CHANGES revision 273149
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 5273149Sjkim Changes between 1.0.1i and 1.0.1j [15 Oct 2014] 6273149Sjkim 7273149Sjkim *) SRTP Memory Leak. 8273149Sjkim 9273149Sjkim A flaw in the DTLS SRTP extension parsing code allows an attacker, who 10273149Sjkim sends a carefully crafted handshake message, to cause OpenSSL to fail 11273149Sjkim to free up to 64k of memory causing a memory leak. This could be 12273149Sjkim exploited in a Denial Of Service attack. This issue affects OpenSSL 13273149Sjkim 1.0.1 server implementations for both SSL/TLS and DTLS regardless of 14273149Sjkim whether SRTP is used or configured. Implementations of OpenSSL that 15273149Sjkim have been compiled with OPENSSL_NO_SRTP defined are not affected. 16273149Sjkim 17273149Sjkim The fix was developed by the OpenSSL team. 18273149Sjkim (CVE-2014-3513) 19273149Sjkim [OpenSSL team] 20273149Sjkim 21273149Sjkim *) Session Ticket Memory Leak. 22273149Sjkim 23273149Sjkim When an OpenSSL SSL/TLS/DTLS server receives a session ticket the 24273149Sjkim integrity of that ticket is first verified. In the event of a session 25273149Sjkim ticket integrity check failing, OpenSSL will fail to free memory 26273149Sjkim causing a memory leak. By sending a large number of invalid session 27273149Sjkim tickets an attacker could exploit this issue in a Denial Of Service 28273149Sjkim attack. 29273149Sjkim (CVE-2014-3567) 30273149Sjkim [Steve Henson] 31273149Sjkim 32273149Sjkim *) Build option no-ssl3 is incomplete. 33273149Sjkim 34273149Sjkim When OpenSSL is configured with "no-ssl3" as a build option, servers 35273149Sjkim could accept and complete a SSL 3.0 handshake, and clients could be 36273149Sjkim configured to send them. 37273149Sjkim (CVE-2014-3568) 38273149Sjkim [Akamai and the OpenSSL team] 39273149Sjkim 40273149Sjkim *) Add support for TLS_FALLBACK_SCSV. 41273149Sjkim Client applications doing fallback retries should call 42273149Sjkim SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV). 43273149Sjkim (CVE-2014-3566) 44273149Sjkim [Adam Langley, Bodo Moeller] 45273149Sjkim 46273149Sjkim *) Add additional DigestInfo checks. 47273149Sjkim 48273149Sjkim Reencode DigestInto in DER and check against the original when 49273149Sjkim verifying RSA signature: this will reject any improperly encoded 50273149Sjkim DigestInfo structures. 51273149Sjkim 52273149Sjkim Note: this is a precautionary measure and no attacks are currently known. 53273149Sjkim 54273149Sjkim [Steve Henson] 55273149Sjkim 56269686Sjkim Changes between 1.0.1h and 1.0.1i [6 Aug 2014] 57269686Sjkim 58269686Sjkim *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the 59269686Sjkim SRP code can be overrun an internal buffer. Add sanity check that 60269686Sjkim g, A, B < N to SRP code. 61269686Sjkim 62269686Sjkim Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC 63269686Sjkim Group for discovering this issue. 64269686Sjkim (CVE-2014-3512) 65269686Sjkim [Steve Henson] 66269686Sjkim 67269686Sjkim *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate 68269686Sjkim TLS 1.0 instead of higher protocol versions when the ClientHello message 69269686Sjkim is badly fragmented. This allows a man-in-the-middle attacker to force a 70269686Sjkim downgrade to TLS 1.0 even if both the server and the client support a 71269686Sjkim higher protocol version, by modifying the client's TLS records. 72269686Sjkim 73269686Sjkim Thanks to David Benjamin and Adam Langley (Google) for discovering and 74269686Sjkim researching this issue. 75269686Sjkim (CVE-2014-3511) 76269686Sjkim [David Benjamin] 77269686Sjkim 78269686Sjkim *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject 79269686Sjkim to a denial of service attack. A malicious server can crash the client 80269686Sjkim with a null pointer dereference (read) by specifying an anonymous (EC)DH 81269686Sjkim ciphersuite and sending carefully crafted handshake messages. 82269686Sjkim 83269686Sjkim Thanks to Felix Gr�bert (Google) for discovering and researching this 84269686Sjkim issue. 85269686Sjkim (CVE-2014-3510) 86269686Sjkim [Emilia K�sper] 87269686Sjkim 88269686Sjkim *) By sending carefully crafted DTLS packets an attacker could cause openssl 89269686Sjkim to leak memory. This can be exploited through a Denial of Service attack. 90269686Sjkim Thanks to Adam Langley for discovering and researching this issue. 91269686Sjkim (CVE-2014-3507) 92269686Sjkim [Adam Langley] 93269686Sjkim 94269686Sjkim *) An attacker can force openssl to consume large amounts of memory whilst 95269686Sjkim processing DTLS handshake messages. This can be exploited through a 96269686Sjkim Denial of Service attack. 97269686Sjkim Thanks to Adam Langley for discovering and researching this issue. 98269686Sjkim (CVE-2014-3506) 99269686Sjkim [Adam Langley] 100269686Sjkim 101269686Sjkim *) An attacker can force an error condition which causes openssl to crash 102269686Sjkim whilst processing DTLS packets due to memory being freed twice. This 103269686Sjkim can be exploited through a Denial of Service attack. 104269686Sjkim Thanks to Adam Langley and Wan-Teh Chang for discovering and researching 105269686Sjkim this issue. 106269686Sjkim (CVE-2014-3505) 107269686Sjkim [Adam Langley] 108269686Sjkim 109269686Sjkim *) If a multithreaded client connects to a malicious server using a resumed 110269686Sjkim session and the server sends an ec point format extension it could write 111269686Sjkim up to 255 bytes to freed memory. 112269686Sjkim 113269686Sjkim Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this 114269686Sjkim issue. 115269686Sjkim (CVE-2014-3509) 116269686Sjkim [Gabor Tyukasz] 117269686Sjkim 118269686Sjkim *) A malicious server can crash an OpenSSL client with a null pointer 119269686Sjkim dereference (read) by specifying an SRP ciphersuite even though it was not 120269686Sjkim properly negotiated with the client. This can be exploited through a 121269686Sjkim Denial of Service attack. 122269686Sjkim 123269686Sjkim Thanks to Joonas Kuorilehto and Riku Hietam�ki (Codenomicon) for 124269686Sjkim discovering and researching this issue. 125269686Sjkim (CVE-2014-5139) 126269686Sjkim [Steve Henson] 127269686Sjkim 128269686Sjkim *) A flaw in OBJ_obj2txt may cause pretty printing functions such as 129269686Sjkim X509_name_oneline, X509_name_print_ex et al. to leak some information 130269686Sjkim from the stack. Applications may be affected if they echo pretty printing 131269686Sjkim output to the attacker. 132269686Sjkim 133269686Sjkim Thanks to Ivan Fratric (Google) for discovering this issue. 134269686Sjkim (CVE-2014-3508) 135269686Sjkim [Emilia K�sper, and Steve Henson] 136269686Sjkim 137269686Sjkim *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) 138269686Sjkim for corner cases. (Certain input points at infinity could lead to 139269686Sjkim bogus results, with non-infinity inputs mapped to infinity too.) 140269686Sjkim [Bodo Moeller] 141269686Sjkim 142267258Sjkim Changes between 1.0.1g and 1.0.1h [5 Jun 2014] 143267258Sjkim 144267258Sjkim *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted 145267258Sjkim handshake can force the use of weak keying material in OpenSSL 146267258Sjkim SSL/TLS clients and servers. 147267258Sjkim 148267258Sjkim Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and 149267258Sjkim researching this issue. (CVE-2014-0224) 150267258Sjkim [KIKUCHI Masashi, Steve Henson] 151267258Sjkim 152267258Sjkim *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an 153267258Sjkim OpenSSL DTLS client the code can be made to recurse eventually crashing 154267258Sjkim in a DoS attack. 155267258Sjkim 156267258Sjkim Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue. 157267258Sjkim (CVE-2014-0221) 158267258Sjkim [Imre Rad, Steve Henson] 159267258Sjkim 160267258Sjkim *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can 161267258Sjkim be triggered by sending invalid DTLS fragments to an OpenSSL DTLS 162267258Sjkim client or server. This is potentially exploitable to run arbitrary 163267258Sjkim code on a vulnerable client or server. 164267258Sjkim 165267258Sjkim Thanks to J�ri Aedla for reporting this issue. (CVE-2014-0195) 166267258Sjkim [J�ri Aedla, Steve Henson] 167267258Sjkim 168267258Sjkim *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites 169267258Sjkim are subject to a denial of service attack. 170267258Sjkim 171267258Sjkim Thanks to Felix Gr�bert and Ivan Fratric at Google for discovering 172267258Sjkim this issue. (CVE-2014-3470) 173267258Sjkim [Felix Gr�bert, Ivan Fratric, Steve Henson] 174267258Sjkim 175267258Sjkim *) Harmonize version and its documentation. -f flag is used to display 176267258Sjkim compilation flags. 177267258Sjkim [mancha <mancha1@zoho.com>] 178267258Sjkim 179267258Sjkim *) Fix eckey_priv_encode so it immediately returns an error upon a failure 180267258Sjkim in i2d_ECPrivateKey. 181267258Sjkim [mancha <mancha1@zoho.com>] 182267258Sjkim 183267258Sjkim *) Fix some double frees. These are not thought to be exploitable. 184267258Sjkim [mancha <mancha1@zoho.com>] 185267258Sjkim 186264331Sjkim Changes between 1.0.1f and 1.0.1g [7 Apr 2014] 187264331Sjkim 188264331Sjkim *) A missing bounds check in the handling of the TLS heartbeat extension 189264331Sjkim can be used to reveal up to 64k of memory to a connected client or 190264331Sjkim server. 191264331Sjkim 192264331Sjkim Thanks for Neel Mehta of Google Security for discovering this bug and to 193264331Sjkim Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for 194264331Sjkim preparing the fix (CVE-2014-0160) 195264331Sjkim [Adam Langley, Bodo Moeller] 196264331Sjkim 197264331Sjkim *) Fix for the attack described in the paper "Recovering OpenSSL 198264331Sjkim ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack" 199264331Sjkim by Yuval Yarom and Naomi Benger. Details can be obtained from: 200264331Sjkim http://eprint.iacr.org/2014/140 201264331Sjkim 202264331Sjkim Thanks to Yuval Yarom and Naomi Benger for discovering this 203264331Sjkim flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076) 204264331Sjkim [Yuval Yarom and Naomi Benger] 205264331Sjkim 206264331Sjkim *) TLS pad extension: draft-agl-tls-padding-03 207264331Sjkim 208264331Sjkim Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the 209264331Sjkim TLS client Hello record length value would otherwise be > 255 and 210264331Sjkim less that 512 pad with a dummy extension containing zeroes so it 211264331Sjkim is at least 512 bytes long. 212264331Sjkim 213264331Sjkim [Adam Langley, Steve Henson] 214264331Sjkim 215264331Sjkim Changes between 1.0.1e and 1.0.1f [6 Jan 2014] 216264331Sjkim 217264331Sjkim *) Fix for TLS record tampering bug. A carefully crafted invalid 218264331Sjkim handshake could crash OpenSSL with a NULL pointer exception. 219264331Sjkim Thanks to Anton Johansson for reporting this issues. 220264331Sjkim (CVE-2013-4353) 221264331Sjkim 222264331Sjkim *) Keep original DTLS digest and encryption contexts in retransmission 223264331Sjkim structures so we can use the previous session parameters if they need 224264331Sjkim to be resent. (CVE-2013-6450) 225264331Sjkim [Steve Henson] 226264331Sjkim 227264331Sjkim *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which 228264331Sjkim avoids preferring ECDHE-ECDSA ciphers when the client appears to be 229264331Sjkim Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for 230264331Sjkim several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug 231264331Sjkim is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing 232264331Sjkim 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer. 233264331Sjkim [Rob Stradling, Adam Langley] 234264331Sjkim 235246772Sjkim Changes between 1.0.1d and 1.0.1e [11 Feb 2013] 236246772Sjkim 237264331Sjkim *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI 238264331Sjkim supporting platforms or when small records were transferred. 239264331Sjkim [Andy Polyakov, Steve Henson] 240246772Sjkim 241246772Sjkim Changes between 1.0.1c and 1.0.1d [5 Feb 2013] 242246772Sjkim 243246772Sjkim *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 244246772Sjkim 245246772Sjkim This addresses the flaw in CBC record processing discovered by 246246772Sjkim Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 247246772Sjkim at: http://www.isg.rhul.ac.uk/tls/ 248246772Sjkim 249246772Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 250246772Sjkim Security Group at Royal Holloway, University of London 251246772Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 252246772Sjkim Emilia K�sper for the initial patch. 253246772Sjkim (CVE-2013-0169) 254246772Sjkim [Emilia K�sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 255246772Sjkim 256246772Sjkim *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode 257246772Sjkim ciphersuites which can be exploited in a denial of service attack. 258246772Sjkim Thanks go to and to Adam Langley <agl@chromium.org> for discovering 259246772Sjkim and detecting this bug and to Wolfgang Ettlinger 260246772Sjkim <wolfgang.ettlinger@gmail.com> for independently discovering this issue. 261246772Sjkim (CVE-2012-2686) 262246772Sjkim [Adam Langley] 263246772Sjkim 264246772Sjkim *) Return an error when checking OCSP signatures when key is NULL. 265246772Sjkim This fixes a DoS attack. (CVE-2013-0166) 266246772Sjkim [Steve Henson] 267246772Sjkim 268246772Sjkim *) Make openssl verify return errors. 269246772Sjkim [Chris Palmer <palmer@google.com> and Ben Laurie] 270246772Sjkim 271246772Sjkim *) Call OCSP Stapling callback after ciphersuite has been chosen, so 272246772Sjkim the right response is stapled. Also change SSL_get_certificate() 273246772Sjkim so it returns the certificate actually sent. 274246772Sjkim See http://rt.openssl.org/Ticket/Display.html?id=2836. 275246772Sjkim [Rob Stradling <rob.stradling@comodo.com>] 276246772Sjkim 277246772Sjkim *) Fix possible deadlock when decoding public keys. 278246772Sjkim [Steve Henson] 279246772Sjkim 280246772Sjkim *) Don't use TLS 1.0 record version number in initial client hello 281246772Sjkim if renegotiating. 282246772Sjkim [Steve Henson] 283246772Sjkim 284238405Sjkim Changes between 1.0.1b and 1.0.1c [10 May 2012] 285237657Sjkim 286238405Sjkim *) Sanity check record length before skipping explicit IV in TLS 287238405Sjkim 1.2, 1.1 and DTLS to fix DoS attack. 288237657Sjkim 289237657Sjkim Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 290237657Sjkim fuzzing as a service testing platform. 291237657Sjkim (CVE-2012-2333) 292237657Sjkim [Steve Henson] 293237657Sjkim 294237657Sjkim *) Initialise tkeylen properly when encrypting CMS messages. 295237657Sjkim Thanks to Solar Designer of Openwall for reporting this issue. 296237657Sjkim [Steve Henson] 297237657Sjkim 298238405Sjkim *) In FIPS mode don't try to use composite ciphers as they are not 299238405Sjkim approved. 300238405Sjkim [Steve Henson] 301237657Sjkim 302238405Sjkim Changes between 1.0.1a and 1.0.1b [26 Apr 2012] 303237657Sjkim 304238405Sjkim *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and 305238405Sjkim 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately 306238405Sjkim mean any application compiled against OpenSSL 1.0.0 headers setting 307238405Sjkim SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng 308238405Sjkim TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to 309238405Sjkim 0x10000000L Any application which was previously compiled against 310238405Sjkim OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1 311238405Sjkim will need to be recompiled as a result. Letting be results in 312238405Sjkim inability to disable specifically TLS 1.1 and in client context, 313238405Sjkim in unlike event, limit maximum offered version to TLS 1.0 [see below]. 314238405Sjkim [Steve Henson] 315237657Sjkim 316238405Sjkim *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not 317238405Sjkim disable just protocol X, but all protocols above X *if* there are 318238405Sjkim protocols *below* X still enabled. In more practical terms it means 319238405Sjkim that if application wants to disable TLS1.0 in favor of TLS1.1 and 320238405Sjkim above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass 321238405Sjkim SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to 322238405Sjkim client side. 323238405Sjkim [Andy Polyakov] 324238405Sjkim 325238405Sjkim Changes between 1.0.1 and 1.0.1a [19 Apr 2012] 326238405Sjkim 327237657Sjkim *) Check for potentially exploitable overflows in asn1_d2i_read_bio 328237657Sjkim BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 329237657Sjkim in CRYPTO_realloc_clean. 330237657Sjkim 331237657Sjkim Thanks to Tavis Ormandy, Google Security Team, for discovering this 332237657Sjkim issue and to Adam Langley <agl@chromium.org> for fixing it. 333237657Sjkim (CVE-2012-2110) 334237657Sjkim [Adam Langley (Google), Tavis Ormandy, Google Security Team] 335237657Sjkim 336238405Sjkim *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections. 337238405Sjkim [Adam Langley] 338237657Sjkim 339238405Sjkim *) Workarounds for some broken servers that "hang" if a client hello 340238405Sjkim record length exceeds 255 bytes. 341238405Sjkim 342238405Sjkim 1. Do not use record version number > TLS 1.0 in initial client 343238405Sjkim hello: some (but not all) hanging servers will now work. 344238405Sjkim 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate 345238405Sjkim the number of ciphers sent in the client hello. This should be 346238405Sjkim set to an even number, such as 50, for example by passing: 347238405Sjkim -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure. 348238405Sjkim Most broken servers should now work. 349238405Sjkim 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable 350238405Sjkim TLS 1.2 client support entirely. 351238405Sjkim [Steve Henson] 352238405Sjkim 353238405Sjkim *) Fix SEGV in Vector Permutation AES module observed in OpenSSH. 354238405Sjkim [Andy Polyakov] 355238405Sjkim 356238405Sjkim Changes between 1.0.0h and 1.0.1 [14 Mar 2012] 357238405Sjkim 358238405Sjkim *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET 359238405Sjkim STRING form instead of a DigestInfo. 360238405Sjkim [Steve Henson] 361238405Sjkim 362238405Sjkim *) The format used for MDC2 RSA signatures is inconsistent between EVP 363238405Sjkim and the RSA_sign/RSA_verify functions. This was made more apparent when 364238405Sjkim OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular 365238405Sjkim those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect 366238405Sjkim the correct format in RSA_verify so both forms transparently work. 367238405Sjkim [Steve Henson] 368238405Sjkim 369238405Sjkim *) Some servers which support TLS 1.0 can choke if we initially indicate 370238405Sjkim support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA 371238405Sjkim encrypted premaster secret. As a workaround use the maximum pemitted 372238405Sjkim client version in client hello, this should keep such servers happy 373238405Sjkim and still work with previous versions of OpenSSL. 374238405Sjkim [Steve Henson] 375238405Sjkim 376238405Sjkim *) Add support for TLS/DTLS heartbeats. 377238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 378238405Sjkim 379238405Sjkim *) Add support for SCTP. 380238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 381238405Sjkim 382238405Sjkim *) Improved PRNG seeding for VOS. 383238405Sjkim [Paul Green <Paul.Green@stratus.com>] 384238405Sjkim 385238405Sjkim *) Extensive assembler packs updates, most notably: 386238405Sjkim 387238405Sjkim - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support; 388238405Sjkim - x86[_64]: SSSE3 support (SHA1, vector-permutation AES); 389238405Sjkim - x86_64: bit-sliced AES implementation; 390238405Sjkim - ARM: NEON support, contemporary platforms optimizations; 391238405Sjkim - s390x: z196 support; 392238405Sjkim - *: GHASH and GF(2^m) multiplication implementations; 393238405Sjkim 394238405Sjkim [Andy Polyakov] 395238405Sjkim 396238405Sjkim *) Make TLS-SRP code conformant with RFC 5054 API cleanup 397238405Sjkim (removal of unnecessary code) 398238405Sjkim [Peter Sylvester <peter.sylvester@edelweb.fr>] 399238405Sjkim 400238405Sjkim *) Add TLS key material exporter from RFC 5705. 401238405Sjkim [Eric Rescorla] 402238405Sjkim 403238405Sjkim *) Add DTLS-SRTP negotiation from RFC 5764. 404238405Sjkim [Eric Rescorla] 405238405Sjkim 406238405Sjkim *) Add Next Protocol Negotiation, 407238405Sjkim http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be 408238405Sjkim disabled with a no-npn flag to config or Configure. Code donated 409238405Sjkim by Google. 410238405Sjkim [Adam Langley <agl@google.com> and Ben Laurie] 411238405Sjkim 412238405Sjkim *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224, 413238405Sjkim NIST-P256, NIST-P521, with constant-time single point multiplication on 414238405Sjkim typical inputs. Compiler support for the nonstandard type __uint128_t is 415238405Sjkim required to use this (present in gcc 4.4 and later, for 64-bit builds). 416238405Sjkim Code made available under Apache License version 2.0. 417238405Sjkim 418238405Sjkim Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command 419238405Sjkim line to include this in your build of OpenSSL, and run "make depend" (or 420238405Sjkim "make update"). This enables the following EC_METHODs: 421238405Sjkim 422238405Sjkim EC_GFp_nistp224_method() 423238405Sjkim EC_GFp_nistp256_method() 424238405Sjkim EC_GFp_nistp521_method() 425238405Sjkim 426238405Sjkim EC_GROUP_new_by_curve_name() will automatically use these (while 427238405Sjkim EC_GROUP_new_curve_GFp() currently prefers the more flexible 428238405Sjkim implementations). 