 644279998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 644379998Skris PRNG state recovery was possible based on the output of 644479998Skris one PRNG request appropriately sized to gain knowledge on 644579998Skris 'md' followed by enough consecutive 1-byte PRNG requests 644679998Skris to traverse all of 'state'. 644779998Skris 644879998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 644979998Skris during PRNG output generation, hash all of the previous 645079998Skris 'md_local' value, not just the half used for PRNG output. 645179998Skris 645279998Skris 2. Make the number of bytes from 'state' included into the hash 645379998Skris independent from the number of PRNG bytes requested. 645479998Skris 645579998Skris The first measure alone would be sufficient to avoid 645679998Skris Markku-Juhani's attack. (Actually it had never occurred 645779998Skris to me that the half of 'md_local' used for chaining was the 645879998Skris half from which PRNG output bytes were taken -- I had always 645979998Skris assumed that the secret half would be used.) The second 646079998Skris measure makes sure that additional data from 'state' is never 646179998Skris mixed into 'md_local' in small portions; this heuristically 646279998Skris further strengthens the PRNG. 646379998Skris [Bodo Moeller] 646479998Skris 646579998Skris *) Fix crypto/bn/asm/mips3.s. 646679998Skris [Andy Polyakov] 646779998Skris 646879998Skris *) When only the key is given to "enc", the IV is undefined. Print out 646979998Skris an error message in this case. 647079998Skris [Lutz Jaenicke] 647179998Skris 647279998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 647379998Skris [Steve Henson] 647479998Skris 647579998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 647679998Skris positive and less than q. 647779998Skris [Bodo Moeller] 647879998Skris 647979998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 648079998Skris used: it isn't thread safe and the add_lock_callback should handle 648179998Skris that itself. 648279998Skris [Paul Rose <Paul.Rose@bridge.com>] 648379998Skris 648479998Skris *) Verify that incoming data obeys the block size in 648579998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 648679998Skris [Bodo Moeller] 648779998Skris 648879998Skris *) Fix OAEP check. 648979998Skris [Ulf M�ller, Bodo M�ller] 649079998Skris 649179998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 649289837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 649379998Skris when fixing the server behaviour for backwards-compatible 'client 649479998Skris hello' messages. (Note that the attack is impractical against 649579998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 649679998Skris means that the probability of guessing a valid ciphertext is 649779998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 649879998Skris paper.) 649979998Skris 650079998Skris Before 0.9.5, the countermeasure (hide the error by generating a 650179998Skris random 'decryption result') did not work properly because 650279998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 650379998Skris detect the supposedly ignored error. 650479998Skris 650579998Skris Both problems are now fixed. 650679998Skris [Bodo Moeller] 650779998Skris 650879998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 650979998Skris (previously it was 1024). 651079998Skris [Bodo Moeller] 651179998Skris 651279998Skris *) Fix for compatibility mode trust settings: ignore trust settings 651379998Skris unless some valid trust or reject settings are present. 651479998Skris [Steve Henson] 651579998Skris 651679998Skris *) Fix for blowfish EVP: its a variable length cipher. 651779998Skris [Steve Henson] 651879998Skris 651979998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 652079998Skris parameters in DSA public key structures and return an error in the 652179998Skris DSA routines if parameters are absent. 652279998Skris [Steve Henson] 652379998Skris 652479998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 652579998Skris in the current directory if neither $RANDFILE nor $HOME was set. 652679998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 652779998Skris caused some confusion to Windows users who haven't defined $HOME. 652879998Skris Thus RAND_file_name() is changed again: e_os.h can define a 652979998Skris DEFAULT_HOME, which will be used if $HOME is not set. 653079998Skris For Windows, we use "C:"; on other platforms, we still require 653179998Skris environment variables. 653279998Skris 653379998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 653479998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 653579998Skris having multiple threads call RAND_poll() concurrently. 653679998Skris [Bodo Moeller] 653779998Skris 653879998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 653979998Skris combination of a flag and a thread ID variable. 654079998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 654179998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 654289837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 654379998Skris that they do not hold after the first thread unsets add_do_not_lock). 654479998Skris [Bodo Moeller] 654579998Skris 654679998Skris *) Change bctest again: '-x' expressions are not available in all 654779998Skris versions of 'test'. 654879998Skris [Bodo Moeller] 654979998Skris 655076866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 655172613Skris 655276866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 655376866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 655476866Skris 655576866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 655676866Skris the default extension for executables, if any. Also, make the perl 655776866Skris scripts that use symlink() to test if it really exists and use "cp" 655876866Skris if it doesn't. All this made OpenSSL compilable and installable in 655976866Skris CygWin. 656076866Skris [Richard Levitte] 656176866Skris 656276866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 656376866Skris If SEQUENCE is length is indefinite just set c->slen to the total 656476866Skris amount of data available. 656576866Skris [Steve Henson, reported by shige@FreeBSD.org] 656676866Skris [This change does not apply to 0.9.7.] 656776866Skris 656876866Skris *) Change bctest to avoid here-documents inside command substitution 656976866Skris (workaround for FreeBSD /bin/sh bug). 657076866Skris For compatibility with Ultrix, avoid shell functions (introduced 657176866Skris in the bctest version that searches along $PATH). 657276866Skris [Bodo Moeller] 657376866Skris 657476866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 657576866Skris with des_encrypt() defined on some operating systems, like Solaris 657676866Skris and UnixWare. 657776866Skris [Richard Levitte] 657876866Skris 657976866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 658076866Skris On the Importance of Eliminating Errors in Cryptographic 658176866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 658276866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 658376866Skris [Ulf Moeller] 658476866Skris 658576866Skris *) MIPS assembler BIGNUM division bug fix. 658676866Skris [Andy Polyakov] 658776866Skris 658876866Skris *) Disabled incorrect Alpha assembler code. 658976866Skris [Richard Levitte] 659076866Skris 659176866Skris *) Fix PKCS#7 decode routines so they correctly update the length 659276866Skris after reading an EOC for the EXPLICIT tag. 659376866Skris [Steve Henson] 659476866Skris [This change does not apply to 0.9.7.] 659576866Skris 659676866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 659776866Skris if a 3DES key was generated with a 0 initial byte. Include 659876866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 659976866Skris (but broken) behaviour. 660076866Skris [Steve Henson] 660176866Skris 660276866Skris *) Enhance bctest to search for a working bc along $PATH and print 660376866Skris it when found. 660476866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 660576866Skris 660676866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 660776866Skris don't write to the wrong index in ERR_set_error_data. 660876866Skris [Bodo Moeller] 660976866Skris 661076866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 661176866Skris did not exist. 661276866Skris [Bodo Moeller] 661376866Skris 661476866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 661576866Skris [Jeremy Cooper <jeremy@baymoo.org>] 661676866Skris 661776866Skris *) Make it possible to reuse SSLv2 sessions. 661876866Skris [Richard Levitte] 661976866Skris 662076866Skris *) In copy_email() check for >= 0 as a return value for 662176866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 662276866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 662376866Skris 662476866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 662576866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 662676866Skris PKCS7_verify() fails with non detached data. 662776866Skris [Steve Henson] 662876866Skris 662976866Skris *) Don't use getenv in library functions when run as setuid/setgid. 663076866Skris New function OPENSSL_issetugid(). 663176866Skris [Ulf Moeller] 663276866Skris 663376866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 663476866Skris due to incorrect handling of multi-threading: 663576866Skris 663676866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 663776866Skris 663876866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 663976866Skris 664076866Skris 3. Count how many times MemCheck_off() has been called so that 664176866Skris nested use can be treated correctly. This also avoids 664276866Skris inband-signalling in the previous code (which relied on the 664376866Skris assumption that thread ID 0 is impossible). 664476866Skris [Bodo Moeller] 664576866Skris 664676866Skris *) Add "-rand" option also to s_client and s_server. 664776866Skris [Lutz Jaenicke] 664876866Skris 664976866Skris *) Fix CPU detection on Irix 6.x. 665076866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 665176866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 665276866Skris 665376866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 665476866Skris was empty. 665576866Skris [Steve Henson] 665676866Skris [This change does not apply to 0.9.7.] 665776866Skris 665876866Skris *) Use the cached encoding of an X509_NAME structure rather than 665976866Skris copying it. This is apparently the reason for the libsafe "errors" 666076866Skris but the code is actually correct. 666176866Skris [Steve Henson] 666276866Skris 666372613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 666472613Skris Bleichenbacher's DSA attack. 666576866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 666676866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 666776866Skris and leaves the highest bit random. 666876866Skris [Ulf Moeller, Bodo Moeller] 666972613Skris 667072613Skris *) In the NCONF_...-based implementations for CONF_... queries 667172613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 667272613Skris a temporary CONF structure with the data component set to NULL 667372613Skris (which gives segmentation faults in lh_retrieve). 667472613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 667572613Skris CONF_get_number (which may use environment variables) and directly 667672613Skris return NULL from CONF_get_section. 667772613Skris [Bodo Moeller] 667872613Skris 667972613Skris *) Fix potential buffer overrun for EBCDIC. 668072613Skris [Ulf Moeller] 668172613Skris 668272613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 668372613Skris keyUsage if basicConstraints absent for a CA. 668472613Skris [Steve Henson] 668572613Skris 668672613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 668772613Skris is more generally accepted (no spaces before the semicolon), since 668872613Skris some programs can't parse those values properly otherwise. Also make 668972613Skris sure BIO's that break lines after each write do not create invalid 669072613Skris headers. 669172613Skris [Richard Levitte] 669272613Skris 669372613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 669472613Skris macros previously used would not encode an empty SEQUENCE OF 669572613Skris and break the signature. 669672613Skris [Steve Henson] 669776866Skris [This change does not apply to 0.9.7.] 669872613Skris 669972613Skris *) Zero the premaster secret after deriving the master secret in 670072613Skris DH ciphersuites. 670172613Skris [Steve Henson] 670272613Skris 670372613Skris *) Add some EVP_add_digest_alias registrations (as found in 670472613Skris OpenSSL_add_all_digests()) to SSL_library_init() 670572613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 670672613Skris compatibility with peers using X.509 certificates 670772613Skris with unconventional AlgorithmIdentifier OIDs. 670872613Skris [Bodo Moeller] 670972613Skris 671072613Skris *) Fix for Irix with NO_ASM. 671172613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 671272613Skris 671372613Skris *) ./config script fixes. 671472613Skris [Ulf Moeller, Richard Levitte] 671572613Skris 671672613Skris *) Fix 'openssl passwd -1'. 671772613Skris [Bodo Moeller] 671872613Skris 671972613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 672072613Skris terminated strings whose length is passed in the passlen 672172613Skris parameter, for example from PEM callbacks. This was done 672272613Skris by adding an extra length parameter to asc2uni(). 672372613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 672472613Skris 672572613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 672672613Skris call failed, free the DSA structure. 672772613Skris [Bodo Moeller] 672872613Skris 672972613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 673072613Skris These are present in some PKCS#12 files. 673172613Skris [Steve Henson] 673272613Skris 673372613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 673472613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 673572613Skris when writing a 32767 byte record. 673672613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 673772613Skris 673872613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 673972613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 674072613Skris 674172613Skris (RSA objects have a reference count access to which is protected 674272613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 674372613Skris so they are meant to be shared between threads.) 674472613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 674572613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 674672613Skris 674772613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 674872613Skris [Bodo Moeller] 674972613Skris 675076866Skris *) Use better test patterns in bntest. 675176866Skris [Ulf M�ller] 675276866Skris 675372613Skris *) rand_win.c fix for Borland C. 675472613Skris [Ulf M�ller] 675572613Skris 675672613Skris *) BN_rshift bugfix for n == 0. 675772613Skris [Bodo Moeller] 675872613Skris 675976866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 676076866Skris so that 'make test' does not abort just because 'bc' is broken. 676176866Skris [Bodo Moeller] 676276866Skris 676372613Skris *) Store verify_result within SSL_SESSION also for client side to 676472613Skris avoid potential security hole. (Re-used sessions on the client side 676572613Skris always resulted in verify_result==X509_V_OK, not using the original 676672613Skris result of the server certificate verification.) 676772613Skris [Lutz Jaenicke] 676872613Skris 676972613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 677072613Skris SSL3_RT_APPLICATION_DATA, return 0. 677172613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 677272613Skris [Bodo Moeller] 677372613Skris 677472613Skris *) Fix SSL_peek: 677572613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 677672613Skris releases, have been re-implemented by renaming the previous 677772613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 677872613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 677972613Skris to them. The new ssl[23]_{read,peek} functions are calls to 678072613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 678172613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 678272613Skris does the actual work for ssl3_read_internal. 678372613Skris [Bodo Moeller] 678472613Skris 678576866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 678676866Skris the method-specific "init()" handler. Also clean up ex_data after 678776866Skris calling the method-specific "finish()" handler. Previously, this was 678876866Skris happening the other way round. 678976866Skris [Geoff Thorpe] 679076866Skris 679172613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 679272613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 679372613Skris [Bodo Moeller] 679472613Skris 679576866Skris *) Make sure that shared libraries get the internal name engine with 679676866Skris the full version number and not just 0. This should mark the 679776866Skris shared libraries as not backward compatible. Of course, this should 679876866Skris be changed again when we can guarantee backward binary compatibility. 679976866Skris [Richard Levitte] 680076866Skris 680172613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 680272613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 680372613Skris 680476866Skris *) Rework the system to generate shared libraries: 680576866Skris 680676866Skris - Make note of the expected extension for the shared libraries and 680776866Skris if there is a need for symbolic links from for example libcrypto.so.0 680876866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 680976866Skris that. 681076866Skris 681176866Skris - Make as few rebuilds of the shared libraries as possible. 681276866Skris 681376866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 681476866Skris 681576866Skris - When installing, install the shared libraries separately from the 681676866Skris static ones. 681776866Skris [Richard Levitte] 681876866Skris 681972613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 682072613Skris 682172613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 682272613Skris and not in SSL_clear because the latter is also used by the 682372613Skris accept/connect functions; previously, the settings made by 682472613Skris SSL_set_read_ahead would be lost during the handshake. 682572613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 682672613Skris 682772613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 682872613Skris Previously, it would create entries for disableed algorithms no 682972613Skris matter what. 683072613Skris [Richard Levitte] 683172613Skris 683272613Skris *) Added several new manual pages for SSL_* function. 683372613Skris [Lutz Jaenicke] 683472613Skris 683568651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 683668651Skris 683768651Skris *) In ssl23_get_client_hello, generate an error message when faced 683868651Skris with an initial SSL 3.0/TLS record that is too small to contain the 683968651Skris first two bytes of the ClientHello message, i.e. client_version. 684068651Skris (Note that this is a pathologic case that probably has never happened 684168651Skris in real life.) The previous approach was to use the version number 684268651Skris from the record header as a substitute; but our protocol choice 684368651Skris should not depend on that one because it is not authenticated 684468651Skris by the Finished messages. 684568651Skris [Bodo Moeller] 684668651Skris 684768651Skris *) More robust randomness gathering functions for Windows. 684868651Skris [Jeffrey Altman <jaltman@columbia.edu>] 684968651Skris 685068651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 685168651Skris not set then we don't setup the error code for issuer check errors 685268651Skris to avoid possibly overwriting other errors which the callback does 685368651Skris handle. If an application does set the flag then we assume it knows 685468651Skris what it is doing and can handle the new informational codes 685568651Skris appropriately. 685668651Skris [Steve Henson] 685768651Skris 685868651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 685968651Skris a general "ANY" type, as such it should be able to decode anything 686068651Skris including tagged types. However it didn't check the class so it would 686168651Skris wrongly interpret tagged types in the same way as their universal 686268651Skris counterpart and unknown types were just rejected. Changed so that the 686368651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 686468651Skris that is the encoding is stored intact. There is also a new type 686568651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 686668651Skris case we have no idea what the actual type is so we just lump them all 686768651Skris together. 686868651Skris [Steve Henson] 686968651Skris 687068651Skris *) On VMS, stdout may very well lead to a file that is written to 687168651Skris in a record-oriented fashion. That means that every write() will 687268651Skris write a separate record, which will be read separately by the 687368651Skris programs trying to read from it. This can be very confusing. 687468651Skris 687568651Skris The solution is to put a BIO filter in the way that will buffer 687668651Skris text until a linefeed is reached, and then write everything a 687768651Skris line at a time, so every record written will be an actual line, 687868651Skris not chunks of lines and not (usually doesn't happen, but I've 687968651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 688068651Skris the answer. 688168651Skris 688268651Skris Currently, it's a VMS-only method, because that's where it has 688368651Skris been tested well enough. 688468651Skris [Richard Levitte] 688568651Skris 688668651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 688768651Skris it can return incorrect results. 688868651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 688968651Skris but it was in 0.9.6-beta[12].) 689068651Skris [Bodo Moeller] 689168651Skris 689268651Skris *) Disable the check for content being present when verifying detached 689368651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 689468651Skris include zero length content when signing messages. 689568651Skris [Steve Henson] 689668651Skris 689768651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 689868651Skris BIO_ctrl (for BIO pairs). 689968651Skris [Bodo M�ller] 690068651Skris 690168651Skris *) Add DSO method for VMS. 690268651Skris [Richard Levitte] 690368651Skris 690468651Skris *) Bug fix: Montgomery multiplication could produce results with the 690568651Skris wrong sign. 690668651Skris [Ulf M�ller] 690768651Skris 690868651Skris *) Add RPM specification openssl.spec and modify it to build three 690968651Skris packages. The default package contains applications, application 691068651Skris documentation and run-time libraries. The devel package contains 691168651Skris include files, static libraries and function documentation. The 691268651Skris doc package contains the contents of the doc directory. The original 691368651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 691468651Skris [Richard Levitte] 691568651Skris 691668651Skris *) Add a large number of documentation files for many SSL routines. 691768651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 691868651Skris 691968651Skris *) Add a configuration entry for Sony News 4. 692068651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 692168651Skris 692268651Skris *) Don't set the two most significant bits to one when generating a 692368651Skris random number < q in the DSA library. 692468651Skris [Ulf M�ller] 692568651Skris 692668651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 692768651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 692868651Skris the underlying transport is blocking) if a handshake took place. 692968651Skris (The default behaviour is needed by applications such as s_client 693068651Skris and s_server that use select() to determine when to use SSL_read; 693168651Skris but for applications that know in advance when to expect data, it 693268651Skris just makes things more complicated.) 693368651Skris [Bodo Moeller] 693468651Skris 693568651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 693668651Skris from EGD. 693768651Skris [Ben Laurie] 693868651Skris 693968651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 694068651Skris work better on such systems. 