429238405Sjkim [Emilia K�sper, Adam Langley, Bodo Moeller (Google)] 430238405Sjkim 431238405Sjkim *) Use type ossl_ssize_t instad of ssize_t which isn't available on 432238405Sjkim all platforms. Move ssize_t definition from e_os.h to the public 433238405Sjkim header file e_os2.h as it now appears in public header file cms.h 434238405Sjkim [Steve Henson] 435238405Sjkim 436238405Sjkim *) New -sigopt option to the ca, req and x509 utilities. Additional 437238405Sjkim signature parameters can be passed using this option and in 438238405Sjkim particular PSS. 439238405Sjkim [Steve Henson] 440238405Sjkim 441238405Sjkim *) Add RSA PSS signing function. This will generate and set the 442238405Sjkim appropriate AlgorithmIdentifiers for PSS based on those in the 443238405Sjkim corresponding EVP_MD_CTX structure. No application support yet. 444238405Sjkim [Steve Henson] 445238405Sjkim 446238405Sjkim *) Support for companion algorithm specific ASN1 signing routines. 447238405Sjkim New function ASN1_item_sign_ctx() signs a pre-initialised 448238405Sjkim EVP_MD_CTX structure and sets AlgorithmIdentifiers based on 449238405Sjkim the appropriate parameters. 450238405Sjkim [Steve Henson] 451238405Sjkim 452238405Sjkim *) Add new algorithm specific ASN1 verification initialisation function 453238405Sjkim to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1 454238405Sjkim handling will be the same no matter what EVP_PKEY_METHOD is used. 455238405Sjkim Add a PSS handler to support verification of PSS signatures: checked 456238405Sjkim against a number of sample certificates. 457238405Sjkim [Steve Henson] 458238405Sjkim 459238405Sjkim *) Add signature printing for PSS. Add PSS OIDs. 460238405Sjkim [Steve Henson, Martin Kaiser <lists@kaiser.cx>] 461238405Sjkim 462238405Sjkim *) Add algorithm specific signature printing. An individual ASN1 method 463238405Sjkim can now print out signatures instead of the standard hex dump. 464238405Sjkim 465238405Sjkim More complex signatures (e.g. PSS) can print out more meaningful 466238405Sjkim information. Include DSA version that prints out the signature 467238405Sjkim parameters r, s. 468238405Sjkim [Steve Henson] 469238405Sjkim 470238405Sjkim *) Password based recipient info support for CMS library: implementing 471238405Sjkim RFC3211. 472238405Sjkim [Steve Henson] 473238405Sjkim 474238405Sjkim *) Split password based encryption into PBES2 and PBKDF2 functions. This 475238405Sjkim neatly separates the code into cipher and PBE sections and is required 476238405Sjkim for some algorithms that split PBES2 into separate pieces (such as 477238405Sjkim password based CMS). 478238405Sjkim [Steve Henson] 479238405Sjkim 480238405Sjkim *) Session-handling fixes: 481238405Sjkim - Fix handling of connections that are resuming with a session ID, 482238405Sjkim but also support Session Tickets. 483238405Sjkim - Fix a bug that suppressed issuing of a new ticket if the client 484238405Sjkim presented a ticket with an expired session. 485238405Sjkim - Try to set the ticket lifetime hint to something reasonable. 486238405Sjkim - Make tickets shorter by excluding irrelevant information. 487238405Sjkim - On the client side, don't ignore renewed tickets. 488238405Sjkim [Adam Langley, Bodo Moeller (Google)] 489238405Sjkim 490238405Sjkim *) Fix PSK session representation. 491238405Sjkim [Bodo Moeller] 492238405Sjkim 493238405Sjkim *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations. 494238405Sjkim 495238405Sjkim This work was sponsored by Intel. 496238405Sjkim [Andy Polyakov] 497238405Sjkim 498238405Sjkim *) Add GCM support to TLS library. Some custom code is needed to split 499238405Sjkim the IV between the fixed (from PRF) and explicit (from TLS record) 500238405Sjkim portions. This adds all GCM ciphersuites supported by RFC5288 and 501238405Sjkim RFC5289. Generalise some AES* cipherstrings to inlclude GCM and 502238405Sjkim add a special AESGCM string for GCM only. 503238405Sjkim [Steve Henson] 504238405Sjkim 505238405Sjkim *) Expand range of ctrls for AES GCM. Permit setting invocation 506238405Sjkim field on decrypt and retrieval of invocation field only on encrypt. 507238405Sjkim [Steve Henson] 508238405Sjkim 509238405Sjkim *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support. 510238405Sjkim As required by RFC5289 these ciphersuites cannot be used if for 511238405Sjkim versions of TLS earlier than 1.2. 512238405Sjkim [Steve Henson] 513238405Sjkim 514238405Sjkim *) For FIPS capable OpenSSL interpret a NULL default public key method 515238405Sjkim as unset and return the appopriate default but do *not* set the default. 516238405Sjkim This means we can return the appopriate method in applications that 517238405Sjkim swicth between FIPS and non-FIPS modes. 518238405Sjkim [Steve Henson] 519238405Sjkim 520238405Sjkim *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an 521238405Sjkim ENGINE is used then we cannot handle that in the FIPS module so we 522238405Sjkim keep original code iff non-FIPS operations are allowed. 523238405Sjkim [Steve Henson] 524238405Sjkim 525238405Sjkim *) Add -attime option to openssl utilities. 526238405Sjkim [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson] 527238405Sjkim 528238405Sjkim *) Redirect DSA and DH operations to FIPS module in FIPS mode. 529238405Sjkim [Steve Henson] 530238405Sjkim 531238405Sjkim *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use 532238405Sjkim FIPS EC methods unconditionally for now. 533238405Sjkim [Steve Henson] 534238405Sjkim 535238405Sjkim *) New build option no-ec2m to disable characteristic 2 code. 536238405Sjkim [Steve Henson] 537238405Sjkim 538238405Sjkim *) Backport libcrypto audit of return value checking from 1.1.0-dev; not 539238405Sjkim all cases can be covered as some introduce binary incompatibilities. 540238405Sjkim [Steve Henson] 541238405Sjkim 542238405Sjkim *) Redirect RSA operations to FIPS module including keygen, 543238405Sjkim encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods. 544238405Sjkim [Steve Henson] 545238405Sjkim 546238405Sjkim *) Add similar low level API blocking to ciphers. 547238405Sjkim [Steve Henson] 548238405Sjkim 549238405Sjkim *) Low level digest APIs are not approved in FIPS mode: any attempt 550238405Sjkim to use these will cause a fatal error. Applications that *really* want 551238405Sjkim to use them can use the private_* version instead. 552238405Sjkim [Steve Henson] 553238405Sjkim 554238405Sjkim *) Redirect cipher operations to FIPS module for FIPS builds. 555238405Sjkim [Steve Henson] 556238405Sjkim 557238405Sjkim *) Redirect digest operations to FIPS module for FIPS builds. 558238405Sjkim [Steve Henson] 559238405Sjkim 560238405Sjkim *) Update build system to add "fips" flag which will link in fipscanister.o 561238405Sjkim for static and shared library builds embedding a signature if needed. 562238405Sjkim [Steve Henson] 563238405Sjkim 564238405Sjkim *) Output TLS supported curves in preference order instead of numerical 565238405Sjkim order. This is currently hardcoded for the highest order curves first. 566238405Sjkim This should be configurable so applications can judge speed vs strength. 567238405Sjkim [Steve Henson] 568238405Sjkim 569238405Sjkim *) Add TLS v1.2 server support for client authentication. 570238405Sjkim [Steve Henson] 571238405Sjkim 572238405Sjkim *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers 573238405Sjkim and enable MD5. 574238405Sjkim [Steve Henson] 575238405Sjkim 576238405Sjkim *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying 577238405Sjkim FIPS modules versions. 578238405Sjkim [Steve Henson] 579238405Sjkim 580238405Sjkim *) Add TLS v1.2 client side support for client authentication. Keep cache 581238405Sjkim of handshake records longer as we don't know the hash algorithm to use 582238405Sjkim until after the certificate request message is received. 583238405Sjkim [Steve Henson] 584238405Sjkim 585238405Sjkim *) Initial TLS v1.2 client support. Add a default signature algorithms 586238405Sjkim extension including all the algorithms we support. Parse new signature 587238405Sjkim format in client key exchange. Relax some ECC signing restrictions for 588238405Sjkim TLS v1.2 as indicated in RFC5246. 589238405Sjkim [Steve Henson] 590238405Sjkim 591238405Sjkim *) Add server support for TLS v1.2 signature algorithms extension. Switch 592238405Sjkim to new signature format when needed using client digest preference. 593238405Sjkim All server ciphersuites should now work correctly in TLS v1.2. No client 594238405Sjkim support yet and no support for client certificates. 595238405Sjkim [Steve Henson] 596238405Sjkim 597238405Sjkim *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch 598238405Sjkim to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based 599238405Sjkim ciphersuites. At present only RSA key exchange ciphersuites work with 600238405Sjkim TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete 601238405Sjkim SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods 602238405Sjkim and version checking. 603238405Sjkim [Steve Henson] 604238405Sjkim 605238405Sjkim *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled 606238405Sjkim with this defined it will not be affected by any changes to ssl internal 607238405Sjkim structures. Add several utility functions to allow openssl application 608238405Sjkim to work with OPENSSL_NO_SSL_INTERN defined. 609238405Sjkim [Steve Henson] 610238405Sjkim 611238405Sjkim *) Add SRP support. 612238405Sjkim [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie] 613238405Sjkim 614238405Sjkim *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id. 615238405Sjkim [Steve Henson] 616238405Sjkim 617238405Sjkim *) Permit abbreviated handshakes when renegotiating using the function 618238405Sjkim SSL_renegotiate_abbreviated(). 619238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 620238405Sjkim 621238405Sjkim *) Add call to ENGINE_register_all_complete() to 622238405Sjkim ENGINE_load_builtin_engines(), so some implementations get used 623238405Sjkim automatically instead of needing explicit application support. 624238405Sjkim [Steve Henson] 625238405Sjkim 626238405Sjkim *) Add support for TLS key exporter as described in RFC5705. 627238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson] 628238405Sjkim 629238405Sjkim *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only 630238405Sjkim a few changes are required: 631238405Sjkim 632238405Sjkim Add SSL_OP_NO_TLSv1_1 flag. 633238405Sjkim Add TLSv1_1 methods. 634238405Sjkim Update version checking logic to handle version 1.1. 635238405Sjkim Add explicit IV handling (ported from DTLS code). 636238405Sjkim Add command line options to s_client/s_server. 637238405Sjkim [Steve Henson] 638238405Sjkim 639264331Sjkim Changes between 1.0.0j and 1.0.0k [5 Feb 2013] 640264331Sjkim 641264331Sjkim *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 642264331Sjkim 643264331Sjkim This addresses the flaw in CBC record processing discovered by 644264331Sjkim Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 645264331Sjkim at: http://www.isg.rhul.ac.uk/tls/ 646264331Sjkim 647264331Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 648264331Sjkim Security Group at Royal Holloway, University of London 649264331Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 650264331Sjkim Emilia K�sper for the initial patch. 651264331Sjkim (CVE-2013-0169) 652264331Sjkim [Emilia K�sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 653264331Sjkim 654264331Sjkim *) Return an error when checking OCSP signatures when key is NULL. 655264331Sjkim This fixes a DoS attack. (CVE-2013-0166) 656264331Sjkim [Steve Henson] 657264331Sjkim 658264331Sjkim *) Call OCSP Stapling callback after ciphersuite has been chosen, so 659264331Sjkim the right response is stapled. Also change SSL_get_certificate() 660264331Sjkim so it returns the certificate actually sent. 661264331Sjkim See http://rt.openssl.org/Ticket/Display.html?id=2836. 662264331Sjkim (This is a backport) 663264331Sjkim [Rob Stradling <rob.stradling@comodo.com>] 664264331Sjkim 665264331Sjkim *) Fix possible deadlock when decoding public keys. 666264331Sjkim [Steve Henson] 667264331Sjkim 668264331Sjkim Changes between 1.0.0i and 1.0.0j [10 May 2012] 669264331Sjkim 670264331Sjkim [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after 671264331Sjkim OpenSSL 1.0.1.] 672264331Sjkim 673264331Sjkim *) Sanity check record length before skipping explicit IV in DTLS 674264331Sjkim to fix DoS attack. 675264331Sjkim 676264331Sjkim Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 677264331Sjkim fuzzing as a service testing platform. 678264331Sjkim (CVE-2012-2333) 679264331Sjkim [Steve Henson] 680264331Sjkim 681264331Sjkim *) Initialise tkeylen properly when encrypting CMS messages. 682264331Sjkim Thanks to Solar Designer of Openwall for reporting this issue. 683264331Sjkim [Steve Henson] 684264331Sjkim 685264331Sjkim Changes between 1.0.0h and 1.0.0i [19 Apr 2012] 686264331Sjkim 687264331Sjkim *) Check for potentially exploitable overflows in asn1_d2i_read_bio 688264331Sjkim BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 689264331Sjkim in CRYPTO_realloc_clean. 690264331Sjkim 691264331Sjkim Thanks to Tavis Ormandy, Google Security Team, for discovering this 692264331Sjkim issue and to Adam Langley <agl@chromium.org> for fixing it. 693264331Sjkim (CVE-2012-2110) 694264331Sjkim [Adam Langley (Google), Tavis Ormandy, Google Security Team] 695264331Sjkim 696238405Sjkim Changes between 1.0.0g and 1.0.0h [12 Mar 2012] 697238405Sjkim 698237657Sjkim *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 699237657Sjkim in CMS and PKCS7 code. When RSA decryption fails use a random key for 700237657Sjkim content decryption and always return the same error. Note: this attack 701237657Sjkim needs on average 2^20 messages so it only affects automated senders. The 702237657Sjkim old behaviour can be reenabled in the CMS code by setting the 703237657Sjkim CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 704237657Sjkim an MMA defence is not necessary. 705237657Sjkim Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 706237657Sjkim this issue. (CVE-2012-0884) 707237657Sjkim [Steve Henson] 708237657Sjkim 709237657Sjkim *) Fix CVE-2011-4619: make sure we really are receiving a 710237657Sjkim client hello before rejecting multiple SGC restarts. Thanks to 711237657Sjkim Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 712237657Sjkim [Steve Henson] 713237657Sjkim 714238405Sjkim Changes between 1.0.0f and 1.0.0g [18 Jan 2012] 715238405Sjkim 716238405Sjkim *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 717238405Sjkim Thanks to Antonio Martin, Enterprise Secure Access Research and 718238405Sjkim Development, Cisco Systems, Inc. for discovering this bug and 719238405Sjkim preparing a fix. (CVE-2012-0050) 720238405Sjkim [Antonio Martin] 721238405Sjkim 722238405Sjkim Changes between 1.0.0e and 1.0.0f [4 Jan 2012] 723238405Sjkim 724238405Sjkim *) Nadhem Alfardan and Kenny Paterson have discovered an extension 725238405Sjkim of the Vaudenay padding oracle attack on CBC mode encryption 726238405Sjkim which enables an efficient plaintext recovery attack against 727238405Sjkim the OpenSSL implementation of DTLS. Their attack exploits timing 728238405Sjkim differences arising during decryption processing. A research 729238405Sjkim paper describing this attack can be found at: 730238405Sjkim http://www.isg.rhul.ac.uk/~kp/dtls.pdf 731238405Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 732238405Sjkim Security Group at Royal Holloway, University of London 733238405Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 734238405Sjkim <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 735238405Sjkim for preparing the fix. (CVE-2011-4108) 736238405Sjkim [Robin Seggelmann, Michael Tuexen] 737238405Sjkim 738238405Sjkim *) Clear bytes used for block padding of SSL 3.0 records. 739238405Sjkim (CVE-2011-4576) 740238405Sjkim [Adam Langley (Google)] 741238405Sjkim 742238405Sjkim *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 743238405Sjkim Kadianakis <desnacked@gmail.com> for discovering this issue and 744238405Sjkim Adam Langley for preparing the fix. (CVE-2011-4619) 745238405Sjkim [Adam Langley (Google)] 746238405Sjkim 747238405Sjkim *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027) 748238405Sjkim [Andrey Kulikov <amdeich@gmail.com>] 749238405Sjkim 750238405Sjkim *) Prevent malformed RFC3779 data triggering an assertion failure. 751238405Sjkim Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 752238405Sjkim and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 753238405Sjkim [Rob Austein <sra@hactrn.net>] 754238405Sjkim 755238405Sjkim *) Improved PRNG seeding for VOS. 756238405Sjkim [Paul Green <Paul.Green@stratus.com>] 757238405Sjkim 758238405Sjkim *) Fix ssl_ciph.c set-up race. 759238405Sjkim [Adam Langley (Google)] 760238405Sjkim 761238405Sjkim *) Fix spurious failures in ecdsatest.c. 762238405Sjkim [Emilia K�sper (Google)] 763238405Sjkim 764238405Sjkim *) Fix the BIO_f_buffer() implementation (which was mixing different 765238405Sjkim interpretations of the '..._len' fields). 766238405Sjkim [Adam Langley (Google)] 767238405Sjkim 768238405Sjkim *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 769238405Sjkim BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 770238405Sjkim threads won't reuse the same blinding coefficients. 771238405Sjkim 772238405Sjkim This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 773238405Sjkim lock to call BN_BLINDING_invert_ex, and avoids one use of 774238405Sjkim BN_BLINDING_update for each BN_BLINDING structure (previously, 775238405Sjkim the last update always remained unused). 776238405Sjkim [Emilia K�sper (Google)] 777238405Sjkim 778238405Sjkim *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf. 779238405Sjkim [Bob Buckholz (Google)] 780238405Sjkim 781238405Sjkim Changes between 1.0.0d and 1.0.0e [6 Sep 2011] 782238405Sjkim 783238405Sjkim *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted 784238405Sjkim by initialising X509_STORE_CTX properly. (CVE-2011-3207) 785238405Sjkim [Kaspar Brand <ossl@velox.ch>] 786238405Sjkim 787238405Sjkim *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 788238405Sjkim for multi-threaded use of ECDH. (CVE-2011-3210) 789238405Sjkim [Adam Langley (Google)] 790238405Sjkim 791238405Sjkim *) Fix x509_name_ex_d2i memory leak on bad inputs. 792238405Sjkim [Bodo Moeller] 793238405Sjkim 794238405Sjkim *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check 795238405Sjkim signature public key algorithm by using OID xref utilities instead. 796238405Sjkim Before this you could only use some ECC ciphersuites with SHA1 only. 797238405Sjkim [Steve Henson] 798238405Sjkim 799238405Sjkim *) Add protection against ECDSA timing attacks as mentioned in the paper 800238405Sjkim by Billy Bob Brumley and Nicola Tuveri, see: 801238405Sjkim 802238405Sjkim http://eprint.iacr.org/2011/232.pdf 803238405Sjkim 804238405Sjkim [Billy Bob Brumley and Nicola Tuveri] 805238405Sjkim 806238405Sjkim Changes between 1.0.0c and 1.0.0d [8 Feb 2011] 807238405Sjkim 808238405Sjkim *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 809238405Sjkim [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 810238405Sjkim 811238405Sjkim *) Fix bug in string printing code: if *any* escaping is enabled we must 812238405Sjkim escape the escape character (backslash) or the resulting string is 813238405Sjkim ambiguous. 814238405Sjkim [Steve Henson] 815238405Sjkim 816238405Sjkim Changes between 1.0.0b and 1.0.0c [2 Dec 2010] 817238405Sjkim 818238405Sjkim *) Disable code workaround for ancient and obsolete Netscape browsers 819238405Sjkim and servers: an attacker can use it in a ciphersuite downgrade attack. 820238405Sjkim Thanks to Martin Rex for discovering this bug. CVE-2010-4180 821238405Sjkim [Steve Henson] 822238405Sjkim 823238405Sjkim *) Fixed J-PAKE implementation error, originally discovered by 824238405Sjkim Sebastien Martini, further info and confirmation from Stefan 825238405Sjkim Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 826238405Sjkim [Ben Laurie] 827238405Sjkim 828238405Sjkim Changes between 1.0.0a and 1.0.0b [16 Nov 2010] 829238405Sjkim 830238405Sjkim *) Fix extension code to avoid race conditions which can result in a buffer 831238405Sjkim overrun vulnerability: resumed sessions must not be modified as they can 832238405Sjkim be shared by multiple threads. CVE-2010-3864 833238405Sjkim [Steve Henson] 834238405Sjkim 835238405Sjkim *) Fix WIN32 build system to correctly link an ENGINE directory into 836238405Sjkim a DLL. 837238405Sjkim [Steve Henson] 838238405Sjkim 839238405Sjkim Changes between 1.0.0 and 1.0.0a [01 Jun 2010] 840238405Sjkim 841238405Sjkim *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover 842238405Sjkim (CVE-2010-1633) 843238405Sjkim [Steve Henson, Peter-Michael Hager <hager@dortmund.net>] 844238405Sjkim 845238405Sjkim Changes between 0.9.8n and 1.0.0 [29 Mar 2010] 846238405Sjkim 847238405Sjkim *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher 848238405Sjkim context. The operation can be customised via the ctrl mechanism in 849238405Sjkim case ENGINEs want to include additional functionality. 850238405Sjkim [Steve Henson] 851238405Sjkim 852238405Sjkim *) Tolerate yet another broken PKCS#8 key format: private key value negative. 853238405Sjkim [Steve Henson] 854238405Sjkim 855238405Sjkim *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to 856238405Sjkim output hashes compatible with older versions of OpenSSL. 857238405Sjkim [Willy Weisz <weisz@vcpc.univie.ac.at>] 858238405Sjkim 859238405Sjkim *) Fix compression algorithm handling: if resuming a session use the 860238405Sjkim compression algorithm of the resumed session instead of determining 861238405Sjkim it from client hello again. Don't allow server to change algorithm. 