694168651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 694268651Skris 694368651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 694468651Skris Update PKCS12_parse() so it copies the friendlyName and the 694568651Skris keyid to the certificates aux info. 694668651Skris [Steve Henson] 694768651Skris 694868651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 694968651Skris if there was more than one signature. 695068651Skris [Sven Uszpelkat <su@celocom.de>] 695168651Skris 695268651Skris *) Major change in util/mkdef.pl to include extra information 695368651Skris about each symbol, as well as presentig variables as well 695468651Skris as functions. This change means that there's n more need 695568651Skris to rebuild the .num files when some algorithms are excluded. 695668651Skris [Richard Levitte] 695768651Skris 695868651Skris *) Allow the verify time to be set by an application, 695968651Skris rather than always using the current time. 696068651Skris [Steve Henson] 696168651Skris 696268651Skris *) Phase 2 verify code reorganisation. The certificate 696368651Skris verify code now looks up an issuer certificate by a 696468651Skris number of criteria: subject name, authority key id 696568651Skris and key usage. It also verifies self signed certificates 696668651Skris by the same criteria. The main comparison function is 696768651Skris X509_check_issued() which performs these checks. 696868651Skris 696968651Skris Lot of changes were necessary in order to support this 697068651Skris without completely rewriting the lookup code. 697168651Skris 697268651Skris Authority and subject key identifier are now cached. 697368651Skris 697468651Skris The LHASH 'certs' is X509_STORE has now been replaced 697568651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 697668651Skris LHASH can't store or retrieve multiple objects with 697768651Skris the same hash value. 697868651Skris 697968651Skris As a result various functions (which were all internal 698068651Skris use only) have changed to handle the new X509_STORE 698168651Skris structure. This will break anything that messed round 698268651Skris with X509_STORE internally. 698368651Skris 698468651Skris The functions X509_STORE_add_cert() now checks for an 698568651Skris exact match, rather than just subject name. 698668651Skris 698768651Skris The X509_STORE API doesn't directly support the retrieval 698868651Skris of multiple certificates matching a given criteria, however 698968651Skris this can be worked round by performing a lookup first 699068651Skris (which will fill the cache with candidate certificates) 699168651Skris and then examining the cache for matches. This is probably 699268651Skris the best we can do without throwing out X509_LOOKUP 699368651Skris entirely (maybe later...). 699468651Skris 699568651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 699668651Skris 699768651Skris All certificate lookup operations now go via a get_issuer() 699868651Skris callback. Although this currently uses an X509_STORE it 699968651Skris can be replaced by custom lookups. This is a simple way 700068651Skris to bypass the X509_STORE hackery necessary to make this 700168651Skris work and makes it possible to use more efficient techniques 700268651Skris in future. A very simple version which uses a simple 700368651Skris STACK for its trusted certificate store is also provided 700468651Skris using X509_STORE_CTX_trusted_stack(). 700568651Skris 700668651Skris The verify_cb() and verify() callbacks now have equivalents 700768651Skris in the X509_STORE_CTX structure. 700868651Skris 700968651Skris X509_STORE_CTX also has a 'flags' field which can be used 701068651Skris to customise the verify behaviour. 701168651Skris [Steve Henson] 701268651Skris 701368651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 701468651Skris excludes S/MIME capabilities. 701568651Skris [Steve Henson] 701668651Skris 701768651Skris *) When a certificate request is read in keep a copy of the 701868651Skris original encoding of the signed data and use it when outputing 701968651Skris again. Signatures then use the original encoding rather than 702068651Skris a decoded, encoded version which may cause problems if the 702168651Skris request is improperly encoded. 702268651Skris [Steve Henson] 702368651Skris 702468651Skris *) For consistency with other BIO_puts implementations, call 702568651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 702668651Skris BIO_write(b, ...). 702768651Skris 702868651Skris In BIO_puts, increment b->num_write as in BIO_write. 702968651Skris [Peter.Sylvester@EdelWeb.fr] 703068651Skris 703168651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 703268651Skris BN_zero, we may not return a BIGNUM with an array consisting of 703368651Skris words set to zero.) 703468651Skris [Bodo Moeller] 703568651Skris 703668651Skris *) Avoid calling abort() from within the library when problems are 703768651Skris detected, except if preprocessor symbols have been defined 703868651Skris (such as REF_CHECK, BN_DEBUG etc.). 703968651Skris [Bodo Moeller] 704068651Skris 704168651Skris *) New openssl application 'rsautl'. This utility can be 704268651Skris used for low level RSA operations. DER public key 704368651Skris BIO/fp routines also added. 704468651Skris [Steve Henson] 704568651Skris 704668651Skris *) New Configure entry and patches for compiling on QNX 4. 704768651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 704868651Skris 704968651Skris *) A demo state-machine implementation was sponsored by 705068651Skris Nuron (http://www.nuron.com/) and is now available in 705168651Skris demos/state_machine. 705268651Skris [Ben Laurie] 705368651Skris 705468651Skris *) New options added to the 'dgst' utility for signature 705568651Skris generation and verification. 705668651Skris [Steve Henson] 705768651Skris 705868651Skris *) Unrecognized PKCS#7 content types are now handled via a 705968651Skris catch all ASN1_TYPE structure. This allows unsupported 706068651Skris types to be stored as a "blob" and an application can 706168651Skris encode and decode it manually. 706268651Skris [Steve Henson] 706368651Skris 706468651Skris *) Fix various signed/unsigned issues to make a_strex.c 706568651Skris compile under VC++. 706668651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 706768651Skris 706868651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 706968651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 707068651Skris if passed a NULL BN and its argument was negative. 707168651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 707268651Skris 707368651Skris *) Modification to PKCS#7 encoding routines to output definite 707468651Skris length encoding. Since currently the whole structures are in 707568651Skris memory there's not real point in using indefinite length 707668651Skris constructed encoding. However if OpenSSL is compiled with 707768651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 707868651Skris [Steve Henson] 707968651Skris 708068651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 708168651Skris [Richard Levitte] 708268651Skris 708368651Skris *) Added more prefixes to parse for in the the strings written 708468651Skris through a logging bio, to cover all the levels that are available 708568651Skris through syslog. The prefixes are now: 708668651Skris 708768651Skris PANIC, EMERG, EMR => LOG_EMERG 708868651Skris ALERT, ALR => LOG_ALERT 708968651Skris CRIT, CRI => LOG_CRIT 709068651Skris ERROR, ERR => LOG_ERR 709168651Skris WARNING, WARN, WAR => LOG_WARNING 709268651Skris NOTICE, NOTE, NOT => LOG_NOTICE 709368651Skris INFO, INF => LOG_INFO 709468651Skris DEBUG, DBG => LOG_DEBUG 709568651Skris 709668651Skris and as before, if none of those prefixes are present at the 709768651Skris beginning of the string, LOG_ERR is chosen. 709868651Skris 709968651Skris On Win32, the LOG_* levels are mapped according to this: 710068651Skris 710168651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 710268651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 710368651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 710468651Skris 710568651Skris [Richard Levitte] 710668651Skris 710768651Skris *) Made it possible to reconfigure with just the configuration 710868651Skris argument "reconf" or "reconfigure". The command line arguments 710968651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 711068651Skris and are retrieved from there when reconfiguring. 711168651Skris [Richard Levitte] 711268651Skris 711368651Skris *) MD4 implemented. 711468651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 711568651Skris 711668651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 711768651Skris [Richard Levitte] 711868651Skris 711968651Skris *) The obj_dat.pl script was messing up the sorting of object 712068651Skris names. The reason was that it compared the quoted version 712168651Skris of strings as a result "OCSP" > "OCSP Signing" because 712268651Skris " > SPACE. Changed script to store unquoted versions of 712368651Skris names and add quotes on output. It was also omitting some 712468651Skris names from the lookup table if they were given a default 712568651Skris value (that is if SN is missing it is given the same 712668651Skris value as LN and vice versa), these are now added on the 712768651Skris grounds that if an object has a name we should be able to 712868651Skris look it up. Finally added warning output when duplicate 712968651Skris short or long names are found. 713068651Skris [Steve Henson] 713168651Skris 713268651Skris *) Changes needed for Tandem NSK. 713368651Skris [Scott Uroff <scott@xypro.com>] 713468651Skris 713568651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 713668651Skris RSA_padding_check_SSLv23(), special padding was never detected 713768651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 713868651Skris version rollback attacks was not effective. 713968651Skris 714068651Skris In s23_clnt.c, don't use special rollback-attack detection padding 714168651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 714268651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 714368651Skris SSL 2.0 is the only protocol enabled in the server. 714468651Skris [Bodo Moeller] 714568651Skris 714668651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 714768651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 714868651Skris BIO_dump_indent() are added. 714968651Skris [Richard Levitte] 715068651Skris 715168651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 715268651Skris these print out strings and name structures based on various 715368651Skris flags including RFC2253 support and proper handling of 715468651Skris multibyte characters. Added options to the 'x509' utility 715568651Skris to allow the various flags to be set. 715668651Skris [Steve Henson] 715768651Skris 715868651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 715968651Skris Also change the functions X509_cmp_current_time() and 716068651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 716168651Skris this will enable certificates using GeneralizedTime in validity 716268651Skris dates to be checked. 716368651Skris [Steve Henson] 716468651Skris 716568651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 716668651Skris negative public key encodings) on by default, 716768651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 716868651Skris [Steve Henson] 716968651Skris 717068651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 717168651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 717268651Skris the encoding can be trivially obtained from the structure. 717368651Skris [Steve Henson] 717468651Skris 717568651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 717668651Skris not read locks (CRYPTO_r_[un]lock). 717768651Skris [Bodo Moeller] 717868651Skris 717968651Skris *) A first attempt at creating official support for shared 718068651Skris libraries through configuration. I've kept it so the 718168651Skris default is static libraries only, and the OpenSSL programs 718268651Skris are always statically linked for now, but there are 718368651Skris preparations for dynamic linking in place. 718489837Skris This has been tested on Linux and Tru64. 718568651Skris [Richard Levitte] 718668651Skris 718768651Skris *) Randomness polling function for Win9x, as described in: 718868651Skris Peter Gutmann, Software Generation of Practically Strong 718968651Skris Random Numbers. 719068651Skris [Ulf M�ller] 719168651Skris 719268651Skris *) Fix so PRNG is seeded in req if using an already existing 719368651Skris DSA key. 719468651Skris [Steve Henson] 719568651Skris 719668651Skris *) New options to smime application. -inform and -outform 719768651Skris allow alternative formats for the S/MIME message including 719868651Skris PEM and DER. The -content option allows the content to be 719968651Skris specified separately. This should allow things like Netscape 720068651Skris form signing output easier to verify. 720168651Skris [Steve Henson] 720268651Skris 720368651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 720468651Skris [Steve Henson] 720568651Skris 720668651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 720768651Skris STRING types. These convert content octets to and from the 720868651Skris underlying type. The actual tag and length octets are 720968651Skris already assumed to have been read in and checked. These 721068651Skris are needed because all other string types have virtually 721168651Skris identical handling apart from the tag. By having versions 721268651Skris of the ASN1 functions that just operate on content octets 721368651Skris IMPLICIT tagging can be handled properly. It also allows 721468651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 721568651Skris and ASN1_INTEGER are identical apart from the tag. 721668651Skris [Steve Henson] 721768651Skris 721868651Skris *) Change the handling of OID objects as follows: 721968651Skris 722068651Skris - New object identifiers are inserted in objects.txt, following 722168651Skris the syntax given in objects.README. 722268651Skris - objects.pl is used to process obj_mac.num and create a new 722368651Skris obj_mac.h. 722468651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 722568651Skris obj_mac.h. 722668651Skris 722768651Skris This is currently kind of a hack, and the perl code in objects.pl 722868651Skris isn't very elegant, but it works as I intended. The simplest way 722968651Skris to check that it worked correctly is to look in obj_dat.h and 723068651Skris check the array nid_objs and make sure the objects haven't moved 723168651Skris around (this is important!). Additions are OK, as well as 723268651Skris consistent name changes. 723368651Skris [Richard Levitte] 723468651Skris 723568651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 723668651Skris [Bodo Moeller] 723768651Skris 723868651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 723968651Skris The given file adds to whatever has already been seeded into the 724068651Skris random pool through the RANDFILE configuration file option or 724168651Skris environment variable, or the default random state file. 724268651Skris [Richard Levitte] 724368651Skris 724468651Skris *) mkstack.pl now sorts each macro group into lexical order. 724568651Skris Previously the output order depended on the order the files 724668651Skris appeared in the directory, resulting in needless rewriting 724768651Skris of safestack.h . 724868651Skris [Steve Henson] 724968651Skris 725068651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 725168651Skris work arounds for the VC++ problem that it treats func() as 725268651Skris func(void). Also stripped out the parts of mkdef.pl that 725368651Skris added extra typesafe functions: these no longer exist. 725468651Skris [Steve Henson] 725568651Skris 725668651Skris *) Reorganisation of the stack code. The macros are now all 725768651Skris collected in safestack.h . Each macro is defined in terms of 725868651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 725968651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 726068651Skris this has the advantage of retaining type safety without the 726168651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 726268651Skris then the non typesafe macros are used instead. Also modified the 726368651Skris mkstack.pl script to handle the new form. Needs testing to see 726468651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 726568651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 726668651Skris and PKCS12_STACK_OF. 726768651Skris [Steve Henson] 726868651Skris 726968651Skris *) When some versions of IIS use the 'NET' form of private key the 727068651Skris key derivation algorithm is different. Normally MD5(password) is 727168651Skris used as a 128 bit RC4 key. In the modified case 727268651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 727368651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 727468651Skris as the old Netscape_RSA functions except they have an additional 727568651Skris 'sgckey' parameter which uses the modified algorithm. Also added 727668651Skris an -sgckey command line option to the rsa utility. Thanks to 727768651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 727868651Skris algorithm to openssl-dev. 727968651Skris [Steve Henson] 728068651Skris 728168651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 728268651Skris invalid expansion on some systems (SCO 5.0.5 for example). 728368651Skris Corrected to 'c.kname'. 728468651Skris [Phillip Porch <root@theporch.com>] 728568651Skris 728668651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 728768651Skris a STACK of email addresses from a certificate or request, these look 728868651Skris in the subject name and the subject alternative name extensions and 728968651Skris omit any duplicate addresses. 729068651Skris [Steve Henson] 729168651Skris 729268651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 729368651Skris This makes DSA verification about 2 % faster. 729468651Skris [Bodo Moeller] 729568651Skris 729668651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 729768651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 729868651Skris plus overhead for 1024 bit moduli). 729968651Skris This makes exponentiations about 0.5 % faster for 1024 bit 730068651Skris exponents (as measured by "openssl speed rsa2048"). 730168651Skris [Bodo Moeller] 730268651Skris 730368651Skris *) Rename memory handling macros to avoid conflicts with other 730468651Skris software: 730568651Skris Malloc => OPENSSL_malloc 730668651Skris Malloc_locked => OPENSSL_malloc_locked 730768651Skris Realloc => OPENSSL_realloc 730868651Skris Free => OPENSSL_free 730968651Skris [Richard Levitte] 731068651Skris 731168651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 731268651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 731368651Skris [Bodo Moeller] 731468651Skris 731568651Skris *) CygWin32 support. 731668651Skris [John Jarvie <jjarvie@newsguy.com>] 731768651Skris 731868651Skris *) The type-safe stack code has been rejigged. It is now only compiled 731968651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 732068651Skris by default all type-specific stack functions are "#define"d back to 732168651Skris standard stack functions. This results in more streamlined output 732268651Skris but retains the type-safety checking possibilities of the original 732368651Skris approach. 732468651Skris [Geoff Thorpe] 732568651Skris 732668651Skris *) The STACK code has been cleaned up, and certain type declarations 732768651Skris that didn't make a lot of sense have been brought in line. This has 732868651Skris also involved a cleanup of sorts in safestack.h to more correctly 732968651Skris map type-safe stack functions onto their plain stack counterparts. 733068651Skris This work has also resulted in a variety of "const"ifications of 733168651Skris lots of the code, especially "_cmp" operations which should normally 733268651Skris be prototyped with "const" parameters anyway. 733368651Skris [Geoff Thorpe] 733468651Skris 733568651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 733668651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 733768651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 733868651Skris where all of 'md' is used each time the PRNG is used, but 'state' 733968651Skris is used only indexed by a cyclic counter. As entropy may not be 734068651Skris well distributed from the beginning, 'md' is important as a 734168651Skris chaining variable. However, the output function chains only half 734268651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 734368651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 734468651Skris in all of 'state' being rewritten, with the new values depending 734568651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 734668651Skris [Bodo Moeller] 734768651Skris 734868651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 734968651Skris the handshake is continued after ssl_verify_cert_chain(); 735068651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 735168651Skris can lead to 'unexplainable' connection aborts later. 735268651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 735368651Skris 735468651Skris *) Major EVP API cipher revision. 735568651Skris Add hooks for extra EVP features. This allows various cipher 735668651Skris parameters to be set in the EVP interface. Support added for variable 735768651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 735868651Skris setting of RC2 and RC5 parameters. 735968651Skris 736068651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 736168651Skris ciphers. 736268651Skris 736368651Skris Remove lots of duplicated code from the EVP library. For example *every* 736468651Skris cipher init() function handles the 'iv' in the same way according to the 736568651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 736668651Skris for CFB and OFB modes they zero ctx->num. 736768651Skris 736868651Skris New functionality allows removal of S/MIME code RC2 hack. 736968651Skris 737068651Skris Most of the routines have the same form and so can be declared in terms 737168651Skris of macros. 737268651Skris 737368651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 737468651Skris all individual ciphers. If the cipher wants to handle IVs or keys 737568651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 737668651Skris flags. 737768651Skris 737868651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 737968651Skris value: although software versions of the algorithms cannot fail 738068651Skris any installed hardware versions can. 738168651Skris [Steve Henson] 738268651Skris 738368651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 738468651Skris this option is set, tolerate broken clients that send the negotiated 738568651Skris protocol version number instead of the requested protocol version 738668651Skris number. 738768651Skris [Bodo Moeller] 738868651Skris 738968651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 739068651Skris i.e. non-zero for export ciphersuites, zero otherwise. 739168651Skris Previous versions had this flag inverted, inconsistent with 739268651Skris rsa_tmp_cb (..._TMP_RSA_CB). 739368651Skris [Bodo Moeller; problem reported by Amit Chopra] 739468651Skris 739568651Skris *) Add missing DSA library text string. Work around for some IIS 739668651Skris key files with invalid SEQUENCE encoding. 739768651Skris [Steve Henson] 739868651Skris 739968651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 740068651Skris and so on that are implemented in OpenSSL. 740168651Skris [Richard Levitte] 740268651Skris 740368651Skris *) Enhance c_rehash script. Old version would mishandle certificates 740468651Skris with the same subject name hash and wouldn't handle CRLs at all. 740568651Skris Added -fingerprint option to crl utility, to support new c_rehash 740668651Skris features. 740768651Skris [Steve Henson] 740868651Skris 740968651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 741068651Skris [Ulf M�ller] 741168651Skris 741268651Skris *) Fix for SSL server purpose checking. Server checking was 741368651Skris rejecting certificates which had extended key usage present 741468651Skris but no ssl client purpose. 741568651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 741668651Skris 741768651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 741868651Skris is a little unclear about how a blank password is handled. 741968651Skris Since the password in encoded as a BMPString with terminating 742068651Skris double NULL a zero length password would end up as just the 742168651Skris double NULL. However no password at all is different and is 742268651Skris handled differently in the PKCS#12 key generation code. NS 742368651Skris treats a blank password as zero length. MSIE treats it as no 742468651Skris password on export: but it will try both on import. We now do 742568651Skris the same: PKCS12_parse() tries zero length and no password if 742668651Skris the password is set to "" or NULL (NULL is now a valid password: 742768651Skris it wasn't before) as does the pkcs12 application. 742868651Skris [Steve Henson] 742968651Skris 743068651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 743168651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 743268651Skris be obtained from the error queue. 743368651Skris [Bodo Moeller] 743468651Skris 743568651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 743668651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 743768651Skris accordingly to avoid race conditions (this is necessary because 743868651Skris thread_hash is no longer constant once set). 743968651Skris [Bodo Moeller] 744068651Skris 744168651Skris *) Bugfix for linux-elf makefile.one. 744268651Skris [Ulf M�ller] 744368651Skris 744468651Skris *) RSA_get_default_method() will now cause a default 744568651Skris RSA_METHOD to be chosen if one doesn't exist already. 744668651Skris Previously this was only set during a call to RSA_new() 744768651Skris or RSA_new_method(NULL) meaning it was possible for 744868651Skris RSA_get_default_method() to return NULL. 744968651Skris [Geoff Thorpe] 745068651Skris 745168651Skris *) Added native name translation to the existing DSO code 745268651Skris that will convert (if the flag to do so is set) filenames 745368651Skris that are sufficiently small and have no path information 745468651Skris into a canonical native form. Eg. "blah" converted to 745568651Skris "libblah.so" or "blah.dll" etc. 745668651Skris [Geoff Thorpe] 745768651Skris 745868651Skris *) New function ERR_error_string_n(e, buf, len) which is like 745968651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 746068651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 746168651Skris may not be NULL. 746268651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 746368651Skris 746468651Skris *) CONF library reworked to become more general. A new CONF 746568651Skris configuration file reader "class" is implemented as well as a 746668651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 746768651Skris old CONF_* functions are still there, but are reimplemented to 746868651Skris work in terms of the new functions. Also, a set of functions 746968651Skris to handle the internal storage of the configuration data is 747068651Skris provided to make it easier to write new configuration file 747168651Skris reader "classes" (I can definitely see something reading a 747268651Skris configuration file in XML format, for example), called _CONF_*, 747368651Skris or "the configuration storage API"... 747468651Skris 747568651Skris The new configuration file reading functions are: 747668651Skris 747768651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 747868651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 747968651Skris 748068651Skris NCONF_default, NCONF_WIN32 748168651Skris 748268651Skris NCONF_dump_fp, NCONF_dump_bio 748368651Skris 748468651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 748568651Skris NCONF_new creates a new CONF object. This works in the same way 748668651Skris as other interfaces in OpenSSL, like the BIO interface. 748768651Skris NCONF_dump_* dump the internal storage of the configuration file, 748868651Skris which is useful for debugging. All other functions take the same 748968651Skris arguments as the old CONF_* functions wth the exception of the 749068651Skris first that must be a `CONF *' instead of a `LHASH *'. 749168651Skris 749268651Skris To make it easer to use the new classes with the old CONF_* functions, 749368651Skris the function CONF_set_default_method is provided. 749468651Skris [Richard Levitte] 749568651Skris 749668651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 749768651Skris mentioned in the documentation but had not been implemented. 749868651Skris (This option is not yet really useful because even the additional 749968651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 750068651Skris [Bodo Moeller] 750168651Skris 750268651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 750368651Skris OpenSSL-based applications) load shared libraries and bind to 750468651Skris them in a portable way. 750568651Skris [Geoff Thorpe, with contributions from Richard Levitte] 750668651Skris 750759191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 750859191Skris 750959191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 751059191Skris 751159191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 751259191Skris (the default implementation of RAND_status). 751359191Skris 751459191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 751559191Skris to '-clrext' (= clear extensions), as intended and documented. 751659191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 751759191Skris <attili@amaxo.com>] 751859191Skris 751959191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 752059191Skris was larger than the MD block size. 752159191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 752259191Skris 752359191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 752459191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 752559191Skris using the passed key: if the passed key was a private key the result 752659191Skris of X509_print(), for example, would be to print out all the private key 752759191Skris components. 752859191Skris [Steve Henson] 752959191Skris 753059191Skris *) des_quad_cksum() byte order bug fix. 753159191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 753259191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 753359191Skris 753459191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 753559191Skris discouraged. 753659191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 753759191Skris 753859191Skris *) For easily testing in shell scripts whether some command 753959191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 754059191Skris returns with exit code 0 iff no command of the given name is available. 754159191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 754259191Skris the output goes to stdout and nothing is printed to stderr. 754359191Skris Additional arguments are always ignored. 754459191Skris 754559191Skris Since for each cipher there is a command of the same name, 754659191Skris the 'no-cipher' compilation switches can be tested this way. 754759191Skris 754859191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 754959191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 755059191Skris [Bodo Moeller] 755159191Skris 755259191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 755359191Skris [Bodo Moeller] 755459191Skris 755559191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 755659191Skris is set; it will be thrown away anyway because each handshake creates 755759191Skris its own key. 755859191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 755959191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 756059191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 756159191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 756259191Skris [Bodo Moeller] 756359191Skris 756459191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 756559191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 756659191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 756759191Skris does not suppress any output. 756859191Skris [Richard Levitte] 756959191Skris 757059191Skris *) Add compatibility options to the purpose and trust code. The 757159191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 757259191Skris accepts a certificate or CA, this was the previous behaviour, 757359191Skris with all the associated security issues. 757459191Skris 757559191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 757659191Skris automatically trust self signed roots in certificate store. A 757759191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 757859191Skris a purpose has no associated trust setting and it should instead 757959191Skris use the value in the default purpose. 758059191Skris [Steve Henson] 758159191Skris 758259191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 758359191Skris and fix a memory leak. 758459191Skris [Steve Henson] 758559191Skris 758659191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 758759191Skris reason strings from the previous version of the .c file, as 758859191Skris the default to have only downcase letters (and digits) in 758959191Skris automatically generated reasons codes is not always appropriate. 759059191Skris [Bodo Moeller] 759159191Skris 759259191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 759359191Skris using strerror. Previously, ERR_reason_error_string() returned 759459191Skris library names as reason strings for SYSerr; but SYSerr is a special 759559191Skris case where small numbers are errno values, not library numbers. 759659191Skris [Bodo Moeller] 759759191Skris 759859191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 759959191Skris converts DSA parameters into DH parameters. (When creating parameters, 760059191Skris DSA_generate_parameters is used.) 760159191Skris [Bodo Moeller] 760259191Skris 760359191Skris *) Include 'length' (recommended exponent length) in C code generated 760459191Skris by 'openssl dhparam -C'. 760559191Skris [Bodo Moeller] 760659191Skris 760759191Skris *) The second argument to set_label in perlasm was already being used 760859191Skris so couldn't be used as a "file scope" flag. Moved to third argument 760959191Skris which was free. 761059191Skris [Steve Henson] 761159191Skris 761259191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 761359191Skris instead of RAND_bytes for encryption IVs and salts. 761459191Skris [Bodo Moeller] 761559191Skris 761659191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 761759191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 761859191Skris RAND_set_rand_method would be impossible. 761959191Skris [Bodo Moeller] 762059191Skris 762159191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 762259191Skris number generation fails. 762359191Skris [Bodo Moeller] 762459191Skris 762559191Skris *) New 'rand' application for creating pseudo-random output. 762659191Skris [Bodo Moeller] 762759191Skris 762859191Skris *) Added configuration support for Linux/IA64 762959191Skris [Rolf Haberrecker <rolf@suse.de>] 763059191Skris 763159191Skris *) Assembler module support for Mingw32. 763259191Skris [Ulf M�ller] 763359191Skris 763459191Skris *) Shared library support for HPUX (in shlib/). 763559191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 763659191Skris 763759191Skris *) Shared library support for Solaris gcc. 763859191Skris [Lutz Behnke <behnke@trustcenter.de>] 763959191Skris 764059191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 764159191Skris 764259191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 764359191Skris were added manually and by SMIME_crlf_copy(). 764459191Skris [Steve Henson] 764559191Skris 764659191Skris *) In bntest.c don't call BN_rand with zero bits argument. 764759191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 764859191Skris 764959191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 765059191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 765159191Skris [Ulf M�ller] 765259191Skris 765359191Skris *) Add an optional second argument to the set_label() in the perl 765459191Skris assembly language builder. If this argument exists and is set 765559191Skris to 1 it signals that the assembler should use a symbol whose 765659191Skris scope is the entire file, not just the current function. This 765759191Skris is needed with MASM which uses the format label:: for this scope. 765859191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 765959191Skris 766059191Skris *) Change the ASN1 types so they are typedefs by default. Before 766159191Skris almost all types were #define'd to ASN1_STRING which was causing 766259191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 766359191Skris for example. 766459191Skris [Steve Henson] 766559191Skris 766659191Skris *) Change names of new functions to the new get1/get0 naming 766759191Skris convention: After 'get1', the caller owns a reference count 766859191Skris and has to call ..._free; 'get0' returns a pointer to some 766959191Skris data structure without incrementing reference counters. 767059191Skris (Some of the existing 'get' functions increment a reference 767159191Skris counter, some don't.) 767259191Skris Similarly, 'set1' and 'add1' functions increase reference 767359191Skris counters or duplicate objects. 767459191Skris [Steve Henson] 767559191Skris 767659191Skris *) Allow for the possibility of temp RSA key generation failure: 767759191Skris the code used to assume it always worked and crashed on failure. 767859191Skris [Steve Henson] 767959191Skris 768059191Skris *) Fix potential buffer overrun problem in BIO_printf(). 768159191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 768259191Skris pointed out by David Sacerdote <das33@cornell.edu>] 768359191Skris 768459191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 768559191Skris RAND_egd() and RAND_status(). In the command line application, 768659191Skris the EGD socket can be specified like a seed file using RANDFILE 768759191Skris or -rand. 768859191Skris [Ulf M�ller] 768959191Skris 769059191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 769159191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 769259191Skris [Steve Henson] 769359191Skris 769459191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 769559191Skris list to exclude them. This means that no special compilation option 769659191Skris is needed to use anonymous DH: it just needs to be included in the 769759191Skris cipher list. 769859191Skris [Steve Henson] 769959191Skris 770059191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 770159191Skris EVP_MD_type. The old functionality is available in a new macro called 770259191Skris EVP_MD_md(). Change code that uses it and update docs. 770359191Skris [Steve Henson] 770459191Skris 770559191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 770659191Skris where the 'void *' argument is replaced by a function pointer argument. 770759191Skris Previously 'void *' was abused to point to functions, which works on 770859191Skris many platforms, but is not correct. As these functions are usually 770959191Skris called by macros defined in OpenSSL header files, most source code 771059191Skris should work without changes. 771159191Skris [Richard Levitte] 771259191Skris 771359191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 771459191Skris sections with information on -D... compiler switches used for 771559191Skris compiling the library so that applications can see them. To enable 771659191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 771759191Skris must be defined. E.g., 771859191Skris #define OPENSSL_ALGORITHM_DEFINES 771959191Skris #include <openssl/opensslconf.h> 772059191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 772159191Skris [Richard Levitte, Ulf and Bodo M�ller] 772259191Skris 772359191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 772459191Skris record layer. 772559191Skris [Bodo Moeller] 772659191Skris 772759191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 772859191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 772959191Skris the required ASN1 format: arbitrary types determined by an OID. 773059191Skris [Steve Henson] 773159191Skris 773259191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 773359191Skris argument to 'req'. This is not because the function is newer or 773459191Skris better than others it just uses the work 'NEW' in the certificate 773559191Skris request header lines. Some software needs this. 773659191Skris [Steve Henson] 773759191Skris 773859191Skris *) Reorganise password command line arguments: now passwords can be 773959191Skris obtained from various sources. Delete the PEM_cb function and make 774059191Skris it the default behaviour: i.e. if the callback is NULL and the 774159191Skris usrdata argument is not NULL interpret it as a null terminated pass 774259191Skris phrase. If usrdata and the callback are NULL then the pass phrase 774359191Skris is prompted for as usual. 774459191Skris [Steve Henson] 774559191Skris 774659191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 774759191Skris the support is automatically enabled. The resulting binaries will 774859191Skris autodetect the card and use it if present. 774959191Skris [Ben Laurie and Compaq Inc.] 775059191Skris 775159191Skris *) Work around for Netscape hang bug. This sends certificate request 775259191Skris and server done in one record. Since this is perfectly legal in the 775359191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 775459191Skris the bugs/SSLv3 entry for more info. 775559191Skris [Steve Henson] 775659191Skris 775759191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 775859191Skris [Andy Polyakov] 775959191Skris 776059191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 776159191Skris of seed file. 776259191Skris [Steve Henson] 776359191Skris 776459191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 776559191Skris [Bodo Moeller] 776659191Skris 776759191Skris *) Add command line password options to the remaining applications. 776859191Skris [Steve Henson] 776959191Skris 777059191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 777159191Skris bits. 777259191Skris [Ulf M�ller] 777359191Skris 777459191Skris *) More tests in bntest.c, and changed test_bn output. 777559191Skris [Ulf M�ller] 777659191Skris 777759191Skris *) ./config recognizes MacOS X now. 777859191Skris [Andy Polyakov] 777959191Skris 778059191Skris *) Bug fix for BN_div() when the first words of num and divsor are 778159191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 778259191Skris [Ulf M�ller] 778359191Skris 778459191Skris *) Add support for various broken PKCS#8 formats, and command line 778559191Skris options to produce them. 778659191Skris [Steve Henson] 778759191Skris 778859191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 778959191Skris get temporary BIGNUMs from a BN_CTX. 779059191Skris [Ulf M�ller] 779159191Skris 779259191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 779359191Skris for p == 0. 779459191Skris [Ulf M�ller] 779559191Skris 779659191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 779759191Skris include a #define from the old name to the new. The original intent 779859191Skris was that statically linked binaries could for example just call 779959191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 780059191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 780159191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 780259191Skris one would link with the other. They are now in separate source files. 780359191Skris [Steve Henson] 780459191Skris 780559191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 780659191Skris [Steve Henson] 780759191Skris 780859191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 780959191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 781059191Skris loop, our standard modexp algorithms are faster). 781159191Skris [Bodo Moeller] 781259191Skris 781359191Skris *) Support for the EBCDIC character set completed. 781459191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 781559191Skris 781659191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 781759191Skris use void * instead of char * in lhash. 781859191Skris [Ulf M�ller] 781959191Skris 782059191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 782159191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 782259191Skris this the server could overwrite ephemeral keys that the client 782359191Skris has already seen). 782459191Skris [Bodo Moeller] 782559191Skris 782659191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 782759191Skris using 50 iterations of the Rabin-Miller test. 782859191Skris 782959191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 783059191Skris iterations of the Rabin-Miller test as required by the appendix 783159191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 783259191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 783359191Skris generation becomes much faster. 783459191Skris 783559191Skris This implies a change for the callback functions in DSA_is_prime 783659191Skris and DSA_generate_parameters: The callback function is called once 783759191Skris for each positive witness in the Rabin-Miller test, not just 783859191Skris occasionally in the inner loop; and the parameters to the 783959191Skris callback function now provide an iteration count for the outer 784059191Skris loop rather than for the current invocation of the inner loop. 784159191Skris DSA_generate_parameters additionally can call the callback 784259191Skris function with an 'iteration count' of -1, meaning that a 784359191Skris candidate has passed the trial division test (when q is generated 784459191Skris from an application-provided seed, trial division is skipped). 784559191Skris [Bodo Moeller] 784659191Skris 784759191Skris *) New function BN_is_prime_fasttest that optionally does trial 784859191Skris division before starting the Rabin-Miller test and has 784959191Skris an additional BN_CTX * argument (whereas BN_is_prime always 785059191Skris has to allocate at least one BN_CTX). 785159191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 785259191Skris trial division stage. 785359191Skris [Bodo Moeller] 785459191Skris 785559191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 785659191Skris as ASN1_TIME. 785759191Skris [Steve Henson] 785859191Skris 785959191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 786059191Skris [Steve Henson] 786159191Skris 786259191Skris *) New function BN_pseudo_rand(). 786359191Skris [Ulf M�ller] 786459191Skris 786559191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 786659191Skris bignum version of BN_from_montgomery() with the working code from 786759191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 786859191Skris the comments. 786959191Skris [Ulf M�ller] 787059191Skris 787159191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 787259191Skris made it impossible to use the same SSL_SESSION data structure in 787359191Skris SSL2 clients in multiple threads. 787459191Skris [Bodo Moeller] 787559191Skris 787659191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 787759191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 787859191Skris to seed the PRNG (previously an explicit byte count was required). 787959191Skris [Ulf M�ller, Bodo M�ller] 788059191Skris 788159191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 788259191Skris used (char *) instead of (void *) and had casts all over the place. 788359191Skris [Steve Henson] 788459191Skris 788559191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 788659191Skris [Ulf M�ller] 788759191Skris 788859191Skris *) Retain source code compatibility for BN_prime_checks macro: 788959191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 789059191Skris BN_prime_checks_for_size to determine the appropriate number of 789159191Skris Rabin-Miller iterations. 789259191Skris [Ulf M�ller] 789359191Skris 789459191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 789559191Skris DH_CHECK_P_NOT_SAFE_PRIME. 789659191Skris (Check if this is true? OpenPGP calls them "strong".) 789759191Skris [Ulf M�ller] 789859191Skris 789959191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 790059191Skris "dhparam". The old programs are retained for now but will handle DH keys 790159191Skris (instead of parameters) in future. 790259191Skris [Steve Henson] 790359191Skris 790459191Skris *) Make the ciphers, s_server and s_client programs check the return values 790559191Skris when a new cipher list is set. 790659191Skris [Steve Henson] 790759191Skris 790859191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 790959191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 791059191Skris wrong. 791159191Skris 791259191Skris The syntax for the cipher sorting has been extended to support sorting by 791359191Skris cipher-strength (using the strength_bits hard coded in the tables). 791459191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 791559191Skris 791659191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 791759191Skris string with an "undefined" symbol (neither command nor alphanumeric 791859191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 791959191Skris an error is flagged. 