862238405Sjkim [Steve Henson] 863238405Sjkim 864238405Sjkim *) Add load_crls() function to apps tidying load_certs() too. Add option 865238405Sjkim to verify utility to allow additional CRLs to be included. 866238405Sjkim [Steve Henson] 867238405Sjkim 868238405Sjkim *) Update OCSP request code to permit adding custom headers to the request: 869238405Sjkim some responders need this. 870238405Sjkim [Steve Henson] 871238405Sjkim 872238405Sjkim *) The function EVP_PKEY_sign() returns <=0 on error: check return code 873238405Sjkim correctly. 874238405Sjkim [Julia Lawall <julia@diku.dk>] 875238405Sjkim 876238405Sjkim *) Update verify callback code in apps/s_cb.c and apps/verify.c, it 877238405Sjkim needlessly dereferenced structures, used obsolete functions and 878238405Sjkim didn't handle all updated verify codes correctly. 879238405Sjkim [Steve Henson] 880238405Sjkim 881238405Sjkim *) Disable MD2 in the default configuration. 882238405Sjkim [Steve Henson] 883238405Sjkim 884238405Sjkim *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to 885238405Sjkim indicate the initial BIO being pushed or popped. This makes it possible 886238405Sjkim to determine whether the BIO is the one explicitly called or as a result 887238405Sjkim of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so 888238405Sjkim it handles reference counts correctly and doesn't zero out the I/O bio 889238405Sjkim when it is not being explicitly popped. WARNING: applications which 890238405Sjkim included workarounds for the old buggy behaviour will need to be modified 891238405Sjkim or they could free up already freed BIOs. 892238405Sjkim [Steve Henson] 893238405Sjkim 894238405Sjkim *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni 895238405Sjkim renaming to all platforms (within the 0.9.8 branch, this was 896238405Sjkim done conditionally on Netware platforms to avoid a name clash). 897238405Sjkim [Guenter <lists@gknw.net>] 898238405Sjkim 899238405Sjkim *) Add ECDHE and PSK support to DTLS. 900238405Sjkim [Michael Tuexen <tuexen@fh-muenster.de>] 901238405Sjkim 902238405Sjkim *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't 903238405Sjkim be used on C++. 904238405Sjkim [Steve Henson] 905238405Sjkim 906238405Sjkim *) Add "missing" function EVP_MD_flags() (without this the only way to 907238405Sjkim retrieve a digest flags is by accessing the structure directly. Update 908238405Sjkim EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest 909238405Sjkim or cipher is registered as in the "from" argument. Print out all 910238405Sjkim registered digests in the dgst usage message instead of manually 911238405Sjkim attempting to work them out. 912238405Sjkim [Steve Henson] 913238405Sjkim 914238405Sjkim *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello: 915238405Sjkim this allows the use of compression and extensions. Change default cipher 916238405Sjkim string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2 917238405Sjkim by default unless an application cipher string requests it. 918238405Sjkim [Steve Henson] 919238405Sjkim 920238405Sjkim *) Alter match criteria in PKCS12_parse(). It used to try to use local 921238405Sjkim key ids to find matching certificates and keys but some PKCS#12 files 922238405Sjkim don't follow the (somewhat unwritten) rules and this strategy fails. 923238405Sjkim Now just gather all certificates together and the first private key 924238405Sjkim then look for the first certificate that matches the key. 925238405Sjkim [Steve Henson] 926238405Sjkim 927238405Sjkim *) Support use of registered digest and cipher names for dgst and cipher 928238405Sjkim commands instead of having to add each one as a special case. So now 929238405Sjkim you can do: 930238405Sjkim 931238405Sjkim openssl sha256 foo 932238405Sjkim 933238405Sjkim as well as: 934238405Sjkim 935238405Sjkim openssl dgst -sha256 foo 936238405Sjkim 937238405Sjkim and this works for ENGINE based algorithms too. 938238405Sjkim 939238405Sjkim [Steve Henson] 940238405Sjkim 941238405Sjkim *) Update Gost ENGINE to support parameter files. 942238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 943238405Sjkim 944238405Sjkim *) Support GeneralizedTime in ca utility. 945238405Sjkim [Oliver Martin <oliver@volatilevoid.net>, Steve Henson] 946238405Sjkim 947238405Sjkim *) Enhance the hash format used for certificate directory links. The new 948238405Sjkim form uses the canonical encoding (meaning equivalent names will work 949238405Sjkim even if they aren't identical) and uses SHA1 instead of MD5. This form 950238405Sjkim is incompatible with the older format and as a result c_rehash should 951238405Sjkim be used to rebuild symbolic links. 952238405Sjkim [Steve Henson] 953238405Sjkim 954238405Sjkim *) Make PKCS#8 the default write format for private keys, replacing the 955238405Sjkim traditional format. This form is standardised, more secure and doesn't 956238405Sjkim include an implicit MD5 dependency. 957238405Sjkim [Steve Henson] 958238405Sjkim 959238405Sjkim *) Add a $gcc_devteam_warn option to Configure. The idea is that any code 960238405Sjkim committed to OpenSSL should pass this lot as a minimum. 961238405Sjkim [Steve Henson] 962238405Sjkim 963238405Sjkim *) Add session ticket override functionality for use by EAP-FAST. 964238405Sjkim [Jouni Malinen <j@w1.fi>] 965238405Sjkim 966238405Sjkim *) Modify HMAC functions to return a value. Since these can be implemented 967238405Sjkim in an ENGINE errors can occur. 968238405Sjkim [Steve Henson] 969238405Sjkim 970238405Sjkim *) Type-checked OBJ_bsearch_ex. 971238405Sjkim [Ben Laurie] 972238405Sjkim 973238405Sjkim *) Type-checked OBJ_bsearch. Also some constification necessitated 974238405Sjkim by type-checking. Still to come: TXT_DB, bsearch(?), 975238405Sjkim OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING, 976238405Sjkim CONF_VALUE. 977238405Sjkim [Ben Laurie] 978238405Sjkim 979238405Sjkim *) New function OPENSSL_gmtime_adj() to add a specific number of days and 980238405Sjkim seconds to a tm structure directly, instead of going through OS 981238405Sjkim specific date routines. This avoids any issues with OS routines such 982238405Sjkim as the year 2038 bug. New *_adj() functions for ASN1 time structures 983238405Sjkim and X509_time_adj_ex() to cover the extended range. The existing 984238405Sjkim X509_time_adj() is still usable and will no longer have any date issues. 985238405Sjkim [Steve Henson] 986238405Sjkim 987238405Sjkim *) Delta CRL support. New use deltas option which will attempt to locate 988238405Sjkim and search any appropriate delta CRLs available. 989238405Sjkim 990238405Sjkim This work was sponsored by Google. 991238405Sjkim [Steve Henson] 992238405Sjkim 993238405Sjkim *) Support for CRLs partitioned by reason code. Reorganise CRL processing 994238405Sjkim code and add additional score elements. Validate alternate CRL paths 995238405Sjkim as part of the CRL checking and indicate a new error "CRL path validation 996238405Sjkim error" in this case. Applications wanting additional details can use 997238405Sjkim the verify callback and check the new "parent" field. If this is not 998238405Sjkim NULL CRL path validation is taking place. Existing applications wont 999238405Sjkim see this because it requires extended CRL support which is off by 1000238405Sjkim default. 1001238405Sjkim 1002238405Sjkim This work was sponsored by Google. 1003238405Sjkim [Steve Henson] 1004238405Sjkim 1005238405Sjkim *) Support for freshest CRL extension. 1006238405Sjkim 1007238405Sjkim This work was sponsored by Google. 1008238405Sjkim [Steve Henson] 1009238405Sjkim 1010238405Sjkim *) Initial indirect CRL support. Currently only supported in the CRLs 1011238405Sjkim passed directly and not via lookup. Process certificate issuer 1012238405Sjkim CRL entry extension and lookup CRL entries by bother issuer name 1013238405Sjkim and serial number. Check and process CRL issuer entry in IDP extension. 1014238405Sjkim 1015238405Sjkim This work was sponsored by Google. 1016238405Sjkim [Steve Henson] 1017238405Sjkim 1018238405Sjkim *) Add support for distinct certificate and CRL paths. The CRL issuer 1019238405Sjkim certificate is validated separately in this case. Only enabled if 1020238405Sjkim an extended CRL support flag is set: this flag will enable additional 1021238405Sjkim CRL functionality in future. 1022238405Sjkim 1023238405Sjkim This work was sponsored by Google. 1024238405Sjkim [Steve Henson] 1025238405Sjkim 1026238405Sjkim *) Add support for policy mappings extension. 1027238405Sjkim 1028238405Sjkim This work was sponsored by Google. 1029238405Sjkim [Steve Henson] 1030238405Sjkim 1031238405Sjkim *) Fixes to pathlength constraint, self issued certificate handling, 1032238405Sjkim policy processing to align with RFC3280 and PKITS tests. 1033238405Sjkim 1034238405Sjkim This work was sponsored by Google. 1035238405Sjkim [Steve Henson] 1036238405Sjkim 1037238405Sjkim *) Support for name constraints certificate extension. DN, email, DNS 1038238405Sjkim and URI types are currently supported. 1039238405Sjkim 1040238405Sjkim This work was sponsored by Google. 1041238405Sjkim [Steve Henson] 1042238405Sjkim 1043238405Sjkim *) To cater for systems that provide a pointer-based thread ID rather 1044238405Sjkim than numeric, deprecate the current numeric thread ID mechanism and 1045238405Sjkim replace it with a structure and associated callback type. This 1046238405Sjkim mechanism allows a numeric "hash" to be extracted from a thread ID in 1047238405Sjkim either case, and on platforms where pointers are larger than 'long', 1048238405Sjkim mixing is done to help ensure the numeric 'hash' is usable even if it 1049238405Sjkim can't be guaranteed unique. The default mechanism is to use "&errno" 1050238405Sjkim as a pointer-based thread ID to distinguish between threads. 1051238405Sjkim 1052238405Sjkim Applications that want to provide their own thread IDs should now use 1053238405Sjkim CRYPTO_THREADID_set_callback() to register a callback that will call 1054238405Sjkim either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer(). 1055238405Sjkim 1056238405Sjkim Note that ERR_remove_state() is now deprecated, because it is tied 1057238405Sjkim to the assumption that thread IDs are numeric. ERR_remove_state(0) 1058238405Sjkim to free the current thread's error state should be replaced by 1059238405Sjkim ERR_remove_thread_state(NULL). 1060238405Sjkim 1061238405Sjkim (This new approach replaces the functions CRYPTO_set_idptr_callback(), 1062238405Sjkim CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in 1063238405Sjkim OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an 1064238405Sjkim application was previously providing a numeric thread callback that 1065238405Sjkim was inappropriate for distinguishing threads, then uniqueness might 1066238405Sjkim have been obtained with &errno that happened immediately in the 1067238405Sjkim intermediate development versions of OpenSSL; this is no longer the 1068238405Sjkim case, the numeric thread callback will now override the automatic use 1069238405Sjkim of &errno.) 1070238405Sjkim [Geoff Thorpe, with help from Bodo Moeller] 1071238405Sjkim 1072238405Sjkim *) Initial support for different CRL issuing certificates. This covers a 1073238405Sjkim simple case where the self issued certificates in the chain exist and 1074238405Sjkim the real CRL issuer is higher in the existing chain. 1075238405Sjkim 1076238405Sjkim This work was sponsored by Google. 1077238405Sjkim [Steve Henson] 1078238405Sjkim 1079238405Sjkim *) Removed effectively defunct crypto/store from the build. 1080238405Sjkim [Ben Laurie] 1081238405Sjkim 1082238405Sjkim *) Revamp of STACK to provide stronger type-checking. Still to come: 1083238405Sjkim TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE, 1084238405Sjkim ASN1_STRING, CONF_VALUE. 1085238405Sjkim [Ben Laurie] 1086238405Sjkim 1087238405Sjkim *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer 1088238405Sjkim RAM on SSL connections. This option can save about 34k per idle SSL. 1089238405Sjkim [Nick Mathewson] 1090238405Sjkim 1091238405Sjkim *) Revamp of LHASH to provide stronger type-checking. Still to come: 1092238405Sjkim STACK, TXT_DB, bsearch, qsort. 1093238405Sjkim [Ben Laurie] 1094238405Sjkim 1095238405Sjkim *) Initial support for Cryptographic Message Syntax (aka CMS) based 1096238405Sjkim on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility, 1097238405Sjkim support for data, signedData, compressedData, digestedData and 1098238405Sjkim encryptedData, envelopedData types included. Scripts to check against 1099238405Sjkim RFC4134 examples draft and interop and consistency checks of many 1100238405Sjkim content types and variants. 1101238405Sjkim [Steve Henson] 1102238405Sjkim 1103238405Sjkim *) Add options to enc utility to support use of zlib compression BIO. 1104238405Sjkim [Steve Henson] 1105238405Sjkim 1106238405Sjkim *) Extend mk1mf to support importing of options and assembly language 1107238405Sjkim files from Configure script, currently only included in VC-WIN32. 1108238405Sjkim The assembly language rules can now optionally generate the source 1109238405Sjkim files from the associated perl scripts. 1110238405Sjkim [Steve Henson] 1111238405Sjkim 1112238405Sjkim *) Implement remaining functionality needed to support GOST ciphersuites. 1113238405Sjkim Interop testing has been performed using CryptoPro implementations. 1114238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 1115238405Sjkim 1116238405Sjkim *) s390x assembler pack. 1117238405Sjkim [Andy Polyakov] 1118238405Sjkim 1119238405Sjkim *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU 1120238405Sjkim "family." 1121238405Sjkim [Andy Polyakov] 1122238405Sjkim 1123238405Sjkim *) Implement Opaque PRF Input TLS extension as specified in 1124238405Sjkim draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an 1125238405Sjkim official specification yet and no extension type assignment by 1126238405Sjkim IANA exists, this extension (for now) will have to be explicitly 1127238405Sjkim enabled when building OpenSSL by providing the extension number 1128238405Sjkim to use. For example, specify an option 1129238405Sjkim 1130238405Sjkim -DTLSEXT_TYPE_opaque_prf_input=0x9527 1131238405Sjkim 1132238405Sjkim to the "config" or "Configure" script to enable the extension, 1133238405Sjkim assuming extension number 0x9527 (which is a completely arbitrary 1134238405Sjkim and unofficial assignment based on the MD5 hash of the Internet 1135238405Sjkim Draft). Note that by doing so, you potentially lose 1136238405Sjkim interoperability with other TLS implementations since these might 1137238405Sjkim be using the same extension number for other purposes. 1138238405Sjkim 1139238405Sjkim SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the 1140238405Sjkim opaque PRF input value to use in the handshake. This will create 1141238405Sjkim an interal copy of the length-'len' string at 'src', and will 1142238405Sjkim return non-zero for success. 1143238405Sjkim 1144238405Sjkim To get more control and flexibility, provide a callback function 1145238405Sjkim by using 1146238405Sjkim 1147238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb) 1148238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg) 1149238405Sjkim 1150238405Sjkim where 1151238405Sjkim 1152238405Sjkim int (*cb)(SSL *, void *peerinput, size_t len, void *arg); 1153238405Sjkim void *arg; 1154238405Sjkim 1155238405Sjkim Callback function 'cb' will be called in handshakes, and is 1156238405Sjkim expected to use SSL_set_tlsext_opaque_prf_input() as appropriate. 1157238405Sjkim Argument 'arg' is for application purposes (the value as given to 1158238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly 1159238405Sjkim be provided to the callback function). The callback function 1160238405Sjkim has to return non-zero to report success: usually 1 to use opaque 1161238405Sjkim PRF input just if possible, or 2 to enforce use of the opaque PRF 1162238405Sjkim input. In the latter case, the library will abort the handshake 1163238405Sjkim if opaque PRF input is not successfully negotiated. 1164238405Sjkim 1165238405Sjkim Arguments 'peerinput' and 'len' given to the callback function 1166238405Sjkim will always be NULL and 0 in the case of a client. A server will 1167238405Sjkim see the client's opaque PRF input through these variables if 1168238405Sjkim available (NULL and 0 otherwise). Note that if the server 1169238405Sjkim provides an opaque PRF input, the length must be the same as the 1170238405Sjkim length of the client's opaque PRF input. 1171238405Sjkim 1172238405Sjkim Note that the callback function will only be called when creating 1173238405Sjkim a new session (session resumption can resume whatever was 1174238405Sjkim previously negotiated), and will not be called in SSL 2.0 1175238405Sjkim handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or 1176238405Sjkim SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended 1177238405Sjkim for applications that need to enforce opaque PRF input. 1178238405Sjkim 1179238405Sjkim [Bodo Moeller] 1180238405Sjkim 1181238405Sjkim *) Update ssl code to support digests other than SHA1+MD5 for handshake 1182238405Sjkim MAC. 1183238405Sjkim 1184238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 1185238405Sjkim 1186238405Sjkim *) Add RFC4507 support to OpenSSL. This includes the corrections in 1187238405Sjkim RFC4507bis. The encrypted ticket format is an encrypted encoded 1188238405Sjkim SSL_SESSION structure, that way new session features are automatically 1189238405Sjkim supported. 1190238405Sjkim 1191238405Sjkim If a client application caches session in an SSL_SESSION structure 1192238405Sjkim support is transparent because tickets are now stored in the encoded 1193238405Sjkim SSL_SESSION. 1194238405Sjkim 1195238405Sjkim The SSL_CTX structure automatically generates keys for ticket 1196238405Sjkim protection in servers so again support should be possible 1197238405Sjkim with no application modification. 1198238405Sjkim 1199238405Sjkim If a client or server wishes to disable RFC4507 support then the option 1200238405Sjkim SSL_OP_NO_TICKET can be set. 1201238405Sjkim 1202238405Sjkim Add a TLS extension debugging callback to allow the contents of any client 1203238405Sjkim or server extensions to be examined. 1204238405Sjkim 1205238405Sjkim This work was sponsored by Google. 1206238405Sjkim [Steve Henson] 1207238405Sjkim 1208238405Sjkim *) Final changes to avoid use of pointer pointer casts in OpenSSL. 1209238405Sjkim OpenSSL should now compile cleanly on gcc 4.2 1210238405Sjkim [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson] 1211238405Sjkim 1212238405Sjkim *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC 1213238405Sjkim support including streaming MAC support: this is required for GOST 1214238405Sjkim ciphersuite support. 1215238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson] 1216238405Sjkim 1217238405Sjkim *) Add option -stream to use PKCS#7 streaming in smime utility. New 1218238405Sjkim function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream() 1219238405Sjkim to output in BER and PEM format. 1220238405Sjkim [Steve Henson] 1221238405Sjkim 1222238405Sjkim *) Experimental support for use of HMAC via EVP_PKEY interface. This 1223238405Sjkim allows HMAC to be handled via the EVP_DigestSign*() interface. The 1224238405Sjkim EVP_PKEY "key" in this case is the HMAC key, potentially allowing 1225238405Sjkim ENGINE support for HMAC keys which are unextractable. New -mac and 1226238405Sjkim -macopt options to dgst utility. 1227238405Sjkim [Steve Henson] 1228238405Sjkim 1229238405Sjkim *) New option -sigopt to dgst utility. Update dgst to use 1230238405Sjkim EVP_Digest{Sign,Verify}*. These two changes make it possible to use 1231238405Sjkim alternative signing paramaters such as X9.31 or PSS in the dgst 1232238405Sjkim utility. 1233238405Sjkim [Steve Henson] 1234238405Sjkim 1235238405Sjkim *) Change ssl_cipher_apply_rule(), the internal function that does 1236238405Sjkim the work each time a ciphersuite string requests enabling 1237238405Sjkim ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or 1238238405Sjkim removing ("!foo+bar") a class of ciphersuites: Now it maintains 1239238405Sjkim the order of disabled ciphersuites such that those ciphersuites 1240238405Sjkim that most recently went from enabled to disabled not only stay 1241238405Sjkim in order with respect to each other, but also have higher priority 1242238405Sjkim than other disabled ciphersuites the next time ciphersuites are 1243238405Sjkim enabled again. 1244238405Sjkim 1245238405Sjkim This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable 1246238405Sjkim the same ciphersuites as with "HIGH" alone, but in a specific 1247238405Sjkim order where the PSK ciphersuites come first (since they are the 1248238405Sjkim most recently disabled ciphersuites when "HIGH" is parsed). 1249238405Sjkim 1250238405Sjkim Also, change ssl_create_cipher_list() (using this new 1251238405Sjkim funcionality) such that between otherwise identical 1252238405Sjkim cihpersuites, ephemeral ECDH is preferred over ephemeral DH in 1253238405Sjkim the default order. 1254238405Sjkim [Bodo Moeller] 1255238405Sjkim 1256238405Sjkim *) Change ssl_create_cipher_list() so that it automatically 1257238405Sjkim arranges the ciphersuites in reasonable order before starting 1258238405Sjkim to process the rule string. Thus, the definition for "DEFAULT" 1259238405Sjkim (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but 1260238405Sjkim remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH". 1261238405Sjkim This makes it much easier to arrive at a reasonable default order 1262238405Sjkim in applications for which anonymous ciphers are OK (meaning 1263238405Sjkim that you can't actually use DEFAULT). 1264238405Sjkim [Bodo Moeller; suggested by Victor Duchovni] 1265238405Sjkim 1266238405Sjkim *) Split the SSL/TLS algorithm mask (as used for ciphersuite string 1267238405Sjkim processing) into multiple integers instead of setting 1268238405Sjkim "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK", 1269238405Sjkim "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer. 1270238405Sjkim (These masks as well as the individual bit definitions are hidden 1271238405Sjkim away into the non-exported interface ssl/ssl_locl.h, so this 1272238405Sjkim change to the definition of the SSL_CIPHER structure shouldn't 1273238405Sjkim affect applications.) This give us more bits for each of these 1274238405Sjkim categories, so there is no longer a need to coagulate AES128 and 1275238405Sjkim AES256 into a single algorithm bit, and to coagulate Camellia128 1276238405Sjkim and Camellia256 into a single algorithm bit, which has led to all 1277238405Sjkim kinds of kludges. 1278238405Sjkim 1279238405Sjkim Thus, among other things, the kludge introduced in 0.9.7m and 1280238405Sjkim 0.9.8e for masking out AES256 independently of AES128 or masking 1281238405Sjkim out Camellia256 independently of AES256 is not needed here in 0.9.9. 1282238405Sjkim 1283238405Sjkim With the change, we also introduce new ciphersuite aliases that 1284238405Sjkim so far were missing: "AES128", "AES256", "CAMELLIA128", and 1285238405Sjkim "CAMELLIA256". 1286238405Sjkim [Bodo Moeller] 1287238405Sjkim 1288238405Sjkim *) Add support for dsa-with-SHA224 and dsa-with-SHA256. 1289238405Sjkim Use the leftmost N bytes of the signature input if the input is 1290238405Sjkim larger than the prime q (with N being the size in bytes of q). 1291238405Sjkim [Nils Larsch] 1292238405Sjkim 1293238405Sjkim *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses 1294238405Sjkim it yet and it is largely untested. 1295238405Sjkim [Steve Henson] 1296238405Sjkim 1297238405Sjkim *) Add support for the ecdsa-with-SHA224/256/384/512 signature types. 1298238405Sjkim [Nils Larsch] 1299238405Sjkim 1300238405Sjkim *) Initial incomplete changes to avoid need for function casts in OpenSSL 1301238405Sjkim some compilers (gcc 4.2 and later) reject their use. Safestack is 1302238405Sjkim reimplemented. Update ASN1 to avoid use of legacy functions. 1303238405Sjkim [Steve Henson] 1304238405Sjkim 1305238405Sjkim *) Win32/64 targets are linked with Winsock2. 1306238405Sjkim [Andy Polyakov] 1307238405Sjkim 1308238405Sjkim *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected 1309238405Sjkim to external functions. This can be used to increase CRL handling 1310238405Sjkim efficiency especially when CRLs are very large by (for example) storing 1311238405Sjkim the CRL revoked certificates in a database. 1312238405Sjkim [Steve Henson] 1313238405Sjkim 1314238405Sjkim *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so 1315238405Sjkim new CRLs added to a directory can be used. New command line option 1316238405Sjkim -verify_return_error to s_client and s_server. This causes real errors 1317238405Sjkim to be returned by the verify callback instead of carrying on no matter 1318238405Sjkim what. This reflects the way a "real world" verify callback would behave. 1319238405Sjkim [Steve Henson] 1320238405Sjkim 1321238405Sjkim *) GOST engine, supporting several GOST algorithms and public key formats. 1322238405Sjkim Kindly donated by Cryptocom. 1323238405Sjkim [Cryptocom] 1324238405Sjkim 1325238405Sjkim *) Partial support for Issuing Distribution Point CRL extension. CRLs 1326238405Sjkim partitioned by DP are handled but no indirect CRL or reason partitioning 1327238405Sjkim (yet). Complete overhaul of CRL handling: now the most suitable CRL is 1328238405Sjkim selected via a scoring technique which handles IDP and AKID in CRLs. 1329238405Sjkim [Steve Henson] 1330238405Sjkim 1331238405Sjkim *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which 1332238405Sjkim will ultimately be used for all verify operations: this will remove the 1333238405Sjkim X509_STORE dependency on certificate verification and allow alternative 1334238405Sjkim lookup methods. X509_STORE based implementations of these two callbacks. 1335238405Sjkim [Steve Henson] 1336238405Sjkim 1337238405Sjkim *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names. 1338238405Sjkim Modify get_crl() to find a valid (unexpired) CRL if possible. 1339238405Sjkim [Steve Henson] 1340238405Sjkim 1341238405Sjkim *) New function X509_CRL_match() to check if two CRLs are identical. Normally 1342238405Sjkim this would be called X509_CRL_cmp() but that name is already used by 1343238405Sjkim a function that just compares CRL issuer names. Cache several CRL 1344238405Sjkim extensions in X509_CRL structure and cache CRLDP in X509. 1345238405Sjkim [Steve Henson] 1346238405Sjkim 1347238405Sjkim *) Store a "canonical" representation of X509_NAME structure (ASN1 Name) 1348238405Sjkim this maps equivalent X509_NAME structures into a consistent structure. 1349238405Sjkim Name comparison can then be performed rapidly using memcmp(). 1350238405Sjkim [Steve Henson] 1351238405Sjkim 1352238405Sjkim *) Non-blocking OCSP request processing. Add -timeout option to ocsp 1353238405Sjkim utility. 1354238405Sjkim [Steve Henson] 1355238405Sjkim 1356238405Sjkim *) Allow digests to supply their own micalg string for S/MIME type using 1357238405Sjkim the ctrl EVP_MD_CTRL_MICALG. 1358238405Sjkim [Steve Henson] 1359238405Sjkim 1360238405Sjkim *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the 1361238405Sjkim EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN 1362238405Sjkim ctrl. It can then customise the structure before and/or after signing 1363238405Sjkim if necessary. 1364238405Sjkim [Steve Henson] 1365238405Sjkim 1366238405Sjkim *) New function OBJ_add_sigid() to allow application defined signature OIDs 1367238405Sjkim to be added to OpenSSLs internal tables. New function OBJ_sigid_free() 1368238405Sjkim to free up any added signature OIDs. 1369238405Sjkim [Steve Henson] 1370238405Sjkim 1371238405Sjkim *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(), 1372238405Sjkim EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal 1373238405Sjkim digest and cipher tables. New options added to openssl utility: 1374238405Sjkim list-message-digest-algorithms and list-cipher-algorithms. 1375238405Sjkim [Steve Henson] 1376238405Sjkim 1377238405Sjkim *) Change the array representation of binary polynomials: the list 1378238405Sjkim of degrees of non-zero coefficients is now terminated with -1. 1379238405Sjkim Previously it was terminated with 0, which was also part of the 1380238405Sjkim value; thus, the array representation was not applicable to 1381238405Sjkim polynomials where t^0 has coefficient zero. This change makes 1382238405Sjkim the array representation useful in a more general context. 1383238405Sjkim [Douglas Stebila] 1384238405Sjkim 1385238405Sjkim *) Various modifications and fixes to SSL/TLS cipher string 1386238405Sjkim handling. For ECC, the code now distinguishes between fixed ECDH 1387238405Sjkim with RSA certificates on the one hand and with ECDSA certificates 1388238405Sjkim on the other hand, since these are separate ciphersuites. The 1389238405Sjkim unused code for Fortezza ciphersuites has been removed. 1390238405Sjkim 1391238405Sjkim For consistency with EDH, ephemeral ECDH is now called "EECDH" 1392238405Sjkim (not "ECDHE"). For consistency with the code for DH 1393238405Sjkim certificates, use of ECDH certificates is now considered ECDH 1394238405Sjkim authentication, not RSA or ECDSA authentication (the latter is 1395238405Sjkim merely the CA's signing algorithm and not actively used in the 1396238405Sjkim protocol). 1397238405Sjkim 1398238405Sjkim The temporary ciphersuite alias "ECCdraft" is no longer 1399238405Sjkim available, and ECC ciphersuites are no longer excluded from "ALL" 1400238405Sjkim and "DEFAULT". The following aliases now exist for RFC 4492 1401238405Sjkim ciphersuites, most of these by analogy with the DH case: 1402238405Sjkim 1403238405Sjkim kECDHr - ECDH cert, signed with RSA 1404238405Sjkim kECDHe - ECDH cert, signed with ECDSA 1405238405Sjkim kECDH - ECDH cert (signed with either RSA or ECDSA) 1406238405Sjkim kEECDH - ephemeral ECDH 1407238405Sjkim ECDH - ECDH cert or ephemeral ECDH 1408238405Sjkim 1409238405Sjkim aECDH - ECDH cert 1410238405Sjkim aECDSA - ECDSA cert 1411238405Sjkim ECDSA - ECDSA cert 1412238405Sjkim 1413238405Sjkim AECDH - anonymous ECDH 1414238405Sjkim EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH") 1415238405Sjkim 1416238405Sjkim [Bodo Moeller] 1417238405Sjkim 1418238405Sjkim *) Add additional S/MIME capabilities for AES and GOST ciphers if supported. 1419238405Sjkim Use correct micalg parameters depending on digest(s) in signed message. 1420238405Sjkim [Steve Henson] 1421238405Sjkim 1422238405Sjkim *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process 1423238405Sjkim an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code. 1424238405Sjkim [Steve Henson] 1425238405Sjkim 1426238405Sjkim *) Initial engine support for EVP_PKEY_METHOD. New functions to permit 1427238405Sjkim an engine to register a method. Add ENGINE lookups for methods and 1428238405Sjkim functional reference processing. 1429238405Sjkim [Steve Henson] 1430238405Sjkim 1431238405Sjkim *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of 1432238405Sjkim EVP_{Sign,Verify}* which allow an application to customise the signature 1433238405Sjkim process. 1434238405Sjkim [Steve Henson] 1435238405Sjkim 1436238405Sjkim *) New -resign option to smime utility. This adds one or more signers 1437238405Sjkim to an existing PKCS#7 signedData structure. Also -md option to use an 1438238405Sjkim alternative message digest algorithm for signing. 1439238405Sjkim [Steve Henson] 1440238405Sjkim 1441238405Sjkim *) Tidy up PKCS#7 routines and add new functions to make it easier to 1442238405Sjkim create PKCS7 structures containing multiple signers. Update smime 1443238405Sjkim application to support multiple signers. 1444238405Sjkim [Steve Henson] 1445238405Sjkim 1446238405Sjkim *) New -macalg option to pkcs12 utility to allow setting of an alternative 1447238405Sjkim digest MAC. 1448238405Sjkim [Steve Henson] 1449238405Sjkim 1450238405Sjkim *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC. 1451238405Sjkim Reorganize PBE internals to lookup from a static table using NIDs, 1452238405Sjkim add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl: 1453238405Sjkim EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative 1454238405Sjkim PRF which will be automatically used with PBES2. 1455238405Sjkim [Steve Henson] 1456238405Sjkim 1457238405Sjkim *) Replace the algorithm specific calls to generate keys in "req" with the 1458238405Sjkim new API. 1459238405Sjkim [Steve Henson] 1460238405Sjkim 1461238405Sjkim *) Update PKCS#7 enveloped data routines to use new API. This is now 1462238405Sjkim supported by any public key method supporting the encrypt operation. A 1463238405Sjkim ctrl is added to allow the public key algorithm to examine or modify 1464238405Sjkim the PKCS#7 RecipientInfo structure if it needs to: for RSA this is 1465238405Sjkim a no op. 1466238405Sjkim [Steve Henson] 1467238405Sjkim 1468238405Sjkim *) Add a ctrl to asn1 method to allow a public key algorithm to express 1469238405Sjkim a default digest type to use. In most cases this will be SHA1 but some 1470238405Sjkim algorithms (such as GOST) need to specify an alternative digest. The 1471238405Sjkim return value indicates how strong the prefernce is 1 means optional and 1472238405Sjkim 2 is mandatory (that is it is the only supported type). Modify 1473238405Sjkim ASN1_item_sign() to accept a NULL digest argument to indicate it should 1474238405Sjkim use the default md. Update openssl utilities to use the default digest 1475238405Sjkim type for signing if it is not explicitly indicated. 1476238405Sjkim [Steve Henson] 1477238405Sjkim 1478238405Sjkim *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New 1479238405Sjkim EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant 1480238405Sjkim signing method from the key type. This effectively removes the link 1481238405Sjkim between digests and public key types. 1482238405Sjkim [Steve Henson] 1483238405Sjkim 1484238405Sjkim *) Add an OID cross reference table and utility functions. Its purpose is to 1485238405Sjkim translate between signature OIDs such as SHA1WithrsaEncryption and SHA1, 1486238405Sjkim rsaEncryption. This will allow some of the algorithm specific hackery 1487238405Sjkim needed to use the correct OID to be removed. 1488238405Sjkim [Steve Henson] 1489238405Sjkim 1490238405Sjkim *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO 1491238405Sjkim structures for PKCS7_sign(). They are now set up by the relevant public 1492238405Sjkim key ASN1 method. 1493238405Sjkim [Steve Henson] 1494238405Sjkim 1495238405Sjkim *) Add provisional EC pkey method with support for ECDSA and ECDH. 1496238405Sjkim [Steve Henson] 1497238405Sjkim 1498238405Sjkim *) Add support for key derivation (agreement) in the API, DH method and 1499238405Sjkim pkeyutl. 1500238405Sjkim [Steve Henson] 1501238405Sjkim 1502238405Sjkim *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support 1503238405Sjkim public and private key formats. As a side effect these add additional 1504238405Sjkim command line functionality not previously available: DSA signatures can be 1505238405Sjkim generated and verified using pkeyutl and DH key support and generation in 1506238405Sjkim pkey, genpkey. 1507238405Sjkim [Steve Henson] 1508238405Sjkim 1509238405Sjkim *) BeOS support. 1510238405Sjkim [Oliver Tappe <zooey@hirschkaefer.de>] 1511238405Sjkim 1512238405Sjkim *) New make target "install_html_docs" installs HTML renditions of the 1513238405Sjkim manual pages. 1514238405Sjkim [Oliver Tappe <zooey@hirschkaefer.de>] 1515238405Sjkim 1516238405Sjkim *) New utility "genpkey" this is analagous to "genrsa" etc except it can 1517238405Sjkim generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to 1518238405Sjkim support key and parameter generation and add initial key generation 1519238405Sjkim functionality for RSA. 1520238405Sjkim [Steve Henson] 1521238405Sjkim 1522238405Sjkim *) Add functions for main EVP_PKEY_method operations. The undocumented 1523238405Sjkim functions EVP_PKEY_{encrypt,decrypt} have been renamed to 1524238405Sjkim EVP_PKEY_{encrypt,decrypt}_old. 1525238405Sjkim [Steve Henson] 1526238405Sjkim 1527238405Sjkim *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public 1528238405Sjkim key API, doesn't do much yet. 1529238405Sjkim [Steve Henson] 1530238405Sjkim 1531238405Sjkim *) New function EVP_PKEY_asn1_get0_info() to retrieve information about 1532238405Sjkim public key algorithms. New option to openssl utility: 1533238405Sjkim "list-public-key-algorithms" to print out info. 1534238405Sjkim [Steve Henson] 1535238405Sjkim 1536238405Sjkim *) Implement the Supported Elliptic Curves Extension for 1537238405Sjkim ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1538238405Sjkim [Douglas Stebila] 1539238405Sjkim 1540238405Sjkim *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or 1541238405Sjkim EVP_CIPHER structures to avoid later problems in EVP_cleanup(). 1542238405Sjkim [Steve Henson] 1543238405Sjkim 1544238405Sjkim *) New utilities pkey and pkeyparam. These are similar to algorithm specific 1545238405Sjkim utilities such as rsa, dsa, dsaparam etc except they process any key 1546238405Sjkim type. 1547238405Sjkim [Steve Henson] 1548238405Sjkim 1549238405Sjkim *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New 1550238405Sjkim functions EVP_PKEY_print_public(), EVP_PKEY_print_private(), 1551238405Sjkim EVP_PKEY_print_param() to print public key data from an EVP_PKEY 1552238405Sjkim structure. 1553238405Sjkim [Steve Henson] 1554238405Sjkim 1555238405Sjkim *) Initial support for pluggable public key ASN1. 1556238405Sjkim De-spaghettify the public key ASN1 handling. Move public and private 1557238405Sjkim key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate 1558238405Sjkim algorithm specific handling to a single module within the relevant 1559238405Sjkim algorithm directory. Add functions to allow (near) opaque processing 1560238405Sjkim of public and private key structures. 1561238405Sjkim [Steve Henson] 1562238405Sjkim 1563238405Sjkim *) Implement the Supported Point Formats Extension for 1564238405Sjkim ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1565238405Sjkim [Douglas Stebila] 1566238405Sjkim 1567238405Sjkim *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members 1568238405Sjkim for the psk identity [hint] and the psk callback functions to the 1569238405Sjkim SSL_SESSION, SSL and SSL_CTX structure. 1570238405Sjkim 1571238405Sjkim New ciphersuites: 1572238405Sjkim PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA, 1573238405Sjkim PSK-AES256-CBC-SHA 1574238405Sjkim 1575238405Sjkim New functions: 1576238405Sjkim SSL_CTX_use_psk_identity_hint 1577238405Sjkim SSL_get_psk_identity_hint 1578238405Sjkim SSL_get_psk_identity 1579238405Sjkim SSL_use_psk_identity_hint 1580238405Sjkim 1581238405Sjkim [Mika Kousa and Pasi Eronen of Nokia Corporation] 1582238405Sjkim 1583238405Sjkim *) Add RFC 3161 compliant time stamp request creation, response generation 1584238405Sjkim and response verification functionality. 1585238405Sjkim [Zolt�n Gl�zik <zglozik@opentsa.org>, The OpenTSA Project] 1586238405Sjkim 1587238405Sjkim *) Add initial support for TLS extensions, specifically for the server_name 1588238405Sjkim extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 1589238405Sjkim have new members for a host name. The SSL data structure has an 1590238405Sjkim additional member SSL_CTX *initial_ctx so that new sessions can be 1591238405Sjkim stored in that context to allow for session resumption, even after the 1592238405Sjkim SSL has been switched to a new SSL_CTX in reaction to a client's 1593238405Sjkim server_name extension. 1594238405Sjkim 1595238405Sjkim New functions (subject to change): 1596238405Sjkim 1597238405Sjkim SSL_get_servername() 1598238405Sjkim SSL_get_servername_type() 1599238405Sjkim SSL_set_SSL_CTX() 1600238405Sjkim 1601238405Sjkim New CTRL codes and macros (subject to change): 1602238405Sjkim 1603238405Sjkim SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 1604238405Sjkim - SSL_CTX_set_tlsext_servername_callback() 1605238405Sjkim SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 1606238405Sjkim - SSL_CTX_set_tlsext_servername_arg() 1607238405Sjkim SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 1608238405Sjkim 1609238405Sjkim openssl s_client has a new '-servername ...' option. 1610238405Sjkim 1611238405Sjkim openssl s_server has new options '-servername_host ...', '-cert2 ...', 1612238405Sjkim '-key2 ...', '-servername_fatal' (subject to change). This allows 1613238405Sjkim testing the HostName extension for a specific single host name ('-cert' 1614238405Sjkim and '-key' remain fallbacks for handshakes without HostName 1615238405Sjkim negotiation). If the unrecogninzed_name alert has to be sent, this by 1616238405Sjkim default is a warning; it becomes fatal with the '-servername_fatal' 1617238405Sjkim option. 1618238405Sjkim 1619238405Sjkim [Peter Sylvester, Remy Allais, Christophe Renou] 1620238405Sjkim 1621238405Sjkim *) Whirlpool hash implementation is added. 1622238405Sjkim [Andy Polyakov] 1623238405Sjkim 1624238405Sjkim *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to 1625238405Sjkim bn(64,32). Because of instruction set limitations it doesn't have 1626238405Sjkim any negative impact on performance. This was done mostly in order 1627238405Sjkim to make it possible to share assembler modules, such as bn_mul_mont 1628238405Sjkim implementations, between 32- and 64-bit builds without hassle. 1629238405Sjkim [Andy Polyakov] 1630238405Sjkim 1631238405Sjkim *) Move code previously exiled into file crypto/ec/ec2_smpt.c 1632238405Sjkim to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP 1633238405Sjkim macro. 1634238405Sjkim [Bodo Moeller] 1635238405Sjkim 1636238405Sjkim *) New candidate for BIGNUM assembler implementation, bn_mul_mont, 1637238405Sjkim dedicated Montgomery multiplication procedure, is introduced. 1638238405Sjkim BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher 1639238405Sjkim "64-bit" performance on certain 32-bit targets. 1640238405Sjkim [Andy Polyakov] 1641238405Sjkim 1642238405Sjkim *) New option SSL_OP_NO_COMP to disable use of compression selectively 1643238405Sjkim in SSL structures. New SSL ctrl to set maximum send fragment size. 1644238405Sjkim Save memory by seeting the I/O buffer sizes dynamically instead of 1645238405Sjkim using the maximum available value. 1646238405Sjkim [Steve Henson] 1647238405Sjkim 1648238405Sjkim *) New option -V for 'openssl ciphers'. This prints the ciphersuite code 1649238405Sjkim in addition to the text details. 1650238405Sjkim [Bodo Moeller] 1651238405Sjkim 1652238405Sjkim *) Very, very preliminary EXPERIMENTAL support for printing of general 1653238405Sjkim ASN1 structures. This currently produces rather ugly output and doesn't 1654238405Sjkim handle several customised structures at all. 1655238405Sjkim [Steve Henson] 1656238405Sjkim 1657238405Sjkim *) Integrated support for PVK file format and some related formats such 1658238405Sjkim as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support 1659238405Sjkim these in the 'rsa' and 'dsa' utilities. 1660238405Sjkim [Steve Henson] 1661238405Sjkim 1662238405Sjkim *) Support for PKCS#1 RSAPublicKey format on rsa utility command line. 1663238405Sjkim [Steve Henson] 1664238405Sjkim 1665238405Sjkim *) Remove the ancient ASN1_METHOD code. This was only ever used in one 1666238405Sjkim place for the (very old) "NETSCAPE" format certificates which are now 1667238405Sjkim handled using new ASN1 code equivalents. 