792059191Skris 792159191Skris Due to the strength-sorting extension, the code of the 792259191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 792359191Skris the readability was also increased :-) 792459191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 792559191Skris 792659191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 792759191Skris for the first serial number and places 2 in the serial number file. This 792859191Skris avoids problems when the root CA is created with serial number zero and 792959191Skris the first user certificate has the same issuer name and serial number 793059191Skris as the root CA. 793159191Skris [Steve Henson] 793259191Skris 793359191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 793459191Skris the new code. Add documentation for this stuff. 793559191Skris [Steve Henson] 793659191Skris 793759191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 793859191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 793959191Skris structures and behave in an analagous way to the X509v3 functions: 794059191Skris they shouldn't be called directly but wrapper functions should be used 794159191Skris instead. 794259191Skris 794359191Skris So we also now have some wrapper functions that call the X509at functions 794459191Skris when passed certificate requests. (TO DO: similar things can be done with 794559191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 794659191Skris things. Some of these need some d2i or i2d and print functionality 794759191Skris because they handle more complex structures.) 794859191Skris [Steve Henson] 794959191Skris 795059191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 795159191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 795259191Skris NO_RSA in ssl/s2*.c. 795359191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 795459191Skris 795559191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 795659191Skris has a return value which indicates the quality of the random data 795759191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 795859191Skris error queue. New function RAND_pseudo_bytes() generates output that is 795959191Skris guaranteed to be unique but not unpredictable. RAND_add is like 796059191Skris RAND_seed, but takes an extra argument for an entropy estimate 796159191Skris (RAND_seed always assumes full entropy). 796259191Skris [Ulf M�ller] 796359191Skris 796459191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 796559191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 796659191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 796759191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 796859191Skris false-positive rate of at most 2^-80 for random input. 796959191Skris [Bodo Moeller] 797059191Skris 797159191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 797259191Skris [Bodo Moeller] 797359191Skris 797459191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 797559191Skris in the 0.9.5 release), this returns the chain 797659191Skris from an X509_CTX structure with a dup of the stack and all 797759191Skris the X509 reference counts upped: so the stack will exist 797859191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 797959191Skris to use this. 798059191Skris 798159191Skris Also make SSL_SESSION_print() print out the verify return 798259191Skris code. 798359191Skris [Steve Henson] 798459191Skris 798559191Skris *) Add manpage for the pkcs12 command. Also change the default 798659191Skris behaviour so MAC iteration counts are used unless the new 798759191Skris -nomaciter option is used. This improves file security and 798859191Skris only older versions of MSIE (4.0 for example) need it. 798959191Skris [Steve Henson] 799059191Skris 799159191Skris *) Honor the no-xxx Configure options when creating .DEF files. 799259191Skris [Ulf M�ller] 799359191Skris 799459191Skris *) Add PKCS#10 attributes to field table: challengePassword, 799559191Skris unstructuredName and unstructuredAddress. These are taken from 799659191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 799759191Skris international characters are used. 799859191Skris 799959191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 800059191Skris based on strings. Remove the 'loc' parameter when adding 800159191Skris attributes because these will be a SET OF encoding which is sorted 800259191Skris in ASN1 order. 800359191Skris [Steve Henson] 800459191Skris 800559191Skris *) Initial changes to the 'req' utility to allow request generation 800659191Skris automation. This will allow an application to just generate a template 800759191Skris file containing all the field values and have req construct the 800859191Skris request. 800959191Skris 801059191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 801159191Skris used all over the place including certificate requests and PKCS#7 801259191Skris structures. They are currently handled manually where necessary with 801359191Skris some primitive wrappers for PKCS#7. The new functions behave in a 801459191Skris manner analogous to the X509 extension functions: they allow 801559191Skris attributes to be looked up by NID and added. 801659191Skris 801759191Skris Later something similar to the X509V3 code would be desirable to 801859191Skris automatically handle the encoding, decoding and printing of the 801959191Skris more complex types. The string types like challengePassword can 802059191Skris be handled by the string table functions. 802159191Skris 802259191Skris Also modified the multi byte string table handling. Now there is 802359191Skris a 'global mask' which masks out certain types. The table itself 802459191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 802559191Skris is useful when for example there is only one permissible type 802659191Skris (as in countryName) and using the mask might result in no valid 802759191Skris types at all. 802859191Skris [Steve Henson] 802959191Skris 803059191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 803159191Skris SSL_get_peer_finished to allow applications to obtain the latest 803259191Skris Finished messages sent to the peer or expected from the peer, 803359191Skris respectively. (SSL_get_peer_finished is usually the Finished message 803459191Skris actually received from the peer, otherwise the protocol will be aborted.) 803559191Skris 803659191Skris As the Finished message are message digests of the complete handshake 803759191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 803859191Skris be used for external authentication procedures when the authentication 803959191Skris provided by SSL/TLS is not desired or is not enough. 804059191Skris [Bodo Moeller] 804159191Skris 804259191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 804359191Skris the host supports BWX extension and if Compaq C is present on the 804459191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 804559191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 804659191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 804759191Skris SHA1. 804859191Skris [Andy Polyakov] 804959191Skris 805059191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 805159191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 805259191Skris weak crypto and after checking the certificate is SGC a second one 805359191Skris with strong crypto. MS SGC stops the first handshake after receiving 805459191Skris the server certificate message and sends a second client hello. Since 805559191Skris a server will typically do all the time consuming operations before 805659191Skris expecting any further messages from the client (server key exchange 805759191Skris is the most expensive) there is little difference between the two. 805859191Skris 805959191Skris To get OpenSSL to support MS SGC we have to permit a second client 806059191Skris hello message after we have sent server done. In addition we have to 806159191Skris reset the MAC if we do get this second client hello. 806259191Skris [Steve Henson] 806359191Skris 806459191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 806559191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 806659191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 806759191Skris format DER encoded private key. Newer code should use PKCS#8 format which 806859191Skris has the key type encoded in the ASN1 structure. Added DER private key 806959191Skris support to pkcs8 application. 807059191Skris [Steve Henson] 807159191Skris 807259191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 807359191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 807459191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 807559191Skris is set, we interpret this as a request to violate the specification 807659191Skris (the worst that can happen is a handshake failure, and 'correct' 807759191Skris behaviour would result in a handshake failure anyway). 807859191Skris [Bodo Moeller] 807959191Skris 808059191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 808159191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 808259191Skris concurrently obtain them from an external cache). 808359191Skris The internal cache can handle only one SSL_SESSION with a given ID, 808459191Skris so if there's a conflict, we now throw out the old one to achieve 808559191Skris consistency. 808659191Skris [Bodo Moeller] 808759191Skris 808859191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 808959191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 809059191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 809159191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 809259191Skris example. 809359191Skris [Steve Henson] 809459191Skris 809559191Skris *) Simplify the trust setting structure and code. Now we just have 809659191Skris two sequences of OIDs for trusted and rejected settings. These will 809759191Skris typically have values the same as the extended key usage extension 809859191Skris and any application specific purposes. 809959191Skris 810059191Skris The trust checking code now has a default behaviour: it will just 810159191Skris check for an object with the same NID as the passed id. Functions can 810259191Skris be provided to override either the default behaviour or the behaviour 810359191Skris for a given id. SSL client, server and email already have functions 810459191Skris in place for compatibility: they check the NID and also return "trusted" 810559191Skris if the certificate is self signed. 810659191Skris [Steve Henson] 810759191Skris 810859191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 810959191Skris traditional format into an EVP_PKEY structure. 811059191Skris [Steve Henson] 811159191Skris 811259191Skris *) Add a password callback function PEM_cb() which either prompts for 811359191Skris a password if usr_data is NULL or otherwise assumes it is a null 811459191Skris terminated password. Allow passwords to be passed on command line 811559191Skris environment or config files in a few more utilities. 811659191Skris [Steve Henson] 811759191Skris 811859191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 811959191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 812059191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 812159191Skris Update documentation. 812259191Skris [Steve Henson] 812359191Skris 812459191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 812559191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 812659191Skris and produce an error if it couldn't. For compatibility we also have 812759191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 812859191Skris don't allocate anything because they don't need to. 812959191Skris [Steve Henson] 813059191Skris 813159191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 813259191Skris for details. 813359191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 813459191Skris 813559191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 813659191Skris possibly others as well. The purpose is to make an interface that 813759191Skris provide hooks so anyone can build a separate set of allocation and 813859191Skris deallocation routines to be used by OpenSSL, for example memory 813959191Skris pool implementations, or something else, which was previously hard 814059191Skris since Malloc(), Realloc() and Free() were defined as macros having 814159191Skris the values malloc, realloc and free, respectively (except for Win32 814259191Skris compilations). The same is provided for memory debugging code. 814359191Skris OpenSSL already comes with functionality to find memory leaks, but 814459191Skris this gives people a chance to debug other memory problems. 814559191Skris 814659191Skris With these changes, a new set of functions and macros have appeared: 814759191Skris 814868651Skris CRYPTO_set_mem_debug_functions() [F] 814968651Skris CRYPTO_get_mem_debug_functions() [F] 815068651Skris CRYPTO_dbg_set_options() [F] 815168651Skris CRYPTO_dbg_get_options() [F] 815268651Skris CRYPTO_malloc_debug_init() [M] 815359191Skris 815459191Skris The memory debug functions are NULL by default, unless the library 815559191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 815659191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 815759191Skris gives the standard debugging functions that come with OpenSSL) or 815859191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 815959191Skris provided by the library user) must be used. When the standard 816059191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 816159191Skris request additional information: 816259191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 816359191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 816459191Skris 816559191Skris Also, things like CRYPTO_set_mem_functions will always give the 816659191Skris expected result (the new set of functions is used for allocation 816759191Skris and deallocation) at all times, regardless of platform and compiler 816859191Skris options. 816959191Skris 817059191Skris To finish it up, some functions that were never use in any other 817159191Skris way than through macros have a new API and new semantic: 817259191Skris 817359191Skris CRYPTO_dbg_malloc() 817459191Skris CRYPTO_dbg_realloc() 817559191Skris CRYPTO_dbg_free() 817659191Skris 817759191Skris All macros of value have retained their old syntax. 817859191Skris [Richard Levitte and Bodo Moeller] 817959191Skris 818059191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 818159191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 818259191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 818359191Skris algorithm. 818459191Skris [Steve Henson] 818559191Skris 818659191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 818759191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 818859191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 818959191Skris 819059191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 819159191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 819259191Skris functionality to handle multipart/signed properly) and a utility 819359191Skris called 'smime' to call all this stuff. This is based on code I 819459191Skris originally wrote for Celo who have kindly allowed it to be 819559191Skris included in OpenSSL. 819659191Skris [Steve Henson] 819759191Skris 819859191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 819959191Skris des_set_key (aka des_key_sched). Global variable des_check_key 820059191Skris decides which of these is called by des_set_key; this way 820159191Skris des_check_key behaves as it always did, but applications and 820259191Skris the library itself, which was buggy for des_check_key == 1, 820359191Skris have a cleaner way to pick the version they need. 820459191Skris [Bodo Moeller] 820559191Skris 820659191Skris *) New function PKCS12_newpass() which changes the password of a 820759191Skris PKCS12 structure. 820859191Skris [Steve Henson] 820959191Skris 821059191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 821159191Skris dynamic mix. In both cases the ids can be used as an index into the 821259191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 821359191Skris functions so they accept a list of the field values and the 821459191Skris application doesn't need to directly manipulate the X509_TRUST 821559191Skris structure. 821659191Skris [Steve Henson] 821759191Skris 821859191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 821959191Skris need initialising. 822059191Skris [Steve Henson] 822159191Skris 822259191Skris *) Modify the way the V3 extension code looks up extensions. This now 822359191Skris works in a similar way to the object code: we have some "standard" 822459191Skris extensions in a static table which is searched with OBJ_bsearch() 822559191Skris and the application can add dynamic ones if needed. The file 822659191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 822759191Skris updated whenever a new extension is added to the core code and kept 822859191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 822959191Skris this. New extensions are not added too often so this file can readily 823059191Skris be maintained manually. 823159191Skris 823259191Skris There are two big advantages in doing things this way. The extensions 823359191Skris can be looked up immediately and no longer need to be "added" using 823459191Skris X509V3_add_standard_extensions(): this function now does nothing. 823559191Skris [Side note: I get *lots* of email saying the extension code doesn't 823659191Skris work because people forget to call this function] 823759191Skris Also no dynamic allocation is done unless new extensions are added: 823859191Skris so if we don't add custom extensions there is no need to call 823959191Skris X509V3_EXT_cleanup(). 824059191Skris [Steve Henson] 824159191Skris 824259191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 824359191Skris magic header, so unsalted files fail gracefully instead of just decrypting 824459191Skris to garbage. This is because not salting is a big security hole, so people 824559191Skris should be discouraged from doing it. 824659191Skris [Ben Laurie] 824759191Skris 824859191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 824959191Skris digest to be passed on the command line but it only used this 825059191Skris parameter when signing a certificate. Modified so all relevant 825159191Skris operations are affected by the digest parameter including the 825259191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 825359191Skris DSA key was used because it didn't fix the digest. 825459191Skris [Steve Henson] 825559191Skris 825659191Skris *) Initial certificate chain verify code. Currently tests the untrusted 825759191Skris certificates for consistency with the verify purpose (which is set 825859191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 825959191Skris 826059191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 826159191Skris this is because it will reject chains with invalid extensions whereas 826259191Skris every previous version of OpenSSL and SSLeay made no checks at all. 826359191Skris 826459191Skris Trust code: checks the root CA for the relevant trust settings. Trust 826559191Skris settings have an initial value consistent with the verify purpose: e.g. 826659191Skris if the verify purpose is for SSL client use it expects the CA to be 826759191Skris trusted for SSL client use. However the default value can be changed to 826859191Skris permit custom trust settings: one example of this would be to only trust 826959191Skris certificates from a specific "secure" set of CAs. 827059191Skris 827159191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 827259191Skris which should be used for version portability: especially since the 827359191Skris verify structure is likely to change more often now. 827459191Skris 827559191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 827659191Skris to set them. If not set then assume SSL clients will verify SSL servers 827759191Skris and vice versa. 827859191Skris 827959191Skris Two new options to the verify program: -untrusted allows a set of 828059191Skris untrusted certificates to be passed in and -purpose which sets the 828159191Skris intended purpose of the certificate. If a purpose is set then the 828259191Skris new chain verify code is used to check extension consistency. 828359191Skris [Steve Henson] 828459191Skris 828559191Skris *) Support for the authority information access extension. 828659191Skris [Steve Henson] 828759191Skris 828859191Skris *) Modify RSA and DSA PEM read routines to transparently handle 828959191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 829059191Skris public keys in a format compatible with certificate 829159191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 829259191Skris functions called *_PublicKey_* which used various odd formats so 829359191Skris these are retained for compatibility: however the DSA variants were 829459191Skris never in a public release so they have been deleted. Changed dsa/rsa 829559191Skris utilities to handle the new format: note no releases ever handled public 829659191Skris keys so we should be OK. 829759191Skris 829859191Skris The primary motivation for this change is to avoid the same fiasco 829959191Skris that dogs private keys: there are several incompatible private key 830059191Skris formats some of which are standard and some OpenSSL specific and 830159191Skris require various evil hacks to allow partial transparent handling and 830259191Skris even then it doesn't work with DER formats. Given the option anything 830359191Skris other than PKCS#8 should be dumped: but the other formats have to 830459191Skris stay in the name of compatibility. 830559191Skris 830659191Skris With public keys and the benefit of hindsight one standard format 830759191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 830859191Skris it clearly returns an error if you try to read the wrong kind of key. 830959191Skris 831059191Skris Added a -pubkey option to the 'x509' utility to output the public key. 831159191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 831259191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 831359191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 831459191Skris that do the same as the EVP_PKEY_assign_*() except they up the 831559191Skris reference count of the added key (they don't "swallow" the 831659191Skris supplied key). 831759191Skris [Steve Henson] 831859191Skris 831959191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 832059191Skris CRLs would fail if the file contained no certificates or no CRLs: 832159191Skris added a new function to read in both types and return the number 832259191Skris read: this means that if none are read it will be an error. The 832359191Skris DER versions of the certificate and CRL reader would always fail 832459191Skris because it isn't possible to mix certificates and CRLs in DER format 832559191Skris without choking one or the other routine. Changed this to just read 832659191Skris a certificate: this is the best we can do. Also modified the code 832759191Skris in apps/verify.c to take notice of return codes: it was previously 832859191Skris attempting to read in certificates from NULL pointers and ignoring 832959191Skris any errors: this is one reason why the cert and CRL reader seemed 833059191Skris to work. It doesn't check return codes from the default certificate 833159191Skris routines: these may well fail if the certificates aren't installed. 833259191Skris [Steve Henson] 833359191Skris 833459191Skris *) Code to support otherName option in GeneralName. 833559191Skris [Steve Henson] 833659191Skris 833759191Skris *) First update to verify code. Change the verify utility 833859191Skris so it warns if it is passed a self signed certificate: 833959191Skris for consistency with the normal behaviour. X509_verify 834059191Skris has been modified to it will now verify a self signed 834159191Skris certificate if *exactly* the same certificate appears 834259191Skris in the store: it was previously impossible to trust a 834359191Skris single self signed certificate. This means that: 834459191Skris openssl verify ss.pem 834559191Skris now gives a warning about a self signed certificate but 834659191Skris openssl verify -CAfile ss.pem ss.pem 834759191Skris is OK. 834859191Skris [Steve Henson] 834959191Skris 835059191Skris *) For servers, store verify_result in SSL_SESSION data structure 835159191Skris (and add it to external session representation). 835259191Skris This is needed when client certificate verifications fails, 835359191Skris but an application-provided verification callback (set by 835459191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 835559191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 835659191Skris but returns 1): When the session is reused, we have to set 835759191Skris ssl->verify_result to the appropriate error code to avoid 835859191Skris security holes. 835959191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 836059191Skris 836159191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 836259191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 836359191Skris didn't contain any existing data because it was being created. 836459191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 836559191Skris 836659191Skris *) Add a salt to the key derivation routines in enc.c. This 836759191Skris forms the first 8 bytes of the encrypted file. Also add a 836859191Skris -S option to allow a salt to be input on the command line. 836959191Skris [Steve Henson] 837059191Skris 837159191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 837259191Skris to compare two certificates. We do this by working out the SHA1 837359191Skris hash and comparing that. X509_cmp() will be needed by the trust 837459191Skris code. 837559191Skris [Steve Henson] 837659191Skris 837759191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 837859191Skris the reference count in the SSL_SESSION returned. 837959191Skris [Geoff Thorpe <geoff@eu.c2.net>] 838059191Skris 838159191Skris *) Fix for 'req': it was adding a null to request attributes. 838259191Skris Also change the X509_LOOKUP and X509_INFO code to handle 838359191Skris certificate auxiliary information. 838459191Skris [Steve Henson] 838559191Skris 838659191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 838759191Skris the 'enc' command. 838859191Skris [Steve Henson] 838959191Skris 839059191Skris *) Add the possibility to add extra information to the memory leak 839159191Skris detecting output, to form tracebacks, showing from where each 839259191Skris allocation was originated: CRYPTO_push_info("constant string") adds 839359191Skris the string plus current file name and line number to a per-thread 839459191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 839559191Skris is like calling CYRPTO_pop_info() until the stack is empty. 839659191Skris Also updated memory leak detection code to be multi-thread-safe. 839759191Skris [Richard Levitte] 839859191Skris 839959191Skris *) Add options -text and -noout to pkcs7 utility and delete the 840059191Skris encryption options which never did anything. Update docs. 840159191Skris [Steve Henson] 840259191Skris 840359191Skris *) Add options to some of the utilities to allow the pass phrase 840459191Skris to be included on either the command line (not recommended on 840559191Skris OSes like Unix) or read from the environment. Update the 840659191Skris manpages and fix a few bugs. 840759191Skris [Steve Henson] 840859191Skris 840959191Skris *) Add a few manpages for some of the openssl commands. 841059191Skris [Steve Henson] 841159191Skris 841259191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 841359191Skris leaking and not finding already revoked certificates. 841459191Skris [Steve Henson] 841559191Skris 841659191Skris *) Extensive changes to support certificate auxiliary information. 841759191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 841859191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 841959191Skris can still read in a certificate file in the usual way but it 842059191Skris will also read in any additional "auxiliary information". By 842159191Skris doing things this way a fair degree of compatibility can be 842259191Skris retained: existing certificates can have this information added 842359191Skris using the new 'x509' options. 842459191Skris 842559191Skris Current auxiliary information includes an "alias" and some trust 842659191Skris settings. The trust settings will ultimately be used in enhanced 842759191Skris certificate chain verification routines: currently a certificate 842859191Skris can only be trusted if it is self signed and then it is trusted 842959191Skris for all purposes. 843059191Skris [Steve Henson] 843159191Skris 843259191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 843359191Skris The problem was that one of the replacement routines had not been working 843459191Skris since SSLeay releases. For now the offending routine has been replaced 843559191Skris with non-optimised assembler. Even so, this now gives around 95% 843659191Skris performance improvement for 1024 bit RSA signs. 843759191Skris [Mark Cox] 843859191Skris 843959191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 844059191Skris handling. Most clients have the effective key size in bits equal to 844159191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 844259191Skris A few however don't do this and instead use the size of the decrypted key 844359191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 844459191Skris the effective key length. In this case the effective key length can still 844559191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 844659191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 844759191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 844859191Skris the key length and effective key length are equal. 844959191Skris [Steve Henson] 845059191Skris 845159191Skris *) Add a bunch of functions that should simplify the creation of 845259191Skris X509_NAME structures. Now you should be able to do: 845359191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 845459191Skris and have it automatically work out the correct field type and fill in 845559191Skris the structures. The more adventurous can try: 845659191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 845759191Skris and it will (hopefully) work out the correct multibyte encoding. 845859191Skris [Steve Henson] 845959191Skris 846059191Skris *) Change the 'req' utility to use the new field handling and multibyte 846159191Skris copy routines. Before the DN field creation was handled in an ad hoc 846259191Skris way in req, ca, and x509 which was rather broken and didn't support 846359191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 846459191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 846559191Skris using the dirstring_type option. See the new comment in the default 846659191Skris openssl.cnf for more info. 846759191Skris [Steve Henson] 846859191Skris 846959191Skris *) Make crypto/rand/md_rand.c more robust: 847059191Skris - Assure unique random numbers after fork(). 847159191Skris - Make sure that concurrent threads access the global counter and 847259191Skris md serializably so that we never lose entropy in them 847359191Skris or use exactly the same state in multiple threads. 847459191Skris Access to the large state is not always serializable because 847559191Skris the additional locking could be a performance killer, and 847659191Skris md should be large enough anyway. 847759191Skris [Bodo Moeller] 847859191Skris 847959191Skris *) New file apps/app_rand.c with commonly needed functionality 848059191Skris for handling the random seed file. 848159191Skris 848259191Skris Use the random seed file in some applications that previously did not: 848359191Skris ca, 848459191Skris dsaparam -genkey (which also ignored its '-rand' option), 848559191Skris s_client, 848659191Skris s_server, 848759191Skris x509 (when signing). 848859191Skris Except on systems with /dev/urandom, it is crucial to have a random 848959191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 849059191Skris for RSA signatures we could do without one. 849159191Skris 849259191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 849359191Skris of each file listed in the '-rand' option. The function as previously 849459191Skris found in genrsa is now in app_rand.c and is used by all programs 849559191Skris that support '-rand'. 849659191Skris [Bodo Moeller] 849759191Skris 849859191Skris *) In RAND_write_file, use mode 0600 for creating files; 849959191Skris don't just chmod when it may be too late. 850059191Skris [Bodo Moeller] 850159191Skris 850259191Skris *) Report an error from X509_STORE_load_locations 850359191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 850459191Skris [Bill Perry] 850559191Skris 850659191Skris *) New function ASN1_mbstring_copy() this copies a string in either 850759191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 850859191Skris into an ASN1_STRING type. A mask of permissible types is passed 850959191Skris and it chooses the "minimal" type to use or an error if not type 851059191Skris is suitable. 851159191Skris [Steve Henson] 851259191Skris 851359191Skris *) Add function equivalents to the various macros in asn1.h. The old 851459191Skris macros are retained with an M_ prefix. Code inside the library can 851559191Skris use the M_ macros. External code (including the openssl utility) 851659191Skris should *NOT* in order to be "shared library friendly". 851759191Skris [Steve Henson] 851859191Skris 851959191Skris *) Add various functions that can check a certificate's extensions 852059191Skris to see if it usable for various purposes such as SSL client, 852159191Skris server or S/MIME and CAs of these types. This is currently 852259191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 852359191Skris verification. Also added a -purpose flag to x509 utility to 852459191Skris print out all the purposes. 852559191Skris [Steve Henson] 852659191Skris 852759191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 852859191Skris functions. 852959191Skris [Steve Henson] 853059191Skris 853159191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 853259191Skris for, obtain and decode and extension and obtain its critical flag. 853359191Skris This allows all the necessary extension code to be handled in a 853459191Skris single function call. 853559191Skris [Steve Henson] 853659191Skris 853759191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 853859191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 853959191Skris [Andy Polyakov] 854059191Skris 854159191Skris *) New -noout option to asn1parse. This causes no output to be produced 854259191Skris its main use is when combined with -strparse and -out to extract data 854359191Skris from a file (which may not be in ASN.1 format). 854459191Skris [Steve Henson] 854559191Skris 854659191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 854759191Skris when producing the local key id. 854859191Skris [Richard Levitte <levitte@stacken.kth.se>] 854959191Skris 855059191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 855159191Skris stated explicitly. If it is not stated then it tries the first server 855259191Skris certificate file. The previous behaviour hard coded the filename 855359191Skris "server.pem". 855459191Skris [Steve Henson] 855559191Skris 855659191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 855759191Skris a public key to be input or output. For example: 855859191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 855959191Skris Also added necessary DSA public key functions to handle this. 856059191Skris [Steve Henson] 856159191Skris 856259191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 856359191Skris in the message. This was handled by allowing 856459191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 856559191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 856659191Skris 856759191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 856859191Skris to the end of the strings whereas this didn't. This would cause problems 856959191Skris if strings read with d2i_ASN1_bytes() were later modified. 857059191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 857159191Skris 857259191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 857359191Skris data and it contains EOF it will end up returning an error. This is 857459191Skris caused by input 46 bytes long. The cause is due to the way base64 857559191Skris BIOs find the start of base64 encoded data. They do this by trying a 857659191Skris trial decode on each line until they find one that works. When they 857759191Skris do a flag is set and it starts again knowing it can pass all the 857859191Skris data directly through the decoder. Unfortunately it doesn't reset 857959191Skris the context it uses. This means that if EOF is reached an attempt 858059191Skris is made to pass two EOFs through the context and this causes the 858159191Skris resulting error. This can also cause other problems as well. As is 858259191Skris usual with these problems it takes *ages* to find and the fix is 858359191Skris trivial: move one line. 858459191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 858559191Skris 858659191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 858759191Skris old code wouldn't work because it needed to select() on sockets and the 858859191Skris tty (for keypresses and to see if data could be written). Win32 only 858959191Skris supports select() on sockets so we select() with a 1s timeout on the 859059191Skris sockets and then see if any characters are waiting to be read, if none 859159191Skris are present then we retry, we also assume we can always write data to 859259191Skris the tty. This isn't nice because the code then blocks until we've 859359191Skris received a complete line of data and it is effectively polling the 859459191Skris keyboard at 1s intervals: however it's quite a bit better than not 859559191Skris working at all :-) A dedicated Windows application might handle this 859659191Skris with an event loop for example. 859759191Skris [Steve Henson] 859859191Skris 859959191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 860059191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 860159191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 860259191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 860359191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 860459191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 860559191Skris This necessitated the support of an extra signature type NID_md5_sha1 860659191Skris for SSL signatures and modifications to the SSL library to use it instead 860759191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 860859191Skris [Steve Henson] 860959191Skris 861059191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 861159191Skris will lookup a CRL issuers certificate and verify the signature in a 861259191Skris similar way to the verify program. Tidy up the crl program so it 861359191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 861459191Skris less strict. It will now permit CRL extensions even if it is not 861559191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 861659191Skris [Steve Henson] 861759191Skris 861859191Skris *) Initialize all non-automatic variables each time one of the openssl 861959191Skris sub-programs is started (this is necessary as they may be started 862059191Skris multiple times from the "OpenSSL>" prompt). 862159191Skris [Lennart Bang, Bodo Moeller] 862259191Skris 862359191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 862459191Skris removing all other RSA functionality (this is what NO_RSA does). This 862559191Skris is so (for example) those in the US can disable those operations covered 862659191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 862759191Skris key generation. 862859191Skris [Steve Henson] 862959191Skris 863059191Skris *) Non-copying interface to BIO pairs. 863159191Skris (still largely untested) 863259191Skris [Bodo Moeller] 863359191Skris 863459191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 863559191Skris ASCII string. This was handled independently in various places before. 863659191Skris [Steve Henson] 863759191Skris 863859191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 863959191Skris UTF8 strings a character at a time. 864059191Skris [Steve Henson] 864159191Skris 864259191Skris *) Use client_version from client hello to select the protocol 864359191Skris (s23_srvr.c) and for RSA client key exchange verification 864459191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 864559191Skris [Bodo Moeller] 864659191Skris 864759191Skris *) Add various utility functions to handle SPKACs, these were previously 864859191Skris handled by poking round in the structure internals. Added new function 864959191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 865059191Skris print, verify and generate SPKACs. Based on an original idea from 865159191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 865259191Skris [Steve Henson] 865359191Skris 865459191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 865559191Skris [Andy Polyakov] 865659191Skris 865759191Skris *) Allow the config file extension section to be overwritten on the 865859191Skris command line. Based on an original idea from Massimiliano Pala 865959191Skris <madwolf@comune.modena.it>. The new option is called -extensions 866059191Skris and can be applied to ca, req and x509. Also -reqexts to override 866159191Skris the request extensions in req and -crlexts to override the crl extensions 866259191Skris in ca. 866359191Skris [Steve Henson] 866459191Skris 866559191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 866659191Skris the same field multiple times by preceding it by "XXXX." for example: 866759191Skris 1.OU="Unit name 1" 866859191Skris 2.OU="Unit name 2" 866959191Skris this is the same syntax as used in the req config file. 867059191Skris [Steve Henson] 867159191Skris 867259191Skris *) Allow certificate extensions to be added to certificate requests. These 867359191Skris are specified in a 'req_extensions' option of the req section of the 867459191Skris config file. They can be printed out with the -text option to req but 867559191Skris are otherwise ignored at present. 867659191Skris [Steve Henson] 867759191Skris 867859191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 867959191Skris data read consists of only the final block it would not decrypted because 868059191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 868159191Skris A misplaced 'break' also meant the decrypted final block might not be 868259191Skris copied until the next read. 868359191Skris [Steve Henson] 868459191Skris 868559191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 868659191Skris a few extra parameters to the DH structure: these will be useful if 868759191Skris for example we want the value of 'q' or implement X9.42 DH. 868859191Skris [Steve Henson] 868959191Skris 869059191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 869159191Skris provides hooks that allow the default DSA functions or functions on a 869259191Skris "per key" basis to be replaced. This allows hardware acceleration and 869359191Skris hardware key storage to be handled without major modification to the 869459191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 869559191Skris associated functions. 869659191Skris [Steve Henson] 869759191Skris 869859191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 869959191Skris as "read only": it can't be written to and the buffer it points to will 870059191Skris not be freed. Reading from a read only BIO is much more efficient than 870159191Skris a normal memory BIO. This was added because there are several times when 870259191Skris an area of memory needs to be read from a BIO. The previous method was 870359191Skris to create a memory BIO and write the data to it, this results in two 870459191Skris copies of the data and an O(n^2) reading algorithm. There is a new 870559191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 870659191Skris an area of memory. Also modified the PKCS#7 routines to use read only 870759191Skris memory BIOs. 870859191Skris [Steve Henson] 870959191Skris 871059191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 871159191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 871259191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 871359191Skris but a retry condition occured while trying to read the rest. 871459191Skris [Bodo Moeller] 871559191Skris 871659191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 871759191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 871859191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 871959191Skris the encrypted data type: this is a more sensible place to put it and it 872059191Skris allows the PKCS#12 code to be tidied up that duplicated this 872159191Skris functionality. 872259191Skris [Steve Henson] 872359191Skris 872459191Skris *) Changed obj_dat.pl script so it takes its input and output files on 872559191Skris the command line. This should avoid shell escape redirection problems 872659191Skris under Win32. 872759191Skris [Steve Henson] 872859191Skris 872959191Skris *) Initial support for certificate extension requests, these are included 873059191Skris in things like Xenroll certificate requests. Included functions to allow 873159191Skris extensions to be obtained and added. 873259191Skris [Steve Henson] 873359191Skris 873459191Skris *) -crlf option to s_client and s_server for sending newlines as 873559191Skris CRLF (as required by many protocols). 873659191Skris [Bodo Moeller] 873759191Skris 873855714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 873955714Skris 874055714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 874155714Skris [Ralf S. Engelschall] 874255714Skris 874355714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 874455714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 874555714Skris 874655714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 874755714Skris program. 874855714Skris [Steve Henson] 874955714Skris 875055714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 875155714Skris DH parameters/keys (q is lost during that conversion, but the resulting 875255714Skris DH parameters contain its length). 875355714Skris 875455714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 875555714Skris much faster than DH_generate_parameters (which creates parameters 875655714Skris where p = 2*q + 1), and also the smaller q makes DH computations 875755714Skris much more efficient (160-bit exponentiation instead of 1024-bit 875855714Skris exponentiation); so this provides a convenient way to support DHE 875955714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 876055714Skris utter importance to use 876155714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 876255714Skris or 876355714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 876455714Skris when such DH parameters are used, because otherwise small subgroup 876555714Skris attacks may become possible! 876655714Skris [Bodo Moeller] 876755714Skris 876855714Skris *) Avoid memory leak in i2d_DHparams. 876955714Skris [Bodo Moeller] 877055714Skris 877155714Skris *) Allow the -k option to be used more than once in the enc program: 877255714Skris this allows the same encrypted message to be read by multiple recipients. 877355714Skris [Steve Henson] 877455714Skris 877555714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 877655714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 877755714Skris it will always use the numerical form of the OID, even if it has a short 877855714Skris or long name. 877955714Skris [Steve Henson] 878055714Skris 878155714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 878255714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 878355714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 878455714Skris no private key components need be present and it might store extra data 878559191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 878659191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 878759191Skris private key operations. 878855714Skris [Steve Henson] 878955714Skris 879055714Skris *) Added support for SPARC Linux. 879155714Skris [Andy Polyakov] 879255714Skris 879355714Skris *) pem_password_cb function type incompatibly changed from 879455714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 879555714Skris to 879655714Skris ....(char *buf, int size, int rwflag, void *userdata); 879755714Skris so that applications can pass data to their callbacks: 879855714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 879955714Skris additional void * argument, which is just handed through whenever 880055714Skris the password callback is called. 880159191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 880255714Skris 880355714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 880455714Skris 880555714Skris Compatibility note: As many C implementations push function arguments 880655714Skris onto the stack in reverse order, the new library version is likely to 880755714Skris interoperate with programs that have been compiled with the old 880855714Skris pem_password_cb definition (PEM_whatever takes some data that 880955714Skris happens to be on the stack as its last argument, and the callback 881055714Skris just ignores this garbage); but there is no guarantee whatsoever that 881155714Skris this will work. 881255714Skris 881355714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 881455714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 881555714Skris problems not only on Windows, but also on some Unix platforms. 