1668238405Sjkim [Steve Henson] 1669238405Sjkim 1670238405Sjkim *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD 1671238405Sjkim pointer and make the SSL_METHOD parameter in SSL_CTX_new, 1672238405Sjkim SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'. 1673238405Sjkim [Nils Larsch] 1674238405Sjkim 1675238405Sjkim *) Modify CRL distribution points extension code to print out previously 1676238405Sjkim unsupported fields. Enhance extension setting code to allow setting of 1677238405Sjkim all fields. 1678238405Sjkim [Steve Henson] 1679238405Sjkim 1680238405Sjkim *) Add print and set support for Issuing Distribution Point CRL extension. 1681238405Sjkim [Steve Henson] 1682238405Sjkim 1683238405Sjkim *) Change 'Configure' script to enable Camellia by default. 1684238405Sjkim [NTT] 1685238405Sjkim 1686264331Sjkim Changes between 0.9.8x and 0.9.8y [5 Feb 2013] 1687264331Sjkim 1688264331Sjkim *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 1689264331Sjkim 1690264331Sjkim This addresses the flaw in CBC record processing discovered by 1691264331Sjkim Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 1692264331Sjkim at: http://www.isg.rhul.ac.uk/tls/ 1693264331Sjkim 1694264331Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 1695264331Sjkim Security Group at Royal Holloway, University of London 1696264331Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 1697264331Sjkim Emilia K�sper for the initial patch. 1698264331Sjkim (CVE-2013-0169) 1699264331Sjkim [Emilia K�sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 1700264331Sjkim 1701264331Sjkim *) Return an error when checking OCSP signatures when key is NULL. 1702264331Sjkim This fixes a DoS attack. (CVE-2013-0166) 1703264331Sjkim [Steve Henson] 1704264331Sjkim 1705264331Sjkim *) Call OCSP Stapling callback after ciphersuite has been chosen, so 1706264331Sjkim the right response is stapled. Also change SSL_get_certificate() 1707264331Sjkim so it returns the certificate actually sent. 1708264331Sjkim See http://rt.openssl.org/Ticket/Display.html?id=2836. 1709264331Sjkim (This is a backport) 1710264331Sjkim [Rob Stradling <rob.stradling@comodo.com>] 1711264331Sjkim 1712264331Sjkim *) Fix possible deadlock when decoding public keys. 1713264331Sjkim [Steve Henson] 1714264331Sjkim 1715264331Sjkim Changes between 0.9.8w and 0.9.8x [10 May 2012] 1716264331Sjkim 1717264331Sjkim *) Sanity check record length before skipping explicit IV in DTLS 1718264331Sjkim to fix DoS attack. 1719264331Sjkim 1720264331Sjkim Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 1721264331Sjkim fuzzing as a service testing platform. 1722264331Sjkim (CVE-2012-2333) 1723264331Sjkim [Steve Henson] 1724264331Sjkim 1725264331Sjkim *) Initialise tkeylen properly when encrypting CMS messages. 1726264331Sjkim Thanks to Solar Designer of Openwall for reporting this issue. 1727264331Sjkim [Steve Henson] 1728264331Sjkim 1729264331Sjkim Changes between 0.9.8v and 0.9.8w [23 Apr 2012] 1730264331Sjkim 1731264331Sjkim *) The fix for CVE-2012-2110 did not take into account that the 1732264331Sjkim 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an 1733264331Sjkim int in OpenSSL 0.9.8, making it still vulnerable. Fix by 1734264331Sjkim rejecting negative len parameter. (CVE-2012-2131) 1735264331Sjkim [Tomas Hoger <thoger@redhat.com>] 1736264331Sjkim 1737264331Sjkim Changes between 0.9.8u and 0.9.8v [19 Apr 2012] 1738264331Sjkim 1739264331Sjkim *) Check for potentially exploitable overflows in asn1_d2i_read_bio 1740264331Sjkim BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 1741264331Sjkim in CRYPTO_realloc_clean. 1742264331Sjkim 1743264331Sjkim Thanks to Tavis Ormandy, Google Security Team, for discovering this 1744264331Sjkim issue and to Adam Langley <agl@chromium.org> for fixing it. 1745264331Sjkim (CVE-2012-2110) 1746264331Sjkim [Adam Langley (Google), Tavis Ormandy, Google Security Team] 1747264331Sjkim 1748264331Sjkim Changes between 0.9.8t and 0.9.8u [12 Mar 2012] 1749264331Sjkim 1750264331Sjkim *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 1751264331Sjkim in CMS and PKCS7 code. When RSA decryption fails use a random key for 1752264331Sjkim content decryption and always return the same error. Note: this attack 1753264331Sjkim needs on average 2^20 messages so it only affects automated senders. The 1754264331Sjkim old behaviour can be reenabled in the CMS code by setting the 1755264331Sjkim CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 1756264331Sjkim an MMA defence is not necessary. 1757264331Sjkim Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 1758264331Sjkim this issue. (CVE-2012-0884) 1759264331Sjkim [Steve Henson] 1760264331Sjkim 1761264331Sjkim *) Fix CVE-2011-4619: make sure we really are receiving a 1762264331Sjkim client hello before rejecting multiple SGC restarts. Thanks to 1763264331Sjkim Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 1764264331Sjkim [Steve Henson] 1765264331Sjkim 1766237657Sjkim Changes between 0.9.8s and 0.9.8t [18 Jan 2012] 1767237657Sjkim 1768237657Sjkim *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 1769237657Sjkim Thanks to Antonio Martin, Enterprise Secure Access Research and 1770237657Sjkim Development, Cisco Systems, Inc. for discovering this bug and 1771237657Sjkim preparing a fix. (CVE-2012-0050) 1772237657Sjkim [Antonio Martin] 1773264331Sjkim 1774237657Sjkim Changes between 0.9.8r and 0.9.8s [4 Jan 2012] 1775237657Sjkim 1776237657Sjkim *) Nadhem Alfardan and Kenny Paterson have discovered an extension 1777237657Sjkim of the Vaudenay padding oracle attack on CBC mode encryption 1778237657Sjkim which enables an efficient plaintext recovery attack against 1779237657Sjkim the OpenSSL implementation of DTLS. Their attack exploits timing 1780237657Sjkim differences arising during decryption processing. A research 1781237657Sjkim paper describing this attack can be found at: 1782237657Sjkim http://www.isg.rhul.ac.uk/~kp/dtls.pdf 1783237657Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 1784237657Sjkim Security Group at Royal Holloway, University of London 1785237657Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 1786237657Sjkim <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 1787237657Sjkim for preparing the fix. (CVE-2011-4108) 1788237657Sjkim [Robin Seggelmann, Michael Tuexen] 1789237657Sjkim 1790237657Sjkim *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109) 1791237657Sjkim [Ben Laurie, Kasper <ekasper@google.com>] 1792237657Sjkim 1793237657Sjkim *) Clear bytes used for block padding of SSL 3.0 records. 1794237657Sjkim (CVE-2011-4576) 1795237657Sjkim [Adam Langley (Google)] 1796237657Sjkim 1797237657Sjkim *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 1798237657Sjkim Kadianakis <desnacked@gmail.com> for discovering this issue and 1799237657Sjkim Adam Langley for preparing the fix. (CVE-2011-4619) 1800237657Sjkim [Adam Langley (Google)] 1801237657Sjkim 1802237657Sjkim *) Prevent malformed RFC3779 data triggering an assertion failure. 1803237657Sjkim Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 1804237657Sjkim and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 1805237657Sjkim [Rob Austein <sra@hactrn.net>] 1806237657Sjkim 1807237657Sjkim *) Fix ssl_ciph.c set-up race. 1808237657Sjkim [Adam Langley (Google)] 1809237657Sjkim 1810237657Sjkim *) Fix spurious failures in ecdsatest.c. 1811237657Sjkim [Emilia K�sper (Google)] 1812237657Sjkim 1813237657Sjkim *) Fix the BIO_f_buffer() implementation (which was mixing different 1814237657Sjkim interpretations of the '..._len' fields). 1815237657Sjkim [Adam Langley (Google)] 1816237657Sjkim 1817237657Sjkim *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 1818237657Sjkim BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 1819237657Sjkim threads won't reuse the same blinding coefficients. 1820237657Sjkim 1821237657Sjkim This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 1822237657Sjkim lock to call BN_BLINDING_invert_ex, and avoids one use of 1823237657Sjkim BN_BLINDING_update for each BN_BLINDING structure (previously, 1824237657Sjkim the last update always remained unused). 1825237657Sjkim [Emilia K�sper (Google)] 1826237657Sjkim 1827237657Sjkim *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 1828237657Sjkim for multi-threaded use of ECDH. 1829237657Sjkim [Adam Langley (Google)] 1830237657Sjkim 1831237657Sjkim *) Fix x509_name_ex_d2i memory leak on bad inputs. 1832237657Sjkim [Bodo Moeller] 1833237657Sjkim 1834237657Sjkim *) Add protection against ECDSA timing attacks as mentioned in the paper 1835237657Sjkim by Billy Bob Brumley and Nicola Tuveri, see: 1836237657Sjkim 1837237657Sjkim http://eprint.iacr.org/2011/232.pdf 1838237657Sjkim 1839237657Sjkim [Billy Bob Brumley and Nicola Tuveri] 1840237657Sjkim 1841237657Sjkim Changes between 0.9.8q and 0.9.8r [8 Feb 2011] 1842237657Sjkim 1843237657Sjkim *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 1844237657Sjkim [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 1845237657Sjkim 1846237657Sjkim *) Fix bug in string printing code: if *any* escaping is enabled we must 1847237657Sjkim escape the escape character (backslash) or the resulting string is 1848237657Sjkim ambiguous. 1849237657Sjkim [Steve Henson] 1850237657Sjkim 1851216166Ssimon Changes between 0.9.8p and 0.9.8q [2 Dec 2010] 1852216166Ssimon 1853216166Ssimon *) Disable code workaround for ancient and obsolete Netscape browsers 1854216166Ssimon and servers: an attacker can use it in a ciphersuite downgrade attack. 1855216166Ssimon Thanks to Martin Rex for discovering this bug. CVE-2010-4180 1856216166Ssimon [Steve Henson] 1857216166Ssimon 1858216166Ssimon *) Fixed J-PAKE implementation error, originally discovered by 1859216166Ssimon Sebastien Martini, further info and confirmation from Stefan 1860216166Ssimon Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 1861216166Ssimon [Ben Laurie] 1862216166Ssimon 1863215697Ssimon Changes between 0.9.8o and 0.9.8p [16 Nov 2010] 1864215697Ssimon 1865215697Ssimon *) Fix extension code to avoid race conditions which can result in a buffer 1866215697Ssimon overrun vulnerability: resumed sessions must not be modified as they can 1867215697Ssimon be shared by multiple threads. CVE-2010-3864 1868215697Ssimon [Steve Henson] 1869215697Ssimon 1870215697Ssimon *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939 1871215697Ssimon [Steve Henson] 1872215697Ssimon 1873215697Ssimon *) Don't reencode certificate when calculating signature: cache and use 1874215697Ssimon the original encoding instead. This makes signature verification of 1875215697Ssimon some broken encodings work correctly. 1876215697Ssimon [Steve Henson] 1877215697Ssimon 1878215697Ssimon *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT 1879215697Ssimon is also one of the inputs. 1880215697Ssimon [Emilia K�sper <emilia.kasper@esat.kuleuven.be> (Google)] 1881215697Ssimon 1882215697Ssimon *) Don't repeatedly append PBE algorithms to table if they already exist. 1883215697Ssimon Sort table on each new add. This effectively makes the table read only 1884215697Ssimon after all algorithms are added and subsequent calls to PKCS12_pbe_add 1885215697Ssimon etc are non-op. 1886215697Ssimon [Steve Henson] 1887215697Ssimon 1888215697Ssimon Changes between 0.9.8n and 0.9.8o [01 Jun 2010] 1889215697Ssimon 1890215697Ssimon [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after 1891215697Ssimon OpenSSL 1.0.0.] 1892215697Ssimon 1893215697Ssimon *) Correct a typo in the CMS ASN1 module which can result in invalid memory 1894215697Ssimon access or freeing data twice (CVE-2010-0742) 1895215697Ssimon [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>] 1896215697Ssimon 1897215697Ssimon *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more 1898215697Ssimon common in certificates and some applications which only call 1899215697Ssimon SSL_library_init and not OpenSSL_add_all_algorithms() will fail. 1900215697Ssimon [Steve Henson] 1901215697Ssimon 1902215697Ssimon *) VMS fixes: 1903215697Ssimon Reduce copying into .apps and .test in makevms.com 1904215697Ssimon Don't try to use blank CA certificate in CA.com 1905215697Ssimon Allow use of C files from original directories in maketests.com 1906215697Ssimon [Steven M. Schweda" <sms@antinode.info>] 1907215697Ssimon 1908206046Ssimon Changes between 0.9.8m and 0.9.8n [24 Mar 2010] 1909206046Ssimon 1910206046Ssimon *) When rejecting SSL/TLS records due to an incorrect version number, never 1911206046Ssimon update s->server with a new major version number. As of 1912206046Ssimon - OpenSSL 0.9.8m if 'short' is a 16-bit type, 1913206046Ssimon - OpenSSL 0.9.8f if 'short' is longer than 16 bits, 1914206046Ssimon the previous behavior could result in a read attempt at NULL when 1915206046Ssimon receiving specific incorrect SSL/TLS records once record payload 1916206046Ssimon protection is active. (CVE-2010-0740) 1917206046Ssimon [Bodo Moeller, Adam Langley <agl@chromium.org>] 1918206046Ssimon 1919206046Ssimon *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL 1920206046Ssimon could be crashed if the relevant tables were not present (e.g. chrooted). 1921206046Ssimon [Tomas Hoger <thoger@redhat.com>] 1922206046Ssimon 1923205128Ssimon Changes between 0.9.8l and 0.9.8m [25 Feb 2010] 1924205128Ssimon 1925205128Ssimon *) Always check bn_wexpend() return values for failure. (CVE-2009-3245) 1926205128Ssimon [Martin Olsson, Neel Mehta] 1927205128Ssimon 1928205128Ssimon *) Fix X509_STORE locking: Every 'objs' access requires a lock (to 1929205128Ssimon accommodate for stack sorting, always a write lock!). 1930205128Ssimon [Bodo Moeller] 1931205128Ssimon 1932205128Ssimon *) On some versions of WIN32 Heap32Next is very slow. This can cause 1933205128Ssimon excessive delays in the RAND_poll(): over a minute. As a workaround 1934205128Ssimon include a time check in the inner Heap32Next loop too. 1935205128Ssimon [Steve Henson] 1936205128Ssimon 1937205128Ssimon *) The code that handled flushing of data in SSL/TLS originally used the 1938205128Ssimon BIO_CTRL_INFO ctrl to see if any data was pending first. This caused 1939205128Ssimon the problem outlined in PR#1949. The fix suggested there however can 1940205128Ssimon trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions 1941205128Ssimon of Apache). So instead simplify the code to flush unconditionally. 1942205128Ssimon This should be fine since flushing with no data to flush is a no op. 1943205128Ssimon [Steve Henson] 1944205128Ssimon 1945205128Ssimon *) Handle TLS versions 2.0 and later properly and correctly use the 1946205128Ssimon highest version of TLS/SSL supported. Although TLS >= 2.0 is some way 1947205128Ssimon off ancient servers have a habit of sticking around for a while... 1948205128Ssimon [Steve Henson] 1949205128Ssimon 1950205128Ssimon *) Modify compression code so it frees up structures without using the 1951205128Ssimon ex_data callbacks. This works around a problem where some applications 1952205128Ssimon call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when 1953205128Ssimon restarting) then use compression (e.g. SSL with compression) later. 1954205128Ssimon This results in significant per-connection memory leaks and 1955205128Ssimon has caused some security issues including CVE-2008-1678 and 1956205128Ssimon CVE-2009-4355. 1957205128Ssimon [Steve Henson] 1958205128Ssimon 1959205128Ssimon *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't 1960205128Ssimon change when encrypting or decrypting. 1961205128Ssimon [Bodo Moeller] 1962205128Ssimon 1963205128Ssimon *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to 1964205128Ssimon connect and renegotiate with servers which do not support RI. 1965205128Ssimon Until RI is more widely deployed this option is enabled by default. 1966205128Ssimon [Steve Henson] 1967205128Ssimon 1968205128Ssimon *) Add "missing" ssl ctrls to clear options and mode. 1969205128Ssimon [Steve Henson] 1970205128Ssimon 1971205128Ssimon *) If client attempts to renegotiate and doesn't support RI respond with 1972205128Ssimon a no_renegotiation alert as required by RFC5746. Some renegotiating 1973205128Ssimon TLS clients will continue a connection gracefully when they receive 1974205128Ssimon the alert. Unfortunately OpenSSL mishandled this alert and would hang 1975205128Ssimon waiting for a server hello which it will never receive. Now we treat a 1976205128Ssimon received no_renegotiation alert as a fatal error. This is because 1977205128Ssimon applications requesting a renegotiation might well expect it to succeed 1978205128Ssimon and would have no code in place to handle the server denying it so the 1979205128Ssimon only safe thing to do is to terminate the connection. 1980205128Ssimon [Steve Henson] 1981205128Ssimon 1982205128Ssimon *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if 1983205128Ssimon peer supports secure renegotiation and 0 otherwise. Print out peer 1984205128Ssimon renegotiation support in s_client/s_server. 1985205128Ssimon [Steve Henson] 1986205128Ssimon 1987205128Ssimon *) Replace the highly broken and deprecated SPKAC certification method with 1988205128Ssimon the updated NID creation version. This should correctly handle UTF8. 1989205128Ssimon [Steve Henson] 1990205128Ssimon 1991205128Ssimon *) Implement RFC5746. Re-enable renegotiation but require the extension 1992205128Ssimon as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION 1993205128Ssimon turns out to be a bad idea. It has been replaced by 1994205128Ssimon SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with 1995205128Ssimon SSL_CTX_set_options(). This is really not recommended unless you 1996205128Ssimon know what you are doing. 1997205128Ssimon [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson] 1998205128Ssimon 1999205128Ssimon *) Fixes to stateless session resumption handling. Use initial_ctx when 2000205128Ssimon issuing and attempting to decrypt tickets in case it has changed during 2001205128Ssimon servername handling. Use a non-zero length session ID when attempting 2002205128Ssimon stateless session resumption: this makes it possible to determine if 2003205128Ssimon a resumption has occurred immediately after receiving server hello 2004205128Ssimon (several places in OpenSSL subtly assume this) instead of later in 2005205128Ssimon the handshake. 2006205128Ssimon [Steve Henson] 2007205128Ssimon 2008205128Ssimon *) The functions ENGINE_ctrl(), OPENSSL_isservice(), 2009205128Ssimon CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error 2010205128Ssimon fixes for a few places where the return code is not checked 2011205128Ssimon correctly. 2012205128Ssimon [Julia Lawall <julia@diku.dk>] 2013205128Ssimon 2014205128Ssimon *) Add --strict-warnings option to Configure script to include devteam 2015205128Ssimon warnings in other configurations. 2016205128Ssimon [Steve Henson] 2017205128Ssimon 2018205128Ssimon *) Add support for --libdir option and LIBDIR variable in makefiles. This 2019205128Ssimon makes it possible to install openssl libraries in locations which 2020205128Ssimon have names other than "lib", for example "/usr/lib64" which some 2021205128Ssimon systems need. 2022205128Ssimon [Steve Henson, based on patch from Jeremy Utley] 2023205128Ssimon 2024205128Ssimon *) Don't allow the use of leading 0x80 in OIDs. This is a violation of 2025205128Ssimon X690 8.9.12 and can produce some misleading textual output of OIDs. 2026205128Ssimon [Steve Henson, reported by Dan Kaminsky] 2027205128Ssimon 2028205128Ssimon *) Delete MD2 from algorithm tables. This follows the recommendation in 2029205128Ssimon several standards that it is not used in new applications due to 2030205128Ssimon several cryptographic weaknesses. For binary compatibility reasons 2031205128Ssimon the MD2 API is still compiled in by default. 2032205128Ssimon [Steve Henson] 2033205128Ssimon 2034205128Ssimon *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved 2035205128Ssimon and restored. 2036205128Ssimon [Steve Henson] 2037205128Ssimon 2038205128Ssimon *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and 2039205128Ssimon OPENSSL_asc2uni conditionally on Netware platforms to avoid a name 2040205128Ssimon clash. 2041205128Ssimon [Guenter <lists@gknw.net>] 2042205128Ssimon 2043205128Ssimon *) Fix the server certificate chain building code to use X509_verify_cert(), 2044205128Ssimon it used to have an ad-hoc builder which was unable to cope with anything 2045205128Ssimon other than a simple chain. 2046205128Ssimon [David Woodhouse <dwmw2@infradead.org>, Steve Henson] 2047205128Ssimon 2048205128Ssimon *) Don't check self signed certificate signatures in X509_verify_cert() 2049205128Ssimon by default (a flag can override this): it just wastes time without 2050205128Ssimon adding any security. As a useful side effect self signed root CAs 2051205128Ssimon with non-FIPS digests are now usable in FIPS mode. 2052205128Ssimon [Steve Henson] 2053205128Ssimon 2054205128Ssimon *) In dtls1_process_out_of_seq_message() the check if the current message 2055205128Ssimon is already buffered was missing. For every new message was memory 2056205128Ssimon allocated, allowing an attacker to perform an denial of service attack 2057205128Ssimon with sending out of seq handshake messages until there is no memory 2058205128Ssimon left. Additionally every future messege was buffered, even if the 2059205128Ssimon sequence number made no sense and would be part of another handshake. 