881655714Skris To avoid problematic command lines, these definitions are now in an 881755714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 881855714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 881955714Skris [Bodo Moeller] 882055714Skris 882155714Skris *) MIPS III/IV assembler module is reimplemented. 882255714Skris [Andy Polyakov] 882355714Skris 882455714Skris *) More DES library cleanups: remove references to srand/rand and 882555714Skris delete an unused file. 882655714Skris [Ulf M�ller] 882755714Skris 882855714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 882955714Skris since not many people have MASM (ml) and it can be hard to obtain. 883055714Skris This is currently experimental but it seems to work OK and pass all 883155714Skris the tests. Check out INSTALL.W32 for info. 883255714Skris [Steve Henson] 883355714Skris 883455714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 883555714Skris without temporary keys kept an extra copy of the server key, 883655714Skris and connections with temporary keys did not free everything in case 883755714Skris of an error. 883855714Skris [Bodo Moeller] 883955714Skris 884055714Skris *) New function RSA_check_key and new openssl rsa option -check 884155714Skris for verifying the consistency of RSA keys. 884255714Skris [Ulf Moeller, Bodo Moeller] 884355714Skris 884455714Skris *) Various changes to make Win32 compile work: 884555714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 884655714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 884755714Skris comparison" warnings. 884855714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 884955714Skris [Steve Henson] 885055714Skris 885155714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 885255714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 885355714Skris derived keys are printed to stderr. 885455714Skris [Steve Henson] 885555714Skris 885655714Skris *) Copy the flags in ASN1_STRING_dup(). 885755714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 885855714Skris 885955714Skris *) The x509 application mishandled signing requests containing DSA 886055714Skris keys when the signing key was also DSA and the parameters didn't match. 886155714Skris 886255714Skris It was supposed to omit the parameters when they matched the signing key: 886355714Skris the verifying software was then supposed to automatically use the CA's 886455714Skris parameters if they were absent from the end user certificate. 886555714Skris 886655714Skris Omitting parameters is no longer recommended. The test was also 886755714Skris the wrong way round! This was probably due to unusual behaviour in 886855714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 886955714Skris This meant that parameters were omitted when they *didn't* match and 887055714Skris the certificate was useless. Certificates signed with 'ca' didn't have 887155714Skris this bug. 887255714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 887355714Skris 887455714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 887555714Skris The interface is as follows: 887655714Skris Applications can use 887755714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 887855714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 887955714Skris "off" is now the default. 888055714Skris The library internally uses 888155714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 888255714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 888355714Skris to disable memory-checking temporarily. 888455714Skris 888555714Skris Some inconsistent states that previously were possible (and were 888655714Skris even the default) are now avoided. 888755714Skris 888855714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 888955714Skris with each memory chunk allocated; this is occasionally more helpful 889055714Skris than just having a counter. 889155714Skris 889255714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 889355714Skris 889455714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 889555714Skris extensions. 889655714Skris [Bodo Moeller] 889755714Skris 889855714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 889955714Skris which largely parallels "options", but is for changing API behaviour, 890055714Skris whereas "options" are about protocol behaviour. 890155714Skris Initial "mode" flags are: 890255714Skris 890355714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 890455714Skris a single record has been written. 890555714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 890655714Skris retries use the same buffer location. 890755714Skris (But all of the contents must be 890855714Skris copied!) 890955714Skris [Bodo Moeller] 891055714Skris 891179998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 891255714Skris worked. 891355714Skris 891455714Skris *) Fix problems with no-hmac etc. 891555714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 891655714Skris 891755714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 891855714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 891955714Skris to mess around with the internals of an RSA structure. 892055714Skris [Steve Henson] 892155714Skris 892255714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 892355714Skris Also really enable memory leak checks in openssl.c and in some 892455714Skris test programs. 892555714Skris [Chad C. Mulligan, Bodo Moeller] 892655714Skris 892755714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 892855714Skris up the length of negative integers. This has now been simplified to just 892955714Skris store the length when it is first determined and use it later, rather 893055714Skris than trying to keep track of where data is copied and updating it to 893155714Skris point to the end. 893255714Skris [Steve Henson, reported by Brien Wheeler 893355714Skris <bwheeler@authentica-security.com>] 893455714Skris 893555714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 893655714Skris of a PKCS#7 signature but with the signing certificate passed to the 893755714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 893855714Skris certificate is present in the PKCS#7 structure. This isn't always the 893955714Skris case: certificates can be omitted from a PKCS#7 structure and be 894055714Skris distributed by "out of band" means (such as a certificate database). 894155714Skris [Steve Henson] 894255714Skris 894355714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 894455714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 894555714Skris necessary function names. 894655714Skris [Steve Henson] 894755714Skris 894855714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 894955714Skris options set by Configure in the top level Makefile, and Configure 895055714Skris was not even able to write more than one option correctly. 895155714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 895255714Skris [Bodo Moeller] 895355714Skris 895455714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 895555714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 895655714Skris for example allow memory BIOs to contain config info. 895755714Skris [Steve Henson] 895855714Skris 895955714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 896055714Skris Whoever hopes to achieve shared-library compatibility across versions 896155714Skris must use this, not the compile-time macro. 896255714Skris (Exercise 0.9.4: Which is the minimum library version required by 896355714Skris such programs?) 896455714Skris Note: All this applies only to multi-threaded programs, others don't 896555714Skris need locks. 896655714Skris [Bodo Moeller] 896755714Skris 896855714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 896955714Skris through a BIO pair triggered the default case, i.e. 897055714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 897155714Skris [Bodo Moeller] 897255714Skris 897355714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 897455714Skris can use the SSL library even if none of the specific BIOs is 897555714Skris appropriate. 897655714Skris [Bodo Moeller] 897755714Skris 897855714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 897955714Skris for the encoded length. 898055714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 898155714Skris 898255714Skris *) Add initial documentation of the X509V3 functions. 898355714Skris [Steve Henson] 898455714Skris 898555714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 898655714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 898755714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 898855714Skris secure PKCS#8 private key format with a high iteration count. 898955714Skris [Steve Henson] 899055714Skris 899155714Skris *) Fix determination of Perl interpreter: A perl or perl5 899255714Skris _directory_ in $PATH was also accepted as the interpreter. 899355714Skris [Ralf S. Engelschall] 899455714Skris 899555714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 899655714Skris wrong with it but it was very old and did things like calling 899755714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 899855714Skris unusual formatting. 899955714Skris [Steve Henson] 900055714Skris 900155714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 900255714Skris to use the new extension code. 900355714Skris [Steve Henson] 900455714Skris 900555714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 900655714Skris with macros. This should make it easier to change their form, add extra 900755714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 900855714Skris constant. 900955714Skris [Steve Henson] 901055714Skris 901155714Skris *) Add to configuration table a new entry that can specify an alternative 901255714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 901355714Skris according to Mark Crispin <MRC@Panda.COM>. 901455714Skris [Bodo Moeller] 901555714Skris 901655714Skris#if 0 901755714Skris *) DES CBC did not update the IV. Weird. 901855714Skris [Ben Laurie] 901955714Skris#else 902055714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 902155714Skris Changing the behaviour of the former might break existing programs -- 902255714Skris where IV updating is needed, des_ncbc_encrypt can be used. 902355714Skris#endif 902455714Skris 902555714Skris *) When bntest is run from "make test" it drives bc to check its 902655714Skris calculations, as well as internally checking them. If an internal check 902755714Skris fails, it needs to cause bc to give a non-zero result or make test carries 902855714Skris on without noticing the failure. Fixed. 902955714Skris [Ben Laurie] 903055714Skris 903155714Skris *) DES library cleanups. 903255714Skris [Ulf M�ller] 903355714Skris 903455714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 903555714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 903655714Skris ciphers. NOTE: although the key derivation function has been verified 903755714Skris against some published test vectors it has not been extensively tested 903855714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 903955714Skris of v2.0. 904055714Skris [Steve Henson] 904155714Skris 904255714Skris *) Instead of "mkdir -p", which is not fully portable, use new 904355714Skris Perl script "util/mkdir-p.pl". 904455714Skris [Bodo Moeller] 904555714Skris 904655714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 904755714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 904855714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 904955714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 905055714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 905155714Skris underlying key generation function so it must do its own ASN1 parsing. 905255714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 905355714Skris 'parameter' argument instead of literal salt and iteration count values 905455714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 905555714Skris [Steve Henson] 905655714Skris 905755714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 905855714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 905955714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 906055714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 906155714Skris value was just used as a "magic string" and not used directly its 906255714Skris value doesn't matter. 906355714Skris [Steve Henson] 906455714Skris 906555714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 906655714Skris support mutable. 906755714Skris [Ben Laurie] 906855714Skris 906955714Skris *) "linux-sparc64" configuration (ultrapenguin). 907055714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 907155714Skris "linux-sparc" configuration. 907255714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 907355714Skris 907455714Skris *) config now generates no-xxx options for missing ciphers. 907555714Skris [Ulf M�ller] 907655714Skris 907755714Skris *) Support the EBCDIC character set (work in progress). 907855714Skris File ebcdic.c not yet included because it has a different license. 907955714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 908055714Skris 908155714Skris *) Support BS2000/OSD-POSIX. 908255714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 908355714Skris 908455714Skris *) Make callbacks for key generation use void * instead of char *. 908555714Skris [Ben Laurie] 908655714Skris 908755714Skris *) Make S/MIME samples compile (not yet tested). 908855714Skris [Ben Laurie] 908955714Skris 909055714Skris *) Additional typesafe stacks. 909155714Skris [Ben Laurie] 909255714Skris 909355714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 909455714Skris [Bodo Moeller] 909555714Skris 909655714Skris 909755714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 909855714Skris 909955714Skris *) New configuration variant "sco5-gcc". 910055714Skris 910155714Skris *) Updated some demos. 910255714Skris [Sean O Riordain, Wade Scholine] 910355714Skris 910455714Skris *) Add missing BIO_free at exit of pkcs12 application. 910555714Skris [Wu Zhigang] 910655714Skris 910755714Skris *) Fix memory leak in conf.c. 910855714Skris [Steve Henson] 910955714Skris 911055714Skris *) Updates for Win32 to assembler version of MD5. 911155714Skris [Steve Henson] 911255714Skris 911355714Skris *) Set #! path to perl in apps/der_chop to where we found it 911455714Skris instead of using a fixed path. 911555714Skris [Bodo Moeller] 911655714Skris 911755714Skris *) SHA library changes for irix64-mips4-cc. 911855714Skris [Andy Polyakov] 911955714Skris 912055714Skris *) Improvements for VMS support. 912155714Skris [Richard Levitte] 912255714Skris 912355714Skris 912455714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 912555714Skris 912655714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 912755714Skris This also avoids the problems with SC4.2 and unpatched SC5. 912855714Skris [Andy Polyakov <appro@fy.chalmers.se>] 912955714Skris 913055714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 913155714Skris These are required because of the typesafe stack would otherwise break 913255714Skris existing code. If old code used a structure member which used to be STACK 913355714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 913455714Skris sk_num or sk_value it would produce an error because the num, data members 913555714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 913655714Skris replaces the old method of assigning a value to sk_value 913755714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 913855714Skris that does this will no longer work (and should use sk_set instead) but 913955714Skris this could be regarded as a "questionable" behaviour anyway. 914055714Skris [Steve Henson] 914155714Skris 914255714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 914355714Skris correctly handle encrypted S/MIME data. 914455714Skris [Steve Henson] 914555714Skris 914655714Skris *) Change type of various DES function arguments from des_cblock 914755714Skris (which means, in function argument declarations, pointer to char) 914855714Skris to des_cblock * (meaning pointer to array with 8 char elements), 914955714Skris which allows the compiler to do more typechecking; it was like 915055714Skris that back in SSLeay, but with lots of ugly casts. 915155714Skris 915255714Skris Introduce new type const_des_cblock. 915355714Skris [Bodo Moeller] 915455714Skris 915555714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 915655714Skris problems: find RecipientInfo structure that matches recipient certificate 915755714Skris and initialise the ASN1 structures properly based on passed cipher. 915855714Skris [Steve Henson] 915955714Skris 916055714Skris *) Belatedly make the BN tests actually check the results. 916155714Skris [Ben Laurie] 916255714Skris 916355714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 916455714Skris to and from BNs: it was completely broken. New compilation option 916555714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 916655714Skris key elements as negative integers. 916755714Skris [Steve Henson] 916855714Skris 916955714Skris *) Reorganize and speed up MD5. 917055714Skris [Andy Polyakov <appro@fy.chalmers.se>] 917155714Skris 917255714Skris *) VMS support. 917355714Skris [Richard Levitte <richard@levitte.org>] 917455714Skris 917555714Skris *) New option -out to asn1parse to allow the parsed structure to be 917655714Skris output to a file. This is most useful when combined with the -strparse 917755714Skris option to examine the output of things like OCTET STRINGS. 917855714Skris [Steve Henson] 917955714Skris 918055714Skris *) Make SSL library a little more fool-proof by not requiring any longer 918155714Skris that SSL_set_{accept,connect}_state be called before 918255714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 918355714Skris in many applications because usually everything *appeared* to work as 918455714Skris intended anyway -- now it really works as intended). 918555714Skris [Bodo Moeller] 918655714Skris 918755714Skris *) Move openssl.cnf out of lib/. 918855714Skris [Ulf M�ller] 918955714Skris 919055714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 919155714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 919255714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 919355714Skris [Ralf S. Engelschall] 919455714Skris 919555714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 919655714Skris handle PKCS#7 enveloped data properly. 919755714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 919855714Skris 919955714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 920055714Skris copying pointers. The cert_st handling is changed by this in 920155714Skris various ways (and thus what used to be known as ctx->default_cert 920255714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 920355714Skris any longer when s->cert does not give us what we need). 920455714Skris ssl_cert_instantiate becomes obsolete by this change. 920555714Skris As soon as we've got the new code right (possibly it already is?), 920655714Skris we have solved a couple of bugs of the earlier code where s->cert 920755714Skris was used as if it could not have been shared with other SSL structures. 920855714Skris 920955714Skris Note that using the SSL API in certain dirty ways now will result 921055714Skris in different behaviour than observed with earlier library versions: 921155714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 921255714Skris does not influence s as it used to. 921355714Skris 921455714Skris In order to clean up things more thoroughly, inside SSL_SESSION 921555714Skris we don't use CERT any longer, but a new structure SESS_CERT 921655714Skris that holds per-session data (if available); currently, this is 921755714Skris the peer's certificate chain and, for clients, the server's certificate 921855714Skris and temporary key. CERT holds only those values that can have 921955714Skris meaningful defaults in an SSL_CTX. 922055714Skris [Bodo Moeller] 922155714Skris 922255714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 922355714Skris from the internal representation. Various PKCS#7 fixes: remove some 922455714Skris evil casts and set the enc_dig_alg field properly based on the signing 922555714Skris key type. 922655714Skris [Steve Henson] 922755714Skris 922855714Skris *) Allow PKCS#12 password to be set from the command line or the 922955714Skris environment. Let 'ca' get its config file name from the environment 923055714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 923155714Skris and 'x509'). 923255714Skris [Steve Henson] 923355714Skris 923455714Skris *) Allow certificate policies extension to use an IA5STRING for the 923555714Skris organization field. This is contrary to the PKIX definition but 923655714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 923755714Skris extension option. 923855714Skris [Steve Henson] 923955714Skris 924055714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 924155714Skris without disallowing inline assembler and the like for non-pedantic builds. 924255714Skris [Ben Laurie] 924355714Skris 924455714Skris *) Support Borland C++ builder. 924555714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 924655714Skris 924755714Skris *) Support Mingw32. 924855714Skris [Ulf M�ller] 924955714Skris 925055714Skris *) SHA-1 cleanups and performance enhancements. 925155714Skris [Andy Polyakov <appro@fy.chalmers.se>] 925255714Skris 925355714Skris *) Sparc v8plus assembler for the bignum library. 925455714Skris [Andy Polyakov <appro@fy.chalmers.se>] 925555714Skris 925655714Skris *) Accept any -xxx and +xxx compiler options in Configure. 925755714Skris [Ulf M�ller] 925855714Skris 925955714Skris *) Update HPUX configuration. 926055714Skris [Anonymous] 926155714Skris 926255714Skris *) Add missing sk_<type>_unshift() function to safestack.h 926355714Skris [Ralf S. Engelschall] 926455714Skris 926555714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 926655714Skris "extra_cert"s in addition to the certificate. (This makes sense 926755714Skris only for "PEM" format files, as chains as a whole are not 926855714Skris DER-encoded.) 926955714Skris [Bodo Moeller] 927055714Skris 927155714Skris *) Support verify_depth from the SSL API. 927255714Skris x509_vfy.c had what can be considered an off-by-one-error: 927355714Skris Its depth (which was not part of the external interface) 927455714Skris was actually counting the number of certificates in a chain; 927555714Skris now it really counts the depth. 927655714Skris [Bodo Moeller] 927755714Skris 927855714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 927955714Skris instead of X509err, which often resulted in confusing error 928055714Skris messages since the error codes are not globally unique 928155714Skris (e.g. an alleged error in ssl3_accept when a certificate 928255714Skris didn't match the private key). 928355714Skris 928455714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 928555714Skris value (so that you don't need SSL_set_session_id_context for each 928655714Skris connection using the SSL_CTX). 928755714Skris [Bodo Moeller] 928855714Skris 928955714Skris *) OAEP decoding bug fix. 929055714Skris [Ulf M�ller] 929155714Skris 929255714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 929355714Skris David Harris. 929455714Skris [Bodo Moeller] 929555714Skris 929655714Skris *) New Configure options "threads" and "no-threads". For systems 929755714Skris where the proper compiler options are known (currently Solaris 929855714Skris and Linux), "threads" is the default. 929955714Skris [Bodo Moeller] 930055714Skris 930155714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 930255714Skris [Bodo Moeller] 930355714Skris 930455714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 930555714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 930655714Skris such as /usr/local/bin. 930755714Skris [Bodo Moeller] 930855714Skris 930955714Skris *) "make linux-shared" to build shared libraries. 931055714Skris [Niels Poppe <niels@netbox.org>] 931155714Skris 931255714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 931355714Skris [Ulf M�ller] 931455714Skris 931555714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 931655714Skris extension adding in x509 utility. 931755714Skris [Steve Henson] 931855714Skris 931955714Skris *) Remove NOPROTO sections and error code comments. 