2060205128Ssimon So only messages with sequence numbers less than 10 in advance will be 2061205128Ssimon buffered. (CVE-2009-1378) 2062205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 2063205128Ssimon 2064205128Ssimon *) Records are buffered if they arrive with a future epoch to be 2065205128Ssimon processed after finishing the corresponding handshake. There is 2066205128Ssimon currently no limitation to this buffer allowing an attacker to perform 2067205128Ssimon a DOS attack with sending records with future epochs until there is no 2068205128Ssimon memory left. This patch adds the pqueue_size() function to detemine 2069205128Ssimon the size of a buffer and limits the record buffer to 100 entries. 2070205128Ssimon (CVE-2009-1377) 2071205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 2072205128Ssimon 2073205128Ssimon *) Keep a copy of frag->msg_header.frag_len so it can be used after the 2074205128Ssimon parent structure is freed. (CVE-2009-1379) 2075205128Ssimon [Daniel Mentz] 2076205128Ssimon 2077205128Ssimon *) Handle non-blocking I/O properly in SSL_shutdown() call. 2078205128Ssimon [Darryl Miles <darryl-mailinglists@netbauds.net>] 2079205128Ssimon 2080205128Ssimon *) Add 2.5.4.* OIDs 2081205128Ssimon [Ilya O. <vrghost@gmail.com>] 2082205128Ssimon 2083205128Ssimon Changes between 0.9.8k and 0.9.8l [5 Nov 2009] 2084205128Ssimon 2085205128Ssimon *) Disable renegotiation completely - this fixes a severe security 2086205128Ssimon problem (CVE-2009-3555) at the cost of breaking all 2087205128Ssimon renegotiation. Renegotiation can be re-enabled by setting 2088205128Ssimon SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at 2089205128Ssimon run-time. This is really not recommended unless you know what 2090205128Ssimon you're doing. 2091205128Ssimon [Ben Laurie] 2092205128Ssimon 2093194206Ssimon Changes between 0.9.8j and 0.9.8k [25 Mar 2009] 2094167612Ssimon 2095194206Ssimon *) Don't set val to NULL when freeing up structures, it is freed up by 2096194206Ssimon underlying code. If sizeof(void *) > sizeof(long) this can result in 2097194206Ssimon zeroing past the valid field. (CVE-2009-0789) 2098194206Ssimon [Paolo Ganci <Paolo.Ganci@AdNovum.CH>] 2099194206Ssimon 2100194206Ssimon *) Fix bug where return value of CMS_SignerInfo_verify_content() was not 2101194206Ssimon checked correctly. This would allow some invalid signed attributes to 2102194206Ssimon appear to verify correctly. (CVE-2009-0591) 2103194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 2104194206Ssimon 2105194206Ssimon *) Reject UniversalString and BMPString types with invalid lengths. This 2106194206Ssimon prevents a crash in ASN1_STRING_print_ex() which assumes the strings have 2107194206Ssimon a legal length. (CVE-2009-0590) 2108194206Ssimon [Steve Henson] 2109194206Ssimon 2110194206Ssimon *) Set S/MIME signing as the default purpose rather than setting it 2111194206Ssimon unconditionally. This allows applications to override it at the store 2112194206Ssimon level. 2113194206Ssimon [Steve Henson] 2114194206Ssimon 2115194206Ssimon *) Permit restricted recursion of ASN1 strings. This is needed in practice 2116194206Ssimon to handle some structures. 2117194206Ssimon [Steve Henson] 2118194206Ssimon 2119194206Ssimon *) Improve efficiency of mem_gets: don't search whole buffer each time 2120194206Ssimon for a '\n' 2121194206Ssimon [Jeremy Shapiro <jnshapir@us.ibm.com>] 2122194206Ssimon 2123194206Ssimon *) New -hex option for openssl rand. 2124194206Ssimon [Matthieu Herrb] 2125194206Ssimon 2126194206Ssimon *) Print out UTF8String and NumericString when parsing ASN1. 2127194206Ssimon [Steve Henson] 2128194206Ssimon 2129194206Ssimon *) Support NumericString type for name components. 2130194206Ssimon [Steve Henson] 2131194206Ssimon 2132194206Ssimon *) Allow CC in the environment to override the automatically chosen 2133194206Ssimon compiler. Note that nothing is done to ensure flags work with the 2134194206Ssimon chosen compiler. 2135194206Ssimon [Ben Laurie] 2136194206Ssimon 2137194206Ssimon Changes between 0.9.8i and 0.9.8j [07 Jan 2009] 2138194206Ssimon 2139194206Ssimon *) Properly check EVP_VerifyFinal() and similar return values 2140194206Ssimon (CVE-2008-5077). 2141194206Ssimon [Ben Laurie, Bodo Moeller, Google Security Team] 2142194206Ssimon 2143194206Ssimon *) Enable TLS extensions by default. 2144194206Ssimon [Ben Laurie] 2145194206Ssimon 2146194206Ssimon *) Allow the CHIL engine to be loaded, whether the application is 2147194206Ssimon multithreaded or not. (This does not release the developer from the 2148194206Ssimon obligation to set up the dynamic locking callbacks.) 2149194206Ssimon [Sander Temme <sander@temme.net>] 2150194206Ssimon 2151194206Ssimon *) Use correct exit code if there is an error in dgst command. 2152194206Ssimon [Steve Henson; problem pointed out by Roland Dirlewanger] 2153194206Ssimon 2154194206Ssimon *) Tweak Configure so that you need to say "experimental-jpake" to enable 2155194206Ssimon JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. 2156194206Ssimon [Bodo Moeller] 2157194206Ssimon 2158194206Ssimon *) Add experimental JPAKE support, including demo authentication in 2159194206Ssimon s_client and s_server. 2160194206Ssimon [Ben Laurie] 2161194206Ssimon 2162194206Ssimon *) Set the comparison function in v3_addr_canonize(). 2163194206Ssimon [Rob Austein <sra@hactrn.net>] 2164194206Ssimon 2165194206Ssimon *) Add support for XMPP STARTTLS in s_client. 2166194206Ssimon [Philip Paeps <philip@freebsd.org>] 2167194206Ssimon 2168194206Ssimon *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior 2169194206Ssimon to ensure that even with this option, only ciphersuites in the 2170194206Ssimon server's preference list will be accepted. (Note that the option 2171194206Ssimon applies only when resuming a session, so the earlier behavior was 2172194206Ssimon just about the algorithm choice for symmetric cryptography.) 2173194206Ssimon [Bodo Moeller] 2174194206Ssimon 2175194206Ssimon Changes between 0.9.8h and 0.9.8i [15 Sep 2008] 2176194206Ssimon 2177205128Ssimon *) Fix NULL pointer dereference if a DTLS server received 2178205128Ssimon ChangeCipherSpec as first record (CVE-2009-1386). 2179205128Ssimon [PR #1679] 2180205128Ssimon 2181194206Ssimon *) Fix a state transitition in s3_srvr.c and d1_srvr.c 2182194206Ssimon (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...). 2183194206Ssimon [Nagendra Modadugu] 2184194206Ssimon 2185194206Ssimon *) The fix in 0.9.8c that supposedly got rid of unsafe 2186194206Ssimon double-checked locking was incomplete for RSA blinding, 2187194206Ssimon addressing just one layer of what turns out to have been 2188194206Ssimon doubly unsafe triple-checked locking. 2189194206Ssimon 2190194206Ssimon So now fix this for real by retiring the MONT_HELPER macro 2191194206Ssimon in crypto/rsa/rsa_eay.c. 2192194206Ssimon 2193194206Ssimon [Bodo Moeller; problem pointed out by Marius Schilder] 2194194206Ssimon 2195194206Ssimon *) Various precautionary measures: 2196194206Ssimon 2197194206Ssimon - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). 2198194206Ssimon 2199194206Ssimon - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). 2200194206Ssimon (NB: This would require knowledge of the secret session ticket key 2201194206Ssimon to exploit, in which case you'd be SOL either way.) 2202194206Ssimon 2203194206Ssimon - Change bn_nist.c so that it will properly handle input BIGNUMs 2204194206Ssimon outside the expected range. 2205194206Ssimon 2206194206Ssimon - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG 2207194206Ssimon builds. 2208194206Ssimon 2209194206Ssimon [Neel Mehta, Bodo Moeller] 2210194206Ssimon 2211194206Ssimon *) Allow engines to be "soft loaded" - i.e. optionally don't die if 2212194206Ssimon the load fails. Useful for distros. 2213194206Ssimon [Ben Laurie and the FreeBSD team] 2214194206Ssimon 2215194206Ssimon *) Add support for Local Machine Keyset attribute in PKCS#12 files. 2216194206Ssimon [Steve Henson] 2217194206Ssimon 2218194206Ssimon *) Fix BN_GF2m_mod_arr() top-bit cleanup code. 2219194206Ssimon [Huang Ying] 2220194206Ssimon 2221194206Ssimon *) Expand ENGINE to support engine supplied SSL client certificate functions. 2222194206Ssimon 2223194206Ssimon This work was sponsored by Logica. 2224194206Ssimon [Steve Henson] 2225194206Ssimon 2226194206Ssimon *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows 2227194206Ssimon keystores. Support for SSL/TLS client authentication too. 2228194206Ssimon Not compiled unless enable-capieng specified to Configure. 2229194206Ssimon 2230194206Ssimon This work was sponsored by Logica. 2231194206Ssimon [Steve Henson] 2232194206Ssimon 2233194206Ssimon *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using 2234194206Ssimon ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain 2235194206Ssimon attribute creation routines such as certifcate requests and PKCS#12 2236194206Ssimon files. 2237194206Ssimon [Steve Henson] 2238194206Ssimon 2239194206Ssimon Changes between 0.9.8g and 0.9.8h [28 May 2008] 2240194206Ssimon 2241194206Ssimon *) Fix flaw if 'Server Key exchange message' is omitted from a TLS 2242194206Ssimon handshake which could lead to a cilent crash as found using the 2243194206Ssimon Codenomicon TLS test suite (CVE-2008-1672) 2244194206Ssimon [Steve Henson, Mark Cox] 2245194206Ssimon 2246194206Ssimon *) Fix double free in TLS server name extensions which could lead to 2247194206Ssimon a remote crash found by Codenomicon TLS test suite (CVE-2008-0891) 2248194206Ssimon [Joe Orton] 2249194206Ssimon 2250194206Ssimon *) Clear error queue in SSL_CTX_use_certificate_chain_file() 2251194206Ssimon 2252194206Ssimon Clear the error queue to ensure that error entries left from 2253194206Ssimon older function calls do not interfere with the correct operation. 2254194206Ssimon [Lutz Jaenicke, Erik de Castro Lopo] 2255194206Ssimon 2256194206Ssimon *) Remove root CA certificates of commercial CAs: 2257194206Ssimon 2258194206Ssimon The OpenSSL project does not recommend any specific CA and does not 2259194206Ssimon have any policy with respect to including or excluding any CA. 2260194206Ssimon Therefore it does not make any sense to ship an arbitrary selection 2261194206Ssimon of root CA certificates with the OpenSSL software. 2262194206Ssimon [Lutz Jaenicke] 2263194206Ssimon 2264194206Ssimon *) RSA OAEP patches to fix two separate invalid memory reads. 2265194206Ssimon The first one involves inputs when 'lzero' is greater than 2266194206Ssimon 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes 2267194206Ssimon before the beginning of from). The second one involves inputs where 2268194206Ssimon the 'db' section contains nothing but zeroes (there is a one-byte 2269194206Ssimon invalid read after the end of 'db'). 2270194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 2271194206Ssimon 2272194206Ssimon *) Partial backport from 0.9.9-dev: 2273194206Ssimon 2274194206Ssimon Introduce bn_mul_mont (dedicated Montgomery multiplication 2275194206Ssimon procedure) as a candidate for BIGNUM assembler implementation. 2276194206Ssimon While 0.9.9-dev uses assembler for various architectures, only 2277194206Ssimon x86_64 is available by default here in the 0.9.8 branch, and 2278194206Ssimon 32-bit x86 is available through a compile-time setting. 2279194206Ssimon 2280194206Ssimon To try the 32-bit x86 assembler implementation, use Configure 2281194206Ssimon option "enable-montasm" (which exists only for this backport). 2282194206Ssimon 2283194206Ssimon As "enable-montasm" for 32-bit x86 disclaims code stability 2284194206Ssimon anyway, in this constellation we activate additional code 2285194206Ssimon backported from 0.9.9-dev for further performance improvements, 2286194206Ssimon namely BN_from_montgomery_word. (To enable this otherwise, 2287194206Ssimon e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".) 2288194206Ssimon 2289194206Ssimon [Andy Polyakov (backport partially by Bodo Moeller)] 2290194206Ssimon 2291194206Ssimon *) Add TLS session ticket callback. This allows an application to set 2292194206Ssimon TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed 2293194206Ssimon values. This is useful for key rollover for example where several key 2294194206Ssimon sets may exist with different names. 2295194206Ssimon [Steve Henson] 2296194206Ssimon 2297194206Ssimon *) Reverse ENGINE-internal logic for caching default ENGINE handles. 2298194206Ssimon This was broken until now in 0.9.8 releases, such that the only way 2299194206Ssimon a registered ENGINE could be used (assuming it initialises 2300194206Ssimon successfully on the host) was to explicitly set it as the default 2301194206Ssimon for the relevant algorithms. This is in contradiction with 0.9.7 2302194206Ssimon behaviour and the documentation. With this fix, when an ENGINE is 2303194206Ssimon registered into a given algorithm's table of implementations, the 2304194206Ssimon 'uptodate' flag is reset so that auto-discovery will be used next 2305194206Ssimon time a new context for that algorithm attempts to select an 2306194206Ssimon implementation. 2307194206Ssimon [Ian Lister (tweaked by Geoff Thorpe)] 2308194206Ssimon 2309194206Ssimon *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 2310194206Ssimon implemention in the following ways: 2311194206Ssimon 2312194206Ssimon Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be 2313194206Ssimon hard coded. 2314194206Ssimon 2315194206Ssimon Lack of BER streaming support means one pass streaming processing is 2316194206Ssimon only supported if data is detached: setting the streaming flag is 2317194206Ssimon ignored for embedded content. 2318194206Ssimon 2319194206Ssimon CMS support is disabled by default and must be explicitly enabled 2320194206Ssimon with the enable-cms configuration option. 2321194206Ssimon [Steve Henson] 2322194206Ssimon 2323194206Ssimon *) Update the GMP engine glue to do direct copies between BIGNUM and 2324194206Ssimon mpz_t when openssl and GMP use the same limb size. Otherwise the 2325194206Ssimon existing "conversion via a text string export" trick is still used. 2326194206Ssimon [Paul Sheer <paulsheer@gmail.com>] 2327194206Ssimon 2328194206Ssimon *) Zlib compression BIO. This is a filter BIO which compressed and 2329194206Ssimon uncompresses any data passed through it. 2330194206Ssimon [Steve Henson] 2331194206Ssimon 2332194206Ssimon *) Add AES_wrap_key() and AES_unwrap_key() functions to implement 2333194206Ssimon RFC3394 compatible AES key wrapping. 2334194206Ssimon [Steve Henson] 2335194206Ssimon 2336194206Ssimon *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): 2337194206Ssimon sets string data without copying. X509_ALGOR_set0() and 2338194206Ssimon X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) 2339194206Ssimon data. Attribute function X509at_get0_data_by_OBJ(): retrieves data 2340194206Ssimon from an X509_ATTRIBUTE structure optionally checking it occurs only 2341194206Ssimon once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied 2342194206Ssimon data. 2343194206Ssimon [Steve Henson] 2344194206Ssimon 2345194206Ssimon *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() 2346194206Ssimon to get the expected BN_FLG_CONSTTIME behavior. 2347194206Ssimon [Bodo Moeller (Google)] 2348194206Ssimon 2349194206Ssimon *) Netware support: 2350194206Ssimon 2351194206Ssimon - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets 2352194206Ssimon - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) 2353194206Ssimon - added some more tests to do_tests.pl 2354194206Ssimon - fixed RunningProcess usage so that it works with newer LIBC NDKs too 2355194206Ssimon - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency 2356194206Ssimon - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, 2357194206Ssimon netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc 2358194206Ssimon - various changes to netware.pl to enable gcc-cross builds on Win32 2359194206Ssimon platform 2360194206Ssimon - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) 2361194206Ssimon - various changes to fix missing prototype warnings 2362194206Ssimon - fixed x86nasm.pl to create correct asm files for NASM COFF output 2363194206Ssimon - added AES, WHIRLPOOL and CPUID assembler code to build files 2364194206Ssimon - added missing AES assembler make rules to mk1mf.pl 2365194206Ssimon - fixed order of includes in apps/ocsp.c so that e_os.h settings apply 2366194206Ssimon [Guenter Knauf <eflash@gmx.net>] 2367194206Ssimon 2368194206Ssimon *) Implement certificate status request TLS extension defined in RFC3546. 2369194206Ssimon A client can set the appropriate parameters and receive the encoded 2370194206Ssimon OCSP response via a callback. A server can query the supplied parameters 2371194206Ssimon and set the encoded OCSP response in the callback. Add simplified examples 2372194206Ssimon to s_client and s_server. 2373194206Ssimon [Steve Henson] 2374194206Ssimon 2375194206Ssimon Changes between 0.9.8f and 0.9.8g [19 Oct 2007] 2376194206Ssimon 2377194206Ssimon *) Fix various bugs: 2378194206Ssimon + Binary incompatibility of ssl_ctx_st structure 2379194206Ssimon + DTLS interoperation with non-compliant servers 2380194206Ssimon + Don't call get_session_cb() without proposed session 2381194206Ssimon + Fix ia64 assembler code 2382194206Ssimon [Andy Polyakov, Steve Henson] 2383194206Ssimon 2384194206Ssimon Changes between 0.9.8e and 0.9.8f [11 Oct 2007] 2385194206Ssimon 2386194206Ssimon *) DTLS Handshake overhaul. There were longstanding issues with 2387194206Ssimon OpenSSL DTLS implementation, which were making it impossible for 2388194206Ssimon RFC 4347 compliant client to communicate with OpenSSL server. 2389194206Ssimon Unfortunately just fixing these incompatibilities would "cut off" 2390194206Ssimon pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e 2391194206Ssimon server keeps tolerating non RFC compliant syntax. The opposite is 2392194206Ssimon not true, 0.9.8f client can not communicate with earlier server. 2393194206Ssimon This update even addresses CVE-2007-4995. 2394194206Ssimon [Andy Polyakov] 2395194206Ssimon 2396194206Ssimon *) Changes to avoid need for function casts in OpenSSL: some compilers 2397194206Ssimon (gcc 4.2 and later) reject their use. 2398194206Ssimon [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, 2399194206Ssimon Steve Henson] 2400194206Ssimon 2401194206Ssimon *) Add RFC4507 support to OpenSSL. This includes the corrections in 2402194206Ssimon RFC4507bis. The encrypted ticket format is an encrypted encoded 2403194206Ssimon SSL_SESSION structure, that way new session features are automatically 2404194206Ssimon supported. 2405194206Ssimon 2406194206Ssimon If a client application caches session in an SSL_SESSION structure 2407194206Ssimon support is transparent because tickets are now stored in the encoded 2408194206Ssimon SSL_SESSION. 2409194206Ssimon 2410194206Ssimon The SSL_CTX structure automatically generates keys for ticket 2411194206Ssimon protection in servers so again support should be possible 2412194206Ssimon with no application modification. 2413194206Ssimon 2414194206Ssimon If a client or server wishes to disable RFC4507 support then the option 2415194206Ssimon SSL_OP_NO_TICKET can be set. 2416194206Ssimon 2417194206Ssimon Add a TLS extension debugging callback to allow the contents of any client 2418194206Ssimon or server extensions to be examined. 2419194206Ssimon 2420194206Ssimon This work was sponsored by Google. 2421194206Ssimon [Steve Henson] 2422194206Ssimon 2423194206Ssimon *) Add initial support for TLS extensions, specifically for the server_name 2424194206Ssimon extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 2425194206Ssimon have new members for a host name. The SSL data structure has an 2426194206Ssimon additional member SSL_CTX *initial_ctx so that new sessions can be 2427194206Ssimon stored in that context to allow for session resumption, even after the 2428194206Ssimon SSL has been switched to a new SSL_CTX in reaction to a client's 2429194206Ssimon server_name extension. 2430194206Ssimon 2431194206Ssimon New functions (subject to change): 2432194206Ssimon 2433194206Ssimon SSL_get_servername() 2434194206Ssimon SSL_get_servername_type() 2435194206Ssimon SSL_set_SSL_CTX() 2436194206Ssimon 2437194206Ssimon New CTRL codes and macros (subject to change): 2438194206Ssimon 2439194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 2440194206Ssimon - SSL_CTX_set_tlsext_servername_callback() 2441194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 2442194206Ssimon - SSL_CTX_set_tlsext_servername_arg() 2443194206Ssimon SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 2444194206Ssimon 2445194206Ssimon openssl s_client has a new '-servername ...' option. 2446194206Ssimon 2447194206Ssimon openssl s_server has new options '-servername_host ...', '-cert2 ...', 2448194206Ssimon '-key2 ...', '-servername_fatal' (subject to change). This allows 2449194206Ssimon testing the HostName extension for a specific single host name ('-cert' 2450194206Ssimon and '-key' remain fallbacks for handshakes without HostName 2451194206Ssimon negotiation). If the unrecogninzed_name alert has to be sent, this by 2452194206Ssimon default is a warning; it becomes fatal with the '-servername_fatal' 2453194206Ssimon option. 2454194206Ssimon 2455194206Ssimon [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson] 2456194206Ssimon 2457194206Ssimon *) Add AES and SSE2 assembly language support to VC++ build. 2458194206Ssimon [Steve Henson] 2459194206Ssimon 2460194206Ssimon *) Mitigate attack on final subtraction in Montgomery reduction. 2461194206Ssimon [Andy Polyakov] 2462194206Ssimon 2463194206Ssimon *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 2464194206Ssimon (which previously caused an internal error). 2465194206Ssimon [Bodo Moeller] 2466194206Ssimon 2467194206Ssimon *) Squeeze another 10% out of IGE mode when in != out. 2468194206Ssimon [Ben Laurie] 2469194206Ssimon 2470194206Ssimon *) AES IGE mode speedup. 2471194206Ssimon [Dean Gaudet (Google)] 2472194206Ssimon 2473194206Ssimon *) Add the Korean symmetric 128-bit cipher SEED (see 2474194206Ssimon http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and 2475194206Ssimon add SEED ciphersuites from RFC 4162: 2476194206Ssimon 2477194206Ssimon TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" 2478194206Ssimon TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" 2479194206Ssimon TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" 2480194206Ssimon TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" 2481194206Ssimon 2482194206Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 2483194206Ssimon series, SEED remains excluded from compilation unless OpenSSL 2484194206Ssimon is configured with 'enable-seed'. 2485194206Ssimon [KISA, Bodo Moeller] 2486194206Ssimon 2487194206Ssimon *) Mitigate branch prediction attacks, which can be practical if a 2488194206Ssimon single processor is shared, allowing a spy process to extract 2489194206Ssimon information. For detailed background information, see 2490194206Ssimon http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron, 2491194206Ssimon J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL 2492194206Ssimon and Necessary Software Countermeasures"). The core of the change 2493194206Ssimon are new versions BN_div_no_branch() and 2494194206Ssimon BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), 2495194206Ssimon respectively, which are slower, but avoid the security-relevant 2496194206Ssimon conditional branches. These are automatically called by BN_div() 2497194206Ssimon and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one 2498194206Ssimon of the input BIGNUMs. Also, BN_is_bit_set() has been changed to 2499194206Ssimon remove a conditional branch. 2500194206Ssimon 2501194206Ssimon BN_FLG_CONSTTIME is the new name for the previous 2502194206Ssimon BN_FLG_EXP_CONSTTIME flag, since it now affects more than just 2503194206Ssimon modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag 2504194206Ssimon in the exponent causes BN_mod_exp_mont() to use the alternative 2505194206Ssimon implementation in BN_mod_exp_mont_consttime().) The old name 2506194206Ssimon remains as a deprecated alias. 2507194206Ssimon 2508194206Ssimon Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general 2509194206Ssimon RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses 2510194206Ssimon constant-time implementations for more than just exponentiation. 2511194206Ssimon Here too the old name is kept as a deprecated alias. 2512194206Ssimon 2513194206Ssimon BN_BLINDING_new() will now use BN_dup() for the modulus so that 2514194206Ssimon the BN_BLINDING structure gets an independent copy of the 2515194206Ssimon modulus. This means that the previous "BIGNUM *m" argument to 2516194206Ssimon BN_BLINDING_new() and to BN_BLINDING_create_param() now 2517194206Ssimon essentially becomes "const BIGNUM *m", although we can't actually 2518194206Ssimon change this in the header file before 0.9.9. It allows 2519194206Ssimon RSA_setup_blinding() to use BN_with_flags() on the modulus to 2520194206Ssimon enable BN_FLG_CONSTTIME. 2521194206Ssimon 2522194206Ssimon [Matthew D Wood (Intel Corp)] 2523194206Ssimon 2524194206Ssimon *) In the SSL/TLS server implementation, be strict about session ID 2525194206Ssimon context matching (which matters if an application uses a single 2526194206Ssimon external cache for different purposes). Previously, 2527194206Ssimon out-of-context reuse was forbidden only if SSL_VERIFY_PEER was 2528194206Ssimon set. This did ensure strict client verification, but meant that, 2529194206Ssimon with applications using a single external cache for quite 2530194206Ssimon different requirements, clients could circumvent ciphersuite 2531194206Ssimon restrictions for a given session ID context by starting a session 2532194206Ssimon in a different context. 2533194206Ssimon [Bodo Moeller] 2534194206Ssimon 2535167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 2536167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 2537167612Ssimon authentication-only ciphersuites. 2538167612Ssimon [Bodo Moeller] 2539167612Ssimon 2540194206Ssimon *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was 2541194206Ssimon not complete and could lead to a possible single byte overflow 2542194206Ssimon (CVE-2007-5135) [Ben Laurie] 2543194206Ssimon 2544194206Ssimon Changes between 0.9.8d and 0.9.8e [23 Feb 2007] 2545194206Ssimon 2546167612Ssimon *) Since AES128 and AES256 (and similarly Camellia128 and 2547167612Ssimon Camellia256) share a single mask bit in the logic of 2548167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 2549167612Ssimon kludge to work properly if AES128 is available and AES256 isn't 2550167612Ssimon (or if Camellia128 is available and Camellia256 isn't). 2551167612Ssimon [Victor Duchovni] 2552167612Ssimon 2553167612Ssimon *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c 2554167612Ssimon (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): 2555167612Ssimon When a point or a seed is encoded in a BIT STRING, we need to 2556167612Ssimon prevent the removal of trailing zero bits to get the proper DER 2557167612Ssimon encoding. (By default, crypto/asn1/a_bitstr.c assumes the case 2558167612Ssimon of a NamedBitList, for which trailing 0 bits need to be removed.) 2559167612Ssimon [Bodo Moeller] 2560167612Ssimon 2561167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 2562167612Ssimon protocol version while receiving ClientHello even if the 2563167612Ssimon ClientHello is fragmented. (The server can't insist on the 2564167612Ssimon particular protocol version it has chosen before the ServerHello 2565167612Ssimon message has informed the client about his choice.) 2566167612Ssimon [Bodo Moeller] 2567167612Ssimon 2568167612Ssimon *) Add RFC 3779 support. 2569167612Ssimon [Rob Austein for ARIN, Ben Laurie] 2570167612Ssimon 2571167612Ssimon *) Load error codes if they are not already present instead of using a 2572167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 2573167612Ssimon Improve header file function name parsing. 2574167612Ssimon [Steve Henson] 2575167612Ssimon 2576167612Ssimon *) extend SMTP and IMAP protocol emulation in s_client to use EHLO 2577167612Ssimon or CAPABILITY handshake as required by RFCs. 2578167612Ssimon [Goetz Babin-Ebell] 2579167612Ssimon 2580162911Ssimon Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 2581162911Ssimon 2582162911Ssimon *) Introduce limits to prevent malicious keys being able to 2583162911Ssimon cause a denial of service. (CVE-2006-2940) 2584162911Ssimon [Steve Henson, Bodo Moeller] 2585162911Ssimon 2586162911Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 2587162911Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 2588162911Ssimon 2589162911Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 2590162911Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 2591162911Ssimon 2592162911Ssimon *) Fix SSL client code which could crash if connecting to a 2593162911Ssimon malicious SSLv2 server. (CVE-2006-4343) 2594162911Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 2595162911Ssimon 2596162911Ssimon *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 2597162911Ssimon match only those. Before that, "AES256-SHA" would be interpreted 2598162911Ssimon as a pattern and match "AES128-SHA" too (since AES128-SHA got 2599162911Ssimon the same strength classification in 0.9.7h) as we currently only 2600162911Ssimon have a single AES bit in the ciphersuite description bitmap. 2601162911Ssimon That change, however, also applied to ciphersuite strings such as 2602162911Ssimon "RC4-MD5" that intentionally matched multiple ciphersuites -- 2603162911Ssimon namely, SSL 2.0 ciphersuites in addition to the more common ones 2604162911Ssimon from SSL 3.0/TLS 1.0. 2605162911Ssimon 2606162911Ssimon So we change the selection algorithm again: Naming an explicit 2607162911Ssimon ciphersuite selects this one ciphersuite, and any other similar 2608162911Ssimon ciphersuite (same bitmap) from *other* protocol versions. 2609162911Ssimon Thus, "RC4-MD5" again will properly select both the SSL 2.0 2610162911Ssimon ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 2611162911Ssimon 2612162911Ssimon Since SSL 2.0 does not have any ciphersuites for which the 2613162911Ssimon 128/256 bit distinction would be relevant, this works for now. 2614162911Ssimon The proper fix will be to use different bits for AES128 and 2615162911Ssimon AES256, which would have avoided the problems from the beginning; 2616162911Ssimon however, bits are scarce, so we can only do this in a new release 2617162911Ssimon (not just a patchlevel) when we can change the SSL_CIPHER 2618162911Ssimon definition to split the single 'unsigned long mask' bitmap into 2619162911Ssimon multiple values to extend the available space. 2620162911Ssimon 2621162911Ssimon [Bodo Moeller] 2622162911Ssimon 2623162911Ssimon Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 2624162911Ssimon 2625162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 2626162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 2627162911Ssimon 2628162911Ssimon *) Add AES IGE and biIGE modes. 2629162911Ssimon [Ben Laurie] 2630162911Ssimon 2631162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 2632162911Ssimon possible instead of select(), since the latter has some 2633162911Ssimon undesirable limitations. 2634162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 2635162911Ssimon 2636162911Ssimon *) Disable "ECCdraft" ciphersuites more thoroughly. Now special 2637162911Ssimon treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 2638162911Ssimon cannot be implicitly activated as part of, e.g., the "AES" alias. 2639162911Ssimon However, please upgrade to OpenSSL 0.9.9[-dev] for 2640162911Ssimon non-experimental use of the ECC ciphersuites to get TLS extension 2641162911Ssimon support, which is required for curve and point format negotiation 2642162911Ssimon to avoid potential handshake problems. 2643162911Ssimon [Bodo Moeller] 2644162911Ssimon 2645162911Ssimon *) Disable rogue ciphersuites: 2646162911Ssimon 2647162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 2648162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 2649162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 2650162911Ssimon 2651162911Ssimon The latter two were purportedly from 2652162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 2653162911Ssimon appear there. 2654162911Ssimon 2655167612Ssimon Also deactivate the remaining ciphersuites from 2656162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 2657162911Ssimon unofficial, and the ID has long expired. 2658162911Ssimon [Bodo Moeller] 2659162911Ssimon 2660162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 2661162911Ssimon dual-core machines) and other potential thread-safety issues. 2662162911Ssimon [Bodo Moeller] 2663162911Ssimon 2664162911Ssimon *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 2665162911Ssimon versions), which is now available for royalty-free use 2666162911Ssimon (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). 2667162911Ssimon Also, add Camellia TLS ciphersuites from RFC 4132. 2668162911Ssimon 2669162911Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 2670162911Ssimon series, Camellia remains excluded from compilation unless OpenSSL 2671162911Ssimon is configured with 'enable-camellia'. 2672162911Ssimon [NTT] 2673162911Ssimon 2674162911Ssimon *) Disable the padding bug check when compression is in use. The padding 2675162911Ssimon bug check assumes the first packet is of even length, this is not 2676162911Ssimon necessarily true if compresssion is enabled and can result in false 2677162911Ssimon positives causing handshake failure. The actual bug test is ancient 2678162911Ssimon code so it is hoped that implementations will either have fixed it by 2679162911Ssimon now or any which still have the bug do not support compression. 2680162911Ssimon [Steve Henson] 2681162911Ssimon 2682160814Ssimon Changes between 0.9.8a and 0.9.8b [04 May 2006] 2683160814Ssimon 2684160814Ssimon *) When applying a cipher rule check to see if string match is an explicit 2685160814Ssimon cipher suite and only match that one cipher suite if it is. 2686160814Ssimon [Steve Henson] 2687160814Ssimon 2688160814Ssimon *) Link in manifests for VC++ if needed. 2689160814Ssimon [Austin Ziegler <halostatue@gmail.com>] 2690160814Ssimon 2691160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 2692160814Ssimon draft-ietf-tls-ecc-12.txt with proposed changes (but without 2693160814Ssimon TLS extensions, which are supported starting with the 0.9.9 2694160814Ssimon branch, not in the OpenSSL 0.9.8 branch). 2695160814Ssimon [Douglas Stebila] 2696160814Ssimon 2697160814Ssimon *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 2698160814Ssimon opaque EVP_CIPHER_CTX handling. 2699160814Ssimon [Steve Henson] 2700160814Ssimon 2701160814Ssimon *) Fixes and enhancements to zlib compression code. We now only use 2702160814Ssimon "zlib1.dll" and use the default __cdecl calling convention on Win32 2703160814Ssimon to conform with the standards mentioned here: 2704160814Ssimon http://www.zlib.net/DLL_FAQ.txt 2705160814Ssimon Static zlib linking now works on Windows and the new --with-zlib-include 2706160814Ssimon --with-zlib-lib options to Configure can be used to supply the location 2707160814Ssimon of the headers and library. Gracefully handle case where zlib library 2708160814Ssimon can't be loaded. 2709160814Ssimon [Steve Henson] 2710160814Ssimon 2711160814Ssimon *) Several fixes and enhancements to the OID generation code. The old code 2712160814Ssimon sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 2713160814Ssimon handle numbers larger than ULONG_MAX, truncated printing and had a 2714160814Ssimon non standard OBJ_obj2txt() behaviour. 2715160814Ssimon [Steve Henson] 2716160814Ssimon 2717160814Ssimon *) Add support for building of engines under engine/ as shared libraries 2718160814Ssimon under VC++ build system. 2719160814Ssimon [Steve Henson] 2720160814Ssimon 2721160814Ssimon *) Corrected the numerous bugs in the Win32 path splitter in DSO. 2722160814Ssimon Hopefully, we will not see any false combination of paths any more. 2723160814Ssimon [Richard Levitte] 2724160814Ssimon 2725160814Ssimon Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 2726160814Ssimon 2727160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 2728160814Ssimon (part of SSL_OP_ALL). This option used to disable the 2729160814Ssimon countermeasure against man-in-the-middle protocol-version 2730160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 2731160814Ssimon idea. (CVE-2005-2969) 2732160814Ssimon 2733160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 2734160814Ssimon for Information Security, National Institute of Advanced Industrial 2735160814Ssimon Science and Technology [AIST], Japan)] 2736160814Ssimon 2737160814Ssimon *) Add two function to clear and return the verify parameter flags. 2738160814Ssimon [Steve Henson] 2739160814Ssimon 2740160814Ssimon *) Keep cipherlists sorted in the source instead of sorting them at 2741160814Ssimon runtime, thus removing the need for a lock. 2742160814Ssimon [Nils Larsch] 2743160814Ssimon 2744160814Ssimon *) Avoid some small subgroup attacks in Diffie-Hellman. 2745160814Ssimon [Nick Mathewson and Ben Laurie] 2746160814Ssimon 2747160814Ssimon *) Add functions for well-known primes. 2748160814Ssimon [Nick Mathewson] 2749160814Ssimon 2750160814Ssimon *) Extended Windows CE support. 2751160814Ssimon [Satoshi Nakamura and Andy Polyakov] 2752160814Ssimon 2753160814Ssimon *) Initialize SSL_METHOD structures at compile time instead of during 2754160814Ssimon runtime, thus removing the need for a lock. 2755160814Ssimon [Steve Henson] 2756160814Ssimon 2757160814Ssimon *) Make PKCS7_decrypt() work even if no certificate is supplied by 2758160814Ssimon attempting to decrypt each encrypted key in turn. Add support to 2759160814Ssimon smime utility. 2760160814Ssimon [Steve Henson] 2761160814Ssimon 2762160814Ssimon Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 2763160814Ssimon 2764162911Ssimon [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 2765162911Ssimon OpenSSL 0.9.8.] 2766162911Ssimon 2767160814Ssimon *) Add libcrypto.pc and libssl.pc for those who feel they need them. 2768160814Ssimon [Richard Levitte] 2769160814Ssimon 2770160814Ssimon *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 2771160814Ssimon key into the same file any more. 2772160814Ssimon [Richard Levitte] 2773160814Ssimon 2774160814Ssimon *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 2775160814Ssimon [Andy Polyakov] 2776160814Ssimon 2777160814Ssimon *) Add -utf8 command line and config file option to 'ca'. 2778160814Ssimon [Stefan <stf@udoma.org] 2779160814Ssimon 2780160814Ssimon *) Removed the macro des_crypt(), as it seems to conflict with some 2781160814Ssimon libraries. Use DES_crypt(). 2782160814Ssimon [Richard Levitte] 2783160814Ssimon 2784160814Ssimon *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 2785160814Ssimon involves renaming the source and generated shared-libs for 2786160814Ssimon both. The engines will accept the corrected or legacy ids 2787160814Ssimon ('ncipher' and '4758_cca' respectively) when binding. NB, 2788160814Ssimon this only applies when building 'shared'. 2789160814Ssimon [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 2790160814Ssimon 2791160814Ssimon *) Add attribute functions to EVP_PKEY structure. Modify 2792160814Ssimon PKCS12_create() to recognize a CSP name attribute and 2793160814Ssimon use it. Make -CSP option work again in pkcs12 utility. 2794160814Ssimon [Steve Henson] 2795160814Ssimon 2796160814Ssimon *) Add new functionality to the bn blinding code: 2797160814Ssimon - automatic re-creation of the BN_BLINDING parameters after 2798160814Ssimon a fixed number of uses (currently 32) 2799160814Ssimon - add new function for parameter creation 2800160814Ssimon - introduce flags to control the update behaviour of the 2801160814Ssimon BN_BLINDING parameters 2802160814Ssimon - hide BN_BLINDING structure 2803160814Ssimon Add a second BN_BLINDING slot to the RSA structure to improve 2804160814Ssimon performance when a single RSA object is shared among several 2805160814Ssimon threads. 2806160814Ssimon [Nils Larsch] 2807160814Ssimon 2808160814Ssimon *) Add support for DTLS. 2809160814Ssimon [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 2810160814Ssimon 2811160814Ssimon *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 2812160814Ssimon to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 2813160814Ssimon [Walter Goulet] 2814160814Ssimon 2815160814Ssimon *) Remove buggy and incompletet DH cert support from 2816160814Ssimon ssl/ssl_rsa.c and ssl/s3_both.c 2817160814Ssimon [Nils Larsch] 2818160814Ssimon 2819160814Ssimon *) Use SHA-1 instead of MD5 as the default digest algorithm for 2820160814Ssimon the apps/openssl applications. 2821160814Ssimon [Nils Larsch] 2822160814Ssimon 2823160814Ssimon *) Compile clean with "-Wall -Wmissing-prototypes 2824160814Ssimon -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 2825160814Ssimon DEBUG_SAFESTACK must also be set. 2826160814Ssimon [Ben Laurie] 2827160814Ssimon 2828160814Ssimon *) Change ./Configure so that certain algorithms can be disabled by default. 2829160814Ssimon The new counterpiece to "no-xxx" is "enable-xxx". 2830160814Ssimon 2831160814Ssimon The patented RC5 and MDC2 algorithms will now be disabled unless 2832160814Ssimon "enable-rc5" and "enable-mdc2", respectively, are specified. 2833160814Ssimon 2834160814Ssimon (IDEA remains enabled despite being patented. This is because IDEA 2835160814Ssimon is frequently required for interoperability, and there is no license 2836160814Ssimon fee for non-commercial use. As before, "no-idea" can be used to 2837160814Ssimon avoid this algorithm.) 2838160814Ssimon 2839160814Ssimon [Bodo Moeller] 2840160814Ssimon 2841160814Ssimon *) Add processing of proxy certificates (see RFC 3820). This work was 2842160814Ssimon sponsored by KTH (The Royal Institute of Technology in Stockholm) and 2843160814Ssimon EGEE (Enabling Grids for E-science in Europe). 2844160814Ssimon [Richard Levitte] 2845160814Ssimon 2846160814Ssimon *) RC4 performance overhaul on modern architectures/implementations, such 2847160814Ssimon as Intel P4, IA-64 and AMD64. 2848160814Ssimon [Andy Polyakov] 2849160814Ssimon 2850160814Ssimon *) New utility extract-section.pl. This can be used specify an alternative 2851160814Ssimon section number in a pod file instead of having to treat each file as 2852160814Ssimon a separate case in Makefile. This can be done by adding two lines to the 2853160814Ssimon pod file: 2854160814Ssimon 2855160814Ssimon =for comment openssl_section:XXX 2856160814Ssimon 2857160814Ssimon The blank line is mandatory. 2858160814Ssimon 2859160814Ssimon [Steve Henson] 2860160814Ssimon 2861160814Ssimon *) New arguments -certform, -keyform and -pass for s_client and s_server 2862160814Ssimon to allow alternative format key and certificate files and passphrase 2863160814Ssimon sources. 2864160814Ssimon [Steve Henson] 2865160814Ssimon 2866160814Ssimon *) New structure X509_VERIFY_PARAM which combines current verify parameters, 2867160814Ssimon update associated structures and add various utility functions. 2868160814Ssimon 2869160814Ssimon Add new policy related verify parameters, include policy checking in 2870160814Ssimon standard verify code. Enhance 'smime' application with extra parameters 2871160814Ssimon to support policy checking and print out. 2872160814Ssimon [Steve Henson] 2873160814Ssimon 2874160814Ssimon *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 2875160814Ssimon Nehemiah processors. These extensions support AES encryption in hardware 2876160814Ssimon as well as RNG (though RNG support is currently disabled). 