932055714Skris [Ulf M�ller] 932155714Skris 932255714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 932355714Skris prototypes. 932455714Skris [Steve Henson] 932555714Skris 932655714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 932755714Skris [Ulf M�ller] 932855714Skris 932955714Skris *) Complete rewrite of the error code script(s). It is all now handled 933055714Skris by one script at the top level which handles error code gathering, 933155714Skris header rewriting and C source file generation. It should be much better 933255714Skris than the old method: it now uses a modified version of Ulf's parser to 933355714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 933455714Skris aren't needed for error creation any more) and do a better job of 933555714Skris translating function codes into names. The old 'ASN1 error code imbedded 933655714Skris in a comment' is no longer necessary and it doesn't use .err files which 933755714Skris have now been deleted. Also the error code call doesn't have to appear all 933855714Skris on one line (which resulted in some large lines...). 933955714Skris [Steve Henson] 934055714Skris 934155714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 934255714Skris [Bodo Moeller] 934355714Skris 934455714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 934555714Skris 0 (which usually indicates a closed connection), but continue reading. 934655714Skris [Bodo Moeller] 934755714Skris 934855714Skris *) Fix some race conditions. 934955714Skris [Bodo Moeller] 935055714Skris 935155714Skris *) Add support for CRL distribution points extension. Add Certificate 935255714Skris Policies and CRL distribution points documentation. 935355714Skris [Steve Henson] 935455714Skris 935555714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 935655714Skris [Ulf M�ller] 935755714Skris 935855714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 935955714Skris 8 of keying material. Merlin has also confirmed interop with this fix 936055714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 936155714Skris [Merlin Hughes <merlin@baltimore.ie>] 936255714Skris 936355714Skris *) Fix lots of warnings. 936455714Skris [Richard Levitte <levitte@stacken.kth.se>] 936555714Skris 936655714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 936755714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 936855714Skris [Richard Levitte <levitte@stacken.kth.se>] 936955714Skris 937055714Skris *) Fix problems with sizeof(long) == 8. 937155714Skris [Andy Polyakov <appro@fy.chalmers.se>] 937255714Skris 937355714Skris *) Change functions to ANSI C. 937455714Skris [Ulf M�ller] 937555714Skris 937655714Skris *) Fix typos in error codes. 937755714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 937855714Skris 937955714Skris *) Remove defunct assembler files from Configure. 938055714Skris [Ulf M�ller] 938155714Skris 938255714Skris *) SPARC v8 assembler BIGNUM implementation. 938355714Skris [Andy Polyakov <appro@fy.chalmers.se>] 938455714Skris 938555714Skris *) Support for Certificate Policies extension: both print and set. 938655714Skris Various additions to support the r2i method this uses. 938755714Skris [Steve Henson] 938855714Skris 938955714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 939055714Skris return a const string when you are expecting an allocated buffer. 939155714Skris [Ben Laurie] 939255714Skris 939355714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 939455714Skris types DirectoryString and DisplayText. 939555714Skris [Steve Henson] 939655714Skris 939755714Skris *) Add code to allow r2i extensions to access the configuration database, 939855714Skris add an LHASH database driver and add several ctx helper functions. 939955714Skris [Steve Henson] 940055714Skris 940155714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 940255714Skris fail when they extended the size of a BIGNUM. 940355714Skris [Steve Henson] 940455714Skris 940555714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 940655714Skris support typesafe stack. 940755714Skris [Steve Henson] 940855714Skris 940955714Skris *) Fix typo in SSL_[gs]et_options(). 941055714Skris [Nils Frostberg <nils@medcom.se>] 941155714Skris 941255714Skris *) Delete various functions and files that belonged to the (now obsolete) 941355714Skris old X509V3 handling code. 941455714Skris [Steve Henson] 941555714Skris 941655714Skris *) New Configure option "rsaref". 941755714Skris [Ulf M�ller] 941855714Skris 941955714Skris *) Don't auto-generate pem.h. 942055714Skris [Bodo Moeller] 942155714Skris 942255714Skris *) Introduce type-safe ASN.1 SETs. 942355714Skris [Ben Laurie] 942455714Skris 942555714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 942655714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 942755714Skris 942855714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 942955714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 943055714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 943155714Skris few STACKed things have been converted already. Feel free to convert more. 943255714Skris In the fullness of time, I'll do away with the STACK type altogether. 943355714Skris [Ben Laurie] 943455714Skris 943555714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 943655714Skris specified in <certfile> by updating the entry in the index.txt file. 943755714Skris This way one no longer has to edit the index.txt file manually for 943855714Skris revoking a certificate. The -revoke option does the gory details now. 943955714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 944055714Skris 944155714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 944255714Skris `-text' option at all and this way the `-noout -text' combination was 944355714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 944455714Skris [Ralf S. Engelschall] 944555714Skris 944655714Skris *) Make sure a corresponding plain text error message exists for the 944755714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 944855714Skris verify callback function determined that a certificate was revoked. 944955714Skris [Ralf S. Engelschall] 945055714Skris 945155714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 945255714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 945355714Skris all available cipers including rc5, which was forgotten until now. 945455714Skris In order to let the testing shell script know which algorithms 945555714Skris are available, a new (up to now undocumented) command 945655714Skris "openssl list-cipher-commands" is used. 945755714Skris [Bodo Moeller] 945855714Skris 945955714Skris *) Bugfix: s_client occasionally would sleep in select() when 946055714Skris it should have checked SSL_pending() first. 946155714Skris [Bodo Moeller] 946255714Skris 946355714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 946455714Skris the raw DSA values prior to ASN.1 encoding. 946555714Skris [Ulf M�ller] 946655714Skris 946755714Skris *) Tweaks to Configure 946855714Skris [Niels Poppe <niels@netbox.org>] 946955714Skris 947055714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 947155714Skris yet... 947255714Skris [Steve Henson] 947355714Skris 947455714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 947555714Skris [Ulf M�ller] 947655714Skris 947755714Skris *) New config option to avoid instructions that are illegal on the 80386. 947855714Skris The default code is faster, but requires at least a 486. 947955714Skris [Ulf M�ller] 948055714Skris 948155714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 948255714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 948355714Skris same as SSL2_VERSION anyway. 948455714Skris [Bodo Moeller] 948555714Skris 948655714Skris *) New "-showcerts" option for s_client. 948755714Skris [Bodo Moeller] 948855714Skris 948955714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 949055714Skris application. Various cleanups and fixes. 949155714Skris [Steve Henson] 949255714Skris 949355714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 949455714Skris modify error routines to work internally. Add error codes and PBE init 949555714Skris to library startup routines. 949655714Skris [Steve Henson] 949755714Skris 949855714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 949955714Skris packing functions to asn1 and evp. Changed function names and error 950055714Skris codes along the way. 950155714Skris [Steve Henson] 950255714Skris 950355714Skris *) PKCS12 integration: and so it begins... First of several patches to 950455714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 950555714Skris objects to objects.h 950655714Skris [Steve Henson] 950755714Skris 950855714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 950955714Skris and display support for Thawte strong extranet extension. 951055714Skris [Steve Henson] 951155714Skris 951255714Skris *) Add LinuxPPC support. 951355714Skris [Jeff Dubrule <igor@pobox.org>] 951455714Skris 951555714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 951655714Skris bn_div_words in alpha.s. 951755714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 951855714Skris 951955714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 952055714Skris OAEP isn't supported when OpenSSL is built with RSAref. 952155714Skris [Ulf Moeller <ulf@fitug.de>] 952255714Skris 952355714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 952455714Skris so they no longer are missing under -DNOPROTO. 952555714Skris [Soren S. Jorvang <soren@t.dk>] 952655714Skris 952755714Skris 952855714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 952955714Skris 953055714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 953155714Skris doesn't work when the session is reused. Coming soon! 953255714Skris [Ben Laurie] 953355714Skris 953455714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 953555714Skris context thus bypassing client cert protection! All software that uses 953655714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 953755714Skris allow session reuse! A fuller solution is in the works. 953855714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 953955714Skris 954055714Skris *) Some more source tree cleanups (removed obsolete files 954155714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 954255714Skris permission on "config" script to be executable) and a fix for the INSTALL 954355714Skris document. 954455714Skris [Ulf Moeller <ulf@fitug.de>] 954555714Skris 954655714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 954755714Skris Malloc, Free. 954855714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 954955714Skris 955055714Skris *) Make rsa_oaep_test return non-zero on error. 955155714Skris [Ulf Moeller <ulf@fitug.de>] 955255714Skris 955355714Skris *) Add support for native Solaris shared libraries. Configure 955455714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 955555714Skris if someone would make that last step automatic. 955655714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 955755714Skris 955855714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 955955714Skris [Ben Laurie] 956055714Skris 956155714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 956255714Skris except NULL ciphers". This means the default cipher list will no longer 956355714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 956455714Skris the string "DEFAULT:eNULL". 956555714Skris [Steve Henson] 956655714Skris 956755714Skris *) Fix to RSA private encryption routines: if p < q then it would 956855714Skris occasionally produce an invalid result. This will only happen with 956955714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 957055714Skris [Steve Henson] 957155714Skris 957255714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 957355714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 957455714Skris because this way one can also use an interpreter named `perl5' (which is 957555714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 957655714Skris installed as `perl'). 957755714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 957855714Skris 957955714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 958055714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 958155714Skris 958255714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 958355714Skris advapi32.lib to Win32 build and change the pem test comparision 958455714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 958555714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 958655714Skris and crypto/des/ede_cbcm_enc.c. 958755714Skris [Steve Henson] 958855714Skris 958955714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 959055714Skris [Ben Laurie] 959155714Skris 959255714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 959355714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 959455714Skris is horrible: I feel ill.... 959555714Skris [Steve Henson] 959655714Skris 959755714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 959855714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 959955714Skris sections: 10 functions were absent from non ANSI section and not exported 960055714Skris from Windows DLLs. Fixed up libeay.num for new functions. 960155714Skris [Steve Henson] 960255714Skris 960355714Skris *) Make `openssl version' output lines consistent. 960455714Skris [Ralf S. Engelschall] 960555714Skris 960655714Skris *) Fix Win32 symbol export lists for BIO functions: Added 960755714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 960855714Skris to ms/libeay{16,32}.def. 960955714Skris [Ralf S. Engelschall] 961055714Skris 961155714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 961255714Skris fine under Unix and passes some trivial tests I've now added. But the 961355714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 961455714Skris added to make sure no one expects that this stuff really works in the 961555714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 961655714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 961755714Skris openssl_bio.xs. 961855714Skris [Ralf S. Engelschall] 961955714Skris 962055714Skris *) Fix the generation of two part addresses in perl. 962155714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 962255714Skris 962355714Skris *) Add config entry for Linux on MIPS. 962455714Skris [John Tobey <jtobey@channel1.com>] 962555714Skris 962655714Skris *) Make links whenever Configure is run, unless we are on Windoze. 962755714Skris [Ben Laurie] 962855714Skris 962955714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 963055714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 963155714Skris in CRLs. 963255714Skris [Steve Henson] 963355714Skris 963455714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 963555714Skris other platforms details on the command line without having to patch the 963655714Skris Configure script everytime: One now can use ``perl Configure 963755714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 963855714Skris to them (seperated by colons). This is treated as there would be a static 963955714Skris pre-configured entry in Configure's %table under key <id> with value 964055714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 964155714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 964255714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 964355714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 964455714Skris [Ralf S. Engelschall] 964555714Skris 964655714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 964755714Skris [Ben Laurie] 964855714Skris 964955714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 965055714Skris on the `perl Configure ...' command line. This way one can compile 965155714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 965255714Skris for linking it into DSOs. 965355714Skris [Ralf S. Engelschall] 965455714Skris 965555714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 965655714Skris Fixed. 965755714Skris [Ben Laurie] 965855714Skris 965955714Skris *) Cleaned up the LICENSE document: The official contact for any license 966055714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 966155714Skris And add a paragraph about the dual-license situation to make sure people 966255714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 966355714Skris to the OpenSSL toolkit. 966455714Skris [Ralf S. Engelschall] 966555714Skris 966655714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 966755714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 966855714Skris Additonally cleaned up the `make links' target: Remove unnecessary 966955714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 967055714Skris to speed processing and no longer clutter the display with confusing 967155714Skris stuff. Instead only the actually done links are displayed. 967255714Skris [Ralf S. Engelschall] 967355714Skris 967455714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 967555714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 967655714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 967755714Skris encryption. 967855714Skris [Ben Laurie] 967955714Skris 968055714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 968155714Skris signed attributes when verifying signatures (this would break them), 968255714Skris the detached data encoding was wrong and public keys obtained using 968355714Skris X509_get_pubkey() weren't freed. 968455714Skris [Steve Henson] 968555714Skris 968655714Skris *) Add text documentation for the BUFFER functions. Also added a work around 968755714Skris to a Win95 console bug. This was triggered by the password read stuff: the 968855714Skris last character typed gets carried over to the next fread(). If you were 968955714Skris generating a new cert request using 'req' for example then the last 969055714Skris character of the passphrase would be CR which would then enter the first 969155714Skris field as blank. 969255714Skris [Steve Henson] 969355714Skris 969455714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 969555714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 969655714Skris button and can be used by applications based on OpenSSL to show the 969755714Skris relationship to the OpenSSL project. 969855714Skris [Ralf S. Engelschall] 969955714Skris 970055714Skris *) Remove confusing variables in function signatures in files 970155714Skris ssl/ssl_lib.c and ssl/ssl.h. 970255714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 970355714Skris 970455714Skris *) Don't install bss_file.c under PREFIX/include/ 970555714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 970655714Skris 970755714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 970855714Skris functions that return function pointers and has support for NT specific 970955714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 971055714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 971155714Skris unsigned to signed types: this was killing the Win32 compile. 971255714Skris [Steve Henson] 971355714Skris 971455714Skris *) Add new certificate file to stack functions, 971555714Skris SSL_add_dir_cert_subjects_to_stack() and 971655714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 971755714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 971855714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 971955714Skris This means that Apache-SSL and similar packages don't have to mess around 972055714Skris to add as many CAs as they want to the preferred list. 972155714Skris [Ben Laurie] 972255714Skris 972355714Skris *) Experiment with doxygen documentation. Currently only partially applied to 972455714Skris ssl/ssl_lib.c. 972555714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 972655714Skris openssl.doxy as the configuration file. 972755714Skris [Ben Laurie] 972855714Skris 972955714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 973055714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 973155714Skris 973255714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 973355714Skris compiled in by default: it has problems with large keys. 973455714Skris [Steve Henson] 973555714Skris 973655714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 973755714Skris DH private keys and/or callback functions which directly correspond to 973855714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 973955714Skris is needed for applications which have to configure certificates on a 974055714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 974155714Skris (e.g. s_server). 974255714Skris For the RSA certificate situation is makes no difference, but 974355714Skris for the DSA certificate situation this fixes the "no shared cipher" 974455714Skris problem where the OpenSSL cipher selection procedure failed because the 974555714Skris temporary keys were not overtaken from the context and the API provided 974655714Skris no way to reconfigure them. 974755714Skris The new functions now let applications reconfigure the stuff and they 974855714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 974955714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 975055714Skris non-public-API function ssl_cert_instantiate() is used as a helper 975155714Skris function and also to reduce code redundancy inside ssl_rsa.c. 975255714Skris [Ralf S. Engelschall] 975355714Skris 975455714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 975555714Skris area because they are useful for the DSA situation and should be 975655714Skris recognized by the users. 975755714Skris [Ralf S. Engelschall] 975855714Skris 975955714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 976055714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 976155714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 976255714Skris already masked variable. 976355714Skris [Richard Levitte <levitte@stacken.kth.se>] 976455714Skris 976555714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 976655714Skris [Richard Levitte <levitte@stacken.kth.se>] 976755714Skris 976855714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 976955714Skris from `int' to `unsigned int' because it's a length and initialized by 977055714Skris EVP_DigestFinal() which expects an `unsigned int *'. 977155714Skris [Richard Levitte <levitte@stacken.kth.se>] 977255714Skris 977355714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 977455714Skris script. Instead use the usual Shell->Perl transition trick. 977555714Skris [Ralf S. Engelschall] 977655714Skris 977755714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 977855714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 977955714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 978055714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 978155714Skris currently the public key is printed (a decision which was already done by 978255714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 978355714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 978455714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 978555714Skris now, too. 978655714Skris [Ralf S. Engelschall] 978755714Skris 978855714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 978955714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 979055714Skris [Arne Ansper <arne@ats.cyber.ee>] 979155714Skris 979255714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 979355714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 979455714Skris config file. 979555714Skris [Steve Henson] 979655714Skris 979755714Skris *) Add cool BIO that does syslog (or event log on NT). 979855714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 979955714Skris 980055714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 980155714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 980255714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 980355714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 980455714Skris [Ben Laurie] 980555714Skris 980655714Skris *) Add preliminary config info for new extension code. 980755714Skris [Steve Henson] 980855714Skris 980955714Skris *) Make RSA_NO_PADDING really use no padding. 981055714Skris [Ulf Moeller <ulf@fitug.de>] 981155714Skris 981255714Skris *) Generate errors when private/public key check is done. 981355714Skris [Ben Laurie] 981455714Skris 981555714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 981655714Skris for some CRL extensions and new objects added. 981755714Skris [Steve Henson] 981855714Skris 981955714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 982055714Skris key usage extension and fuller support for authority key id. 982155714Skris [Steve Henson] 982255714Skris 982355714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 982455714Skris padding method for RSA, which is recommended for new applications in PKCS 982555714Skris #1 v2.0 (RFC 2437, October 1998). 982655714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 982755714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 982855714Skris against Bleichbacher's attack on RSA. 982955714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 983055714Skris Ben Laurie] 983155714Skris 983255714Skris *) Updates to the new SSL compression code 983355714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 983455714Skris 983555714Skris *) Fix so that the version number in the master secret, when passed 983655714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 983755714Skris (because the server will not accept higher), that the version number 983855714Skris is 0x03,0x01, not 0x03,0x00 983955714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 984055714Skris 984155714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 984255714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 984355714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 984455714Skris [Steve Henson] 984555714Skris 984655714Skris *) Support for RAW extensions where an arbitrary extension can be 984755714Skris created by including its DER encoding. See apps/openssl.cnf for 984855714Skris an example. 984955714Skris [Steve Henson] 985055714Skris 985155714Skris *) Make sure latest Perl versions don't interpret some generated C array 985255714Skris code as Perl array code in the crypto/err/err_genc.pl script. 985355714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 985455714Skris 985555714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 985655714Skris not many people have the assembler. Various Win32 compilation fixes and 985755714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 985855714Skris build instructions. 985955714Skris [Steve Henson] 986055714Skris 986155714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 986255714Skris file under Win32 and also build pem.h from pem.org. New script 986355714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 986455714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 986555714Skris [Steve Henson] 986655714Skris 986755714Skris *) Major rework of DES function declarations, in the pursuit of correctness 986855714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 986955714Skris too. You may find this causes warnings in your code. Zapping your evil 987055714Skris casts will probably fix them. Mostly. 987155714Skris [Ben Laurie] 987255714Skris 987355714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 987455714Skris obj_dat.pl. It considered a zero in an object definition to mean 987555714Skris "end of object": none of the objects in objects.h have any zeros 987655714Skris so it wasn't spotted. 987755714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 987855714Skris 987955714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 988055714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 988155714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 988255714Skris vectors if you have them. 988355714Skris [Ben Laurie] 988455714Skris 988555714Skris *) Correct calculation of key length for export ciphers (too much space was 988655714Skris allocated for null ciphers). This has not been tested! 988755714Skris [Ben Laurie] 988855714Skris 988955714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 989055714Skris message is now correct (it understands "crypto" and "ssl" on its 989155714Skris command line). There is also now an "update" option. This will update 989255714Skris the util/ssleay.num and util/libeay.num files with any new functions. 989355714Skris If you do a: 989455714Skris perl util/mkdef.pl crypto ssl update 989555714Skris it will update them. 989655714Skris [Steve Henson] 989755714Skris 989855714Skris *) Overhauled the Perl interface (perl/*): 989955714Skris - ported BN stuff to OpenSSL's different BN library 990055714Skris - made the perl/ source tree CVS-aware 990155714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 990255714Skris their history because I've copied them in the repository) 990355714Skris - removed obsolete files (the test scripts will be replaced 990455714Skris by better Test::Harness variants in the future) 990555714Skris [Ralf S. Engelschall] 990655714Skris 990755714Skris *) First cut for a very conservative source tree cleanup: 990855714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 990955714Skris where we collect the old documents and readme texts. 991055714Skris 2. remove the first part of files where I'm already sure that we no 991155714Skris longer need them because of three reasons: either they are just temporary 991255714Skris files which were left by Eric or they are preserved original files where 991355714Skris I've verified that the diff is also available in the CVS via "cvs diff 991455714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 991555714Skris the crypto/md/ stuff). 991655714Skris [Ralf S. Engelschall] 991755714Skris 991855714Skris *) More extension code. Incomplete support for subject and issuer alt 991955714Skris name, issuer and authority key id. Change the i2v function parameters 992055714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 992155714Skris what that's for :-) Fix to ASN1 macro which messed up 992255714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 992355714Skris [Steve Henson] 992455714Skris 992555714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 992655714Skris INTEGER code. 992755714Skris [Steve Henson] 992855714Skris 992955714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 993055714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 993155714Skris 993255714Skris *) Make sure `make rehash' target really finds the `openssl' program. 993355714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 993455714Skris 993555714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 993655714Skris like to hear about it if this slows down other processors. 993755714Skris [Ben Laurie] 993855714Skris 993955714Skris *) Add CygWin32 platform information to Configure script. 994055714Skris [Alan Batie <batie@aahz.jf.intel.com>] 994155714Skris 994255714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 994355714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 994455714Skris 994555714Skris *) New program nseq to manipulate netscape certificate sequences 994655714Skris [Steve Henson] 994755714Skris 994855714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 994955714Skris few typos. 995055714Skris [Steve Henson] 995155714Skris 995255714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 995355714Skris but the BN code had some problems that would cause failures when 995455714Skris doing certificate verification and some other functions. 995555714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 995655714Skris 995755714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 995855714Skris [Steve Henson] 995955714Skris 996055714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 996155714Skris [Steve Henson] 996255714Skris 996355714Skris *) Add several PKIX and private extended key usage OIDs. 996455714Skris [Steve Henson] 996555714Skris 996655714Skris *) Modify the 'ca' program to handle the new extension code. Modify 996755714Skris openssl.cnf for new extension format, add comments. 996855714Skris [Steve Henson] 996955714Skris 997055714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 997155714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 997255714Skris CA extensions. 997355714Skris [Steve Henson] 997455714Skris 997555714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 997655714Skris error code, add initial support to X509_print() and x509 application. 997755714Skris [Steve Henson] 997855714Skris 997955714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 998055714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 998155714Skris stuff is currently isolated and isn't even compiled yet. 998255714Skris [Steve Henson] 998355714Skris 998455714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 998555714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 998655714Skris Removed the versions check from X509 routines when loading extensions: 998755714Skris this allows certain broken certificates that don't set the version 998855714Skris properly to be processed. 998955714Skris [Steve Henson] 999055714Skris 999155714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 999255714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 999355714Skris can still be regenerated with "make depend". 999455714Skris [Ben Laurie] 999555714Skris 999655714Skris *) Spelling mistake in C version of CAST-128. 999755714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 999855714Skris 999955714Skris *) Changes to the error generation code. The perl script err-code.pl 1000055714Skris now reads in the old error codes and retains the old numbers, only 1000155714Skris adding new ones if necessary. It also only changes the .err files if new 1000255714Skris codes are added. The makefiles have been modified to only insert errors 1000355714Skris when needed (to avoid needlessly modifying header files). This is done 1000455714Skris by only inserting errors if the .err file is newer than the auto generated 1000555714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 1000655714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 1000755714Skris or delete all the .err files. 1000855714Skris [Steve Henson] 1000955714Skris 1001055714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 1001155714Skris been fixed, but is untested. The assembler versions are also fixed, but 1001255714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 1001355714Skris to regenerate it if needed. 1001455714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 1001555714Skris Hagino <itojun@kame.net>] 1001655714Skris 1001755714Skris *) File was opened incorrectly in randfile.c. 1001855714Skris [Ulf M�ller <ulf@fitug.de>] 1001955714Skris 1002055714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 1002155714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 1002255714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 1002355714Skris al: it's just almost always a UTCTime. Note this patch adds new error 1002455714Skris codes so do a "make errors" if there are problems. 1002555714Skris [Steve Henson] 1002655714Skris 1002755714Skris *) Correct Linux 1 recognition in config. 1002855714Skris [Ulf M�ller <ulf@fitug.de>] 1002955714Skris 1003055714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 1003155714Skris [Anonymous <nobody@replay.com>] 1003255714Skris 1003355714Skris *) Generate an error if given an empty string as a cert directory. Also 1003455714Skris generate an error if handed NULL (previously returned 0 to indicate an 1003555714Skris error, but didn't set one). 1003655714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 1003755714Skris 1003855714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 1003955714Skris [Ben Laurie] 1004055714Skris 1004155714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 1004255714Skris parameters. This was causing a warning which killed off the Win32 compile. 1004355714Skris [Steve Henson] 1004455714Skris 1004555714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 1004655714Skris [Neil Costigan <neil.costigan@celocom.com>] 1004755714Skris 1004855714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 1004955714Skris based on a text string, looking up short and long names and finally 1005055714Skris "dot" format. The "dot" format stuff didn't work. Added new function 1005155714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 1005255714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 1005355714Skris OID is not part of the table. 1005455714Skris [Steve Henson] 1005555714Skris 1005655714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 1005755714Skris X509_LOOKUP_by_alias(). 1005855714Skris [Ben Laurie] 1005955714Skris 1006055714Skris *) Sort openssl functions by name. 1006155714Skris [Ben Laurie] 1006255714Skris 1006355714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 1006455714Skris encryption from sample DSA keys (in case anyone is interested the password 1006555714Skris was "1234"). 1006655714Skris [Steve Henson] 1006755714Skris 1006855714Skris *) Make _all_ *_free functions accept a NULL pointer. 1006955714Skris [Frans Heymans <fheymans@isaserver.be>] 1007055714Skris 1007155714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 1007255714Skris NULL pointers. 1007355714Skris [Anonymous <nobody@replay.com>] 1007455714Skris 1007555714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 1007655714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 1007755714Skris 1007855714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 1007955714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 1008055714Skris 1008155714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 1008255714Skris [Anonymous <nobody@replay.com>] 1008355714Skris 1008455714Skris *) Add prototype for temp key callback functions 1008555714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 1008655714Skris [Ben Laurie] 1008755714Skris 1008855714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 1008955714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 1009055714Skris [Steve Henson] 1009155714Skris 1009255714Skris *) X509_name_add_entry() freed the wrong thing after an error. 1009355714Skris [Arne Ansper <arne@ats.cyber.ee>] 1009455714Skris 1009555714Skris *) rsa_eay.c would attempt to free a NULL context. 1009655714Skris [Arne Ansper <arne@ats.cyber.ee>] 1009755714Skris 1009855714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 1009955714Skris [Arne Ansper <arne@ats.cyber.ee>] 1010055714Skris 1010155714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 1010255714Skris [Arne Ansper <arne@ats.cyber.ee>] 1010355714Skris 1010455714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 1010555714Skris in X509_STORE_new(), but document the fact that this variable is still 1010655714Skris unused in the certificate verification process. 1010755714Skris [Ralf S. Engelschall] 1010855714Skris 1010955714Skris *) Fix the various library and apps files to free up pkeys obtained from 1011055714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 1011155714Skris [Steve Henson] 1011255714Skris 1011355714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 1011455714Skris demos/maurice/example2.c work, amongst others, probably. 1011555714Skris [Steve Henson and Ben Laurie] 1011655714Skris 1011755714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 1011855714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 1011955714Skris are no longer created. This way we have a single and consistent command 1012055714Skris line interface `openssl <command>', similar to `cvs <command>'. 1012155714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 1012255714Skris 1012355714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 1012455714Skris BIT STRING wrapper always have zero unused bits. 1012555714Skris [Steve Henson] 1012655714Skris 1012755714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 1012855714Skris [Steve Henson] 1012955714Skris 1013055714Skris *) Make the top-level INSTALL documentation easier to understand. 1013155714Skris [Paul Sutton] 1013255714Skris 1013355714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 1013455714Skris make (including if user presses ^C) [Paul Sutton] 1013555714Skris 1013655714Skris *) Make Montgomery context stuff explicit in RSA data structure. 1013755714Skris [Ben Laurie] 1013855714Skris 1013955714Skris *) Fix build order of pem and err to allow for generated pem.h. 1014055714Skris [Ben Laurie] 1014155714Skris 1014255714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 1014355714Skris [Ben Laurie] 1014455714Skris 1014555714Skris *) Enhanced the err-ins.pl script so it makes the error library number 1014655714Skris global and can add a library name. This is needed for external ASN1 and 1014755714Skris other error libraries. 1014855714Skris [Steve Henson] 1014955714Skris 1015055714Skris *) Fixed sk_insert which never worked properly. 1015155714Skris [Steve Henson] 1015255714Skris 1015355714Skris *) Fix ASN1 macros so they can handle indefinite length construted 1015455714Skris EXPLICIT tags. Some non standard certificates use these: they can now 1015555714Skris be read in. 1015655714Skris [Steve Henson] 1015755714Skris 1015855714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 1015955714Skris into a single doc/ssleay.txt bundle. This way the information is still 1016055714Skris preserved but no longer messes up this directory. Now it's new room for 1016155714Skris the new set of documenation files. 1016255714Skris [Ralf S. Engelschall] 1016355714Skris 1016455714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 1016555714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 1016655714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 1016755714Skris number of arguments. 1016855714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 1016955714Skris 1017055714Skris *) Fix test data to work with the above. 1017155714Skris [Ben Laurie] 1017255714Skris 1017355714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 1017455714Skris was already fixed by Eric for 0.9.1 it seems. 1017555714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 1017655714Skris 1017755714Skris *) Autodetect FreeBSD3. 1017855714Skris [Ben Laurie] 1017955714Skris 1018055714Skris *) Fix various bugs in Configure. This affects the following platforms: 1018155714Skris nextstep 1018255714Skris ncr-scde 1018355714Skris unixware-2.0 1018455714Skris unixware-2.0-pentium 1018555714Skris sco5-cc. 1018655714Skris [Ben Laurie] 1018755714Skris 1018855714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 1018955714Skris before they are needed. 1019055714Skris [Ben Laurie] 1019155714Skris 1019255714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 1019355714Skris [Ben Laurie] 1019455714Skris 1019555714Skris 1019655714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 1019755714Skris 1019855714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 1019955714Skris changed SSLeay to OpenSSL in version strings. 1020055714Skris [Ralf S. Engelschall] 1020155714Skris 1020255714Skris *) Some fixups to the top-level documents. 1020355714Skris [Paul Sutton] 1020455714Skris 1020555714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 1020655714Skris because the symlink to include/ was missing. 1020755714Skris [Ralf S. Engelschall] 1020855714Skris 1020955714Skris *) Incorporated the popular no-RSA/DSA-only patches 1021055714Skris which allow to compile a RSA-free SSLeay. 1021155714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 1021255714Skris 1021355714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 1021455714Skris when "ssleay" is still not found. 1021555714Skris [Ralf S. Engelschall] 1021655714Skris 1021755714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 1021855714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 1021955714Skris 1022055714Skris *) Updated the README file. 1022155714Skris [Ralf S. Engelschall] 1022255714Skris 1022355714Skris *) Added various .cvsignore files in the CVS repository subdirs 1022455714Skris to make a "cvs update" really silent. 1022555714Skris [Ralf S. Engelschall] 1022655714Skris 1022755714Skris *) Recompiled the error-definition header files and added 1022855714Skris missing symbols to the Win32 linker tables. 1022955714Skris [Ralf S. Engelschall] 1023055714Skris 1023155714Skris *) Cleaned up the top-level documents; 1023255714Skris o new files: CHANGES and LICENSE 1023355714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 1023455714Skris o merged COPYRIGHT into LICENSE 1023555714Skris o removed obsolete TODO file 1023655714Skris o renamed MICROSOFT to INSTALL.W32 1023755714Skris [Ralf S. Engelschall] 1023855714Skris 1023955714Skris *) Removed dummy files from the 0.9.1b source tree: 1024055714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 1024155714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 1024255714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 1024355714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 1024455714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 1024555714Skris [Ralf S. Engelschall] 1024655714Skris 1024755714Skris *) Added various platform portability fixes. 1024855714Skris [Mark J. Cox] 1024955714Skris 1025055714Skris *) The Genesis of the OpenSSL rpject: 1025155714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 1025255714Skris Young and Tim J. Hudson created while they were working for C2Net until 1025355714Skris summer 1998. 1025455714Skris [The OpenSSL Project] 1025555714Skris 1025655714Skris 1025755714Skris Changes between 0.9.0b and 0.9.1b [not released] 1025855714Skris 1025955714Skris *) Updated a few CA certificates under certs/ 1026055714Skris [Eric A. Young] 1026155714Skris 1026255714Skris *) Changed some BIGNUM api stuff. 1026355714Skris [Eric A. Young] 1026455714Skris 1026555714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 1026655714Skris DGUX x86, Linux Alpha, etc. 1026755714Skris [Eric A. Young] 1026855714Skris 1026955714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 1027055714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 1027155714Skris available). 1027255714Skris [Eric A. Young] 1027355714Skris 1027455714Skris *) Add -strparse option to asn1pars program which parses nested 1027555714Skris binary structures 1027655714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 1027755714Skris 1027855714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 1027955714Skris [Eric A. Young] 1028055714Skris 1028155714Skris *) DSA fix for "ca" program. 1028255714Skris [Eric A. Young] 1028355714Skris 1028455714Skris *) Added "-genkey" option to "dsaparam" program. 1028555714Skris [Eric A. Young] 1028655714Skris 1028755714Skris *) Added RIPE MD160 (rmd160) message digest. 1028855714Skris [Eric A. Young] 1028955714Skris 1029055714Skris *) Added -a (all) option to "ssleay version" command. 1029155714Skris [Eric A. Young] 1029255714Skris 1029355714Skris *) Added PLATFORM define which is the id given to Configure. 1029455714Skris [Eric A. Young] 1029555714Skris 1029655714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 1029755714Skris [Eric A. Young] 1029855714Skris 1029955714Skris *) Extended the ASN.1 parser routines. 1030055714Skris [Eric A. Young] 1030155714Skris 1030255714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 1030355714Skris [Eric A. Young] 1030455714Skris 1030555714Skris *) Added a BN_CTX to the BN library. 1030655714Skris [Eric A. Young] 1030755714Skris 1030855714Skris *) Fixed the weak key values in DES library 1030955714Skris [Eric A. Young] 1031055714Skris 1031155714Skris *) Changed API in EVP library for cipher aliases. 1031255714Skris [Eric A. Young] 1031355714Skris 1031455714Skris *) Added support for RC2/64bit cipher. 1031555714Skris [Eric A. Young] 1031655714Skris 1031755714Skris *) Converted the lhash library to the crypto/mem.c functions. 1031855714Skris [Eric A. Young] 1031955714Skris 1032055714Skris *) Added more recognized ASN.1 object ids. 1032155714Skris [Eric A. Young] 1032255714Skris 1032355714Skris *) Added more RSA padding checks for SSL/TLS. 1032455714Skris [Eric A. Young] 1032555714Skris 1032655714Skris *) Added BIO proxy/filter functionality. 1032755714Skris [Eric A. Young] 1032855714Skris 1032955714Skris *) Added extra_certs to SSL_CTX which can be used 1033055714Skris send extra CA certificates to the client in the CA cert chain sending 1033155714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 1033255714Skris [Eric A. Young] 1033355714Skris 1033455714Skris *) Now Fortezza is denied in the authentication phase because 1033555714Skris this is key exchange mechanism is not supported by SSLeay at all. 1033655714Skris [Eric A. Young] 1033755714Skris 1033855714Skris *) Additional PKCS1 checks. 1033955714Skris [Eric A. Young] 1034055714Skris 1034155714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 1034255714Skris [Eric A. Young] 1034355714Skris 1034455714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 1034555714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 1034655714Skris [Eric A. Young] 1034755714Skris 1034855714Skris *) Fixed a few memory leaks. 1034955714Skris [Eric A. Young] 1035055714Skris 1035155714Skris *) Fixed various code and comment typos. 1035255714Skris [Eric A. Young] 1035355714Skris 1035455714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 1035555714Skris bytes sent in the client random. 1035655714Skris [Edward Bishop <ebishop@spyglass.com>] 1035755714Skris 10358