2877160814Ssimon [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 2878160814Ssimon 2879160814Ssimon *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 2880160814Ssimon [Geoff Thorpe] 2881160814Ssimon 2882160814Ssimon *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 2883160814Ssimon [Andy Polyakov and a number of other people] 2884160814Ssimon 2885160814Ssimon *) Improved PowerPC platform support. Most notably BIGNUM assembler 2886160814Ssimon implementation contributed by IBM. 2887160814Ssimon [Suresh Chari, Peter Waltenberg, Andy Polyakov] 2888160814Ssimon 2889160814Ssimon *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 2890160814Ssimon exponent rather than 'unsigned long'. There is a corresponding change to 2891160814Ssimon the new 'rsa_keygen' element of the RSA_METHOD structure. 2892160814Ssimon [Jelte Jansen, Geoff Thorpe] 2893160814Ssimon 2894160814Ssimon *) Functionality for creating the initial serial number file is now 2895160814Ssimon moved from CA.pl to the 'ca' utility with a new option -create_serial. 2896160814Ssimon 2897160814Ssimon (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 2898160814Ssimon number file to 1, which is bound to cause problems. To avoid 2899160814Ssimon the problems while respecting compatibility between different 0.9.7 2900160814Ssimon patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 2901160814Ssimon CA.pl for serial number initialization. With the new release 0.9.8, 2902160814Ssimon we can fix the problem directly in the 'ca' utility.) 2903160814Ssimon [Steve Henson] 2904160814Ssimon 2905160814Ssimon *) Reduced header interdepencies by declaring more opaque objects in 2906160814Ssimon ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 2907160814Ssimon give fewer recursive includes, which could break lazy source code - so 2908160814Ssimon this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 2909160814Ssimon developers should define this symbol when building and using openssl to 2910160814Ssimon ensure they track the recommended behaviour, interfaces, [etc], but 2911160814Ssimon backwards-compatible behaviour prevails when this isn't defined. 2912160814Ssimon [Geoff Thorpe] 2913160814Ssimon 2914160814Ssimon *) New function X509_POLICY_NODE_print() which prints out policy nodes. 2915160814Ssimon [Steve Henson] 2916160814Ssimon 2917160814Ssimon *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 2918160814Ssimon This will generate a random key of the appropriate length based on the 2919160814Ssimon cipher context. The EVP_CIPHER can provide its own random key generation 2920160814Ssimon routine to support keys of a specific form. This is used in the des and 2921160814Ssimon 3des routines to generate a key of the correct parity. Update S/MIME 2922160814Ssimon code to use new functions and hence generate correct parity DES keys. 2923160814Ssimon Add EVP_CHECK_DES_KEY #define to return an error if the key is not 2924160814Ssimon valid (weak or incorrect parity). 2925160814Ssimon [Steve Henson] 2926160814Ssimon 2927160814Ssimon *) Add a local set of CRLs that can be used by X509_verify_cert() as well 2928160814Ssimon as looking them up. This is useful when the verified structure may contain 2929160814Ssimon CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 2930160814Ssimon present unless the new PKCS7_NO_CRL flag is asserted. 2931160814Ssimon [Steve Henson] 2932160814Ssimon 2933160814Ssimon *) Extend ASN1 oid configuration module. It now additionally accepts the 2934160814Ssimon syntax: 2935160814Ssimon 2936160814Ssimon shortName = some long name, 1.2.3.4 2937160814Ssimon [Steve Henson] 2938160814Ssimon 2939160814Ssimon *) Reimplemented the BN_CTX implementation. There is now no more static 2940160814Ssimon limitation on the number of variables it can handle nor the depth of the 2941160814Ssimon "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 2942160814Ssimon information can now expand as required, and rather than having a single 2943160814Ssimon static array of bignums, BN_CTX now uses a linked-list of such arrays 2944160814Ssimon allowing it to expand on demand whilst maintaining the usefulness of 2945160814Ssimon BN_CTX's "bundling". 2946160814Ssimon [Geoff Thorpe] 2947160814Ssimon 2948160814Ssimon *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 2949160814Ssimon to allow all RSA operations to function using a single BN_CTX. 2950160814Ssimon [Geoff Thorpe] 2951160814Ssimon 2952160814Ssimon *) Preliminary support for certificate policy evaluation and checking. This 2953160814Ssimon is initially intended to pass the tests outlined in "Conformance Testing 2954160814Ssimon of Relying Party Client Certificate Path Processing Logic" v1.07. 2955160814Ssimon [Steve Henson] 2956160814Ssimon 2957160814Ssimon *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 2958160814Ssimon remained unused and not that useful. A variety of other little bignum 2959160814Ssimon tweaks and fixes have also been made continuing on from the audit (see 2960160814Ssimon below). 2961160814Ssimon [Geoff Thorpe] 2962160814Ssimon 2963160814Ssimon *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 2964160814Ssimon associated ASN1, EVP and SSL functions and old ASN1 macros. 2965160814Ssimon [Richard Levitte] 2966160814Ssimon 2967160814Ssimon *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 2968160814Ssimon and this should never fail. So the return value from the use of 2969160814Ssimon BN_set_word() (which can fail due to needless expansion) is now deprecated; 2970160814Ssimon if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 2971160814Ssimon [Geoff Thorpe] 2972160814Ssimon 2973160814Ssimon *) BN_CTX_get() should return zero-valued bignums, providing the same 2974160814Ssimon initialised value as BN_new(). 2975160814Ssimon [Geoff Thorpe, suggested by Ulf M�ller] 2976160814Ssimon 2977160814Ssimon *) Support for inhibitAnyPolicy certificate extension. 2978160814Ssimon [Steve Henson] 2979160814Ssimon 2980160814Ssimon *) An audit of the BIGNUM code is underway, for which debugging code is 2981160814Ssimon enabled when BN_DEBUG is defined. This makes stricter enforcements on what 2982160814Ssimon is considered valid when processing BIGNUMs, and causes execution to 2983160814Ssimon assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 2984160814Ssimon further steps are taken to deliberately pollute unused data in BIGNUM 2985160814Ssimon structures to try and expose faulty code further on. For now, openssl will 2986160814Ssimon (in its default mode of operation) continue to tolerate the inconsistent 2987160814Ssimon forms that it has tolerated in the past, but authors and packagers should 2988160814Ssimon consider trying openssl and their own applications when compiled with 2989160814Ssimon these debugging symbols defined. It will help highlight potential bugs in 2990160814Ssimon their own code, and will improve the test coverage for OpenSSL itself. At 2991160814Ssimon some point, these tighter rules will become openssl's default to improve 2992160814Ssimon maintainability, though the assert()s and other overheads will remain only 2993160814Ssimon in debugging configurations. See bn.h for more details. 2994160814Ssimon [Geoff Thorpe, Nils Larsch, Ulf M�ller] 2995160814Ssimon 2996160814Ssimon *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 2997160814Ssimon that can only be obtained through BN_CTX_new() (which implicitly 2998160814Ssimon initialises it). The presence of this function only made it possible 2999160814Ssimon to overwrite an existing structure (and cause memory leaks). 3000160814Ssimon [Geoff Thorpe] 3001160814Ssimon 3002160814Ssimon *) Because of the callback-based approach for implementing LHASH as a 3003160814Ssimon template type, lh_insert() adds opaque objects to hash-tables and 3004160814Ssimon lh_doall() or lh_doall_arg() are typically used with a destructor callback 3005160814Ssimon to clean up those corresponding objects before destroying the hash table 3006160814Ssimon (and losing the object pointers). So some over-zealous constifications in 3007160814Ssimon LHASH have been relaxed so that lh_insert() does not take (nor store) the 3008160814Ssimon objects as "const" and the lh_doall[_arg] callback wrappers are not 3009160814Ssimon prototyped to have "const" restrictions on the object pointers they are 3010160814Ssimon given (and so aren't required to cast them away any more). 3011160814Ssimon [Geoff Thorpe] 3012160814Ssimon 3013160814Ssimon *) The tmdiff.h API was so ugly and minimal that our own timing utility 3014160814Ssimon (speed) prefers to use its own implementation. The two implementations 3015160814Ssimon haven't been consolidated as yet (volunteers?) but the tmdiff API has had 3016160814Ssimon its object type properly exposed (MS_TM) instead of casting to/from "char 3017160814Ssimon *". This may still change yet if someone realises MS_TM and "ms_time_***" 3018160814Ssimon aren't necessarily the greatest nomenclatures - but this is what was used 3019160814Ssimon internally to the implementation so I've used that for now. 3020160814Ssimon [Geoff Thorpe] 3021160814Ssimon 3022160814Ssimon *) Ensure that deprecated functions do not get compiled when 3023160814Ssimon OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 3024160814Ssimon the self-tests were still using deprecated key-generation functions so 3025160814Ssimon these have been updated also. 3026160814Ssimon [Geoff Thorpe] 3027160814Ssimon 3028160814Ssimon *) Reorganise PKCS#7 code to separate the digest location functionality 3029160814Ssimon into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 3030160814Ssimon New function PKCS7_set_digest() to set the digest type for PKCS#7 3031160814Ssimon digestedData type. Add additional code to correctly generate the 3032160814Ssimon digestedData type and add support for this type in PKCS7 initialization 3033160814Ssimon functions. 3034160814Ssimon [Steve Henson] 3035160814Ssimon 3036160814Ssimon *) New function PKCS7_set0_type_other() this initializes a PKCS7 3037160814Ssimon structure of type "other". 3038160814Ssimon [Steve Henson] 3039160814Ssimon 3040160814Ssimon *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 3041160814Ssimon sure the loop does correctly stop and breaking ("division by zero") 3042160814Ssimon modulus operations are not performed. The (pre-generated) prime 3043160814Ssimon table crypto/bn/bn_prime.h was already correct, but it could not be 3044160814Ssimon re-generated on some platforms because of the "division by zero" 3045160814Ssimon situation in the script. 3046160814Ssimon [Ralf S. Engelschall] 3047160814Ssimon 3048160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 3049160814Ssimon draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 3050160814Ssimon SHA-1 now is only used for "small" curves (where the 3051160814Ssimon representation of a field element takes up to 24 bytes); for 3052160814Ssimon larger curves, the field element resulting from ECDH is directly 3053160814Ssimon used as premaster secret. 3054160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 3055160814Ssimon 3056160814Ssimon *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 3057160814Ssimon curve secp160r1 to the tests. 3058160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 3059160814Ssimon 3060160814Ssimon *) Add the possibility to load symbols globally with DSO. 3061160814Ssimon [G�tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 3062160814Ssimon 3063160814Ssimon *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 3064160814Ssimon control of the error stack. 3065160814Ssimon [Richard Levitte] 3066160814Ssimon 3067160814Ssimon *) Add support for STORE in ENGINE. 3068160814Ssimon [Richard Levitte] 3069160814Ssimon 3070160814Ssimon *) Add the STORE type. The intention is to provide a common interface 3071160814Ssimon to certificate and key stores, be they simple file-based stores, or 3072160814Ssimon HSM-type store, or LDAP stores, or... 3073160814Ssimon NOTE: The code is currently UNTESTED and isn't really used anywhere. 3074160814Ssimon [Richard Levitte] 3075160814Ssimon 3076160814Ssimon *) Add a generic structure called OPENSSL_ITEM. This can be used to 3077160814Ssimon pass a list of arguments to any function as well as provide a way 3078160814Ssimon for a function to pass data back to the caller. 3079160814Ssimon [Richard Levitte] 3080160814Ssimon 3081160814Ssimon *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 3082160814Ssimon works like BUF_strdup() but can be used to duplicate a portion of 3083160814Ssimon a string. The copy gets NUL-terminated. BUF_memdup() duplicates 3084160814Ssimon a memory area. 3085160814Ssimon [Richard Levitte] 3086160814Ssimon 3087160814Ssimon *) Add the function sk_find_ex() which works like sk_find(), but will 3088160814Ssimon return an index to an element even if an exact match couldn't be 3089160814Ssimon found. The index is guaranteed to point at the element where the 3090160814Ssimon searched-for key would be inserted to preserve sorting order. 3091160814Ssimon [Richard Levitte] 3092160814Ssimon 3093160814Ssimon *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 3094160814Ssimon takes an extra flags argument for optional functionality. Currently, 3095160814Ssimon the following flags are defined: 3096160814Ssimon 3097160814Ssimon OBJ_BSEARCH_VALUE_ON_NOMATCH 3098160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 3099160814Ssimon element where the comparing function returns a negative or zero 3100160814Ssimon number. 3101160814Ssimon 3102160814Ssimon OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 3103160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 3104160814Ssimon element where the comparing function returns zero. This is useful 3105160814Ssimon if there are more than one element where the comparing function 3106160814Ssimon returns zero. 3107160814Ssimon [Richard Levitte] 3108160814Ssimon 3109160814Ssimon *) Make it possible to create self-signed certificates with 'openssl ca' 3110160814Ssimon in such a way that the self-signed certificate becomes part of the 3111160814Ssimon CA database and uses the same mechanisms for serial number generation 3112160814Ssimon as all other certificate signing. The new flag '-selfsign' enables 3113160814Ssimon this functionality. Adapt CA.sh and CA.pl.in. 3114160814Ssimon [Richard Levitte] 3115160814Ssimon 3116160814Ssimon *) Add functionality to check the public key of a certificate request 3117160814Ssimon against a given private. This is useful to check that a certificate 3118160814Ssimon request can be signed by that key (self-signing). 3119160814Ssimon [Richard Levitte] 3120160814Ssimon 3121160814Ssimon *) Make it possible to have multiple active certificates with the same 3122160814Ssimon subject in the CA index file. This is done only if the keyword 3123160814Ssimon 'unique_subject' is set to 'no' in the main CA section (default 3124160814Ssimon if 'CA_default') of the configuration file. The value is saved 3125160814Ssimon with the database itself in a separate index attribute file, 3126160814Ssimon named like the index file with '.attr' appended to the name. 3127160814Ssimon [Richard Levitte] 3128160814Ssimon 3129160814Ssimon *) Generate muti valued AVAs using '+' notation in config files for 3130160814Ssimon req and dirName. 3131160814Ssimon [Steve Henson] 3132160814Ssimon 3133160814Ssimon *) Support for nameConstraints certificate extension. 3134160814Ssimon [Steve Henson] 3135160814Ssimon 3136160814Ssimon *) Support for policyConstraints certificate extension. 3137160814Ssimon [Steve Henson] 3138160814Ssimon 3139160814Ssimon *) Support for policyMappings certificate extension. 3140160814Ssimon [Steve Henson] 3141160814Ssimon 3142160814Ssimon *) Make sure the default DSA_METHOD implementation only uses its 3143160814Ssimon dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 3144160814Ssimon and change its own handlers to be NULL so as to remove unnecessary 3145160814Ssimon indirection. This lets alternative implementations fallback to the 3146160814Ssimon default implementation more easily. 3147160814Ssimon [Geoff Thorpe] 3148160814Ssimon 3149160814Ssimon *) Support for directoryName in GeneralName related extensions 3150160814Ssimon in config files. 3151160814Ssimon [Steve Henson] 3152160814Ssimon 3153160814Ssimon *) Make it possible to link applications using Makefile.shared. 3154160814Ssimon Make that possible even when linking against static libraries! 3155160814Ssimon [Richard Levitte] 3156160814Ssimon 3157160814Ssimon *) Support for single pass processing for S/MIME signing. This now 3158160814Ssimon means that S/MIME signing can be done from a pipe, in addition 3159160814Ssimon cleartext signing (multipart/signed type) is effectively streaming 3160160814Ssimon and the signed data does not need to be all held in memory. 3161160814Ssimon 3162160814Ssimon This is done with a new flag PKCS7_STREAM. When this flag is set 3163160814Ssimon PKCS7_sign() only initializes the PKCS7 structure and the actual signing 3164160814Ssimon is done after the data is output (and digests calculated) in 3165160814Ssimon SMIME_write_PKCS7(). 3166160814Ssimon [Steve Henson] 3167160814Ssimon 3168160814Ssimon *) Add full support for -rpath/-R, both in shared libraries and 3169160814Ssimon applications, at least on the platforms where it's known how 3170160814Ssimon to do it. 3171160814Ssimon [Richard Levitte] 3172160814Ssimon 3173160814Ssimon *) In crypto/ec/ec_mult.c, implement fast point multiplication with 3174160814Ssimon precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 3175160814Ssimon will now compute a table of multiples of the generator that 3176160814Ssimon makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 3177160814Ssimon faster (notably in the case of a single point multiplication, 3178160814Ssimon scalar * generator). 3179160814Ssimon [Nils Larsch, Bodo Moeller] 3180160814Ssimon 3181160814Ssimon *) IPv6 support for certificate extensions. The various extensions 3182160814Ssimon which use the IP:a.b.c.d can now take IPv6 addresses using the 3183160814Ssimon formats of RFC1884 2.2 . IPv6 addresses are now also displayed 3184160814Ssimon correctly. 3185160814Ssimon [Steve Henson] 3186160814Ssimon 3187160814Ssimon *) Added an ENGINE that implements RSA by performing private key 3188160814Ssimon exponentiations with the GMP library. The conversions to and from 3189160814Ssimon GMP's mpz_t format aren't optimised nor are any montgomery forms 3190160814Ssimon cached, and on x86 it appears OpenSSL's own performance has caught up. 3191160814Ssimon However there are likely to be other architectures where GMP could 3192160814Ssimon provide a boost. This ENGINE is not built in by default, but it can be 3193160814Ssimon specified at Configure time and should be accompanied by the necessary 3194160814Ssimon linker additions, eg; 3195160814Ssimon ./config -DOPENSSL_USE_GMP -lgmp 3196160814Ssimon [Geoff Thorpe] 3197160814Ssimon 3198160814Ssimon *) "openssl engine" will not display ENGINE/DSO load failure errors when 3199160814Ssimon testing availability of engines with "-t" - the old behaviour is 3200160814Ssimon produced by increasing the feature's verbosity with "-tt". 3201160814Ssimon [Geoff Thorpe] 3202160814Ssimon 3203160814Ssimon *) ECDSA routines: under certain error conditions uninitialized BN objects 3204160814Ssimon could be freed. Solution: make sure initialization is performed early 3205160814Ssimon enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 3206160814Ssimon via PR#459) 3207160814Ssimon [Lutz Jaenicke] 3208160814Ssimon 3209160814Ssimon *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 3210160814Ssimon and DH_METHOD (eg. by ENGINE implementations) to override the normal 3211160814Ssimon software implementations. For DSA and DH, parameter generation can 3212160814Ssimon also be overriden by providing the appropriate method callbacks. 3213160814Ssimon [Geoff Thorpe] 3214160814Ssimon 3215160814Ssimon *) Change the "progress" mechanism used in key-generation and 3216160814Ssimon primality testing to functions that take a new BN_GENCB pointer in 3217160814Ssimon place of callback/argument pairs. The new API functions have "_ex" 3218160814Ssimon postfixes and the older functions are reimplemented as wrappers for 3219160814Ssimon the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 3220160814Ssimon declarations of the old functions to help (graceful) attempts to 3221160814Ssimon migrate to the new functions. Also, the new key-generation API 3222160814Ssimon functions operate on a caller-supplied key-structure and return 3223160814Ssimon success/failure rather than returning a key or NULL - this is to 3224160814Ssimon help make "keygen" another member function of RSA_METHOD etc. 3225160814Ssimon 3226160814Ssimon Example for using the new callback interface: 3227160814Ssimon 3228160814Ssimon int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 3229160814Ssimon void *my_arg = ...; 3230160814Ssimon BN_GENCB my_cb; 3231160814Ssimon 3232160814Ssimon BN_GENCB_set(&my_cb, my_callback, my_arg); 3233160814Ssimon 3234160814Ssimon return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 3235160814Ssimon /* For the meaning of a, b in calls to my_callback(), see the 3236160814Ssimon * documentation of the function that calls the callback. 3237160814Ssimon * cb will point to my_cb; my_arg can be retrieved as cb->arg. 3238160814Ssimon * my_callback should return 1 if it wants BN_is_prime_ex() 3239160814Ssimon * to continue, or 0 to stop. 3240160814Ssimon */ 3241160814Ssimon 3242160814Ssimon [Geoff Thorpe] 3243160814Ssimon 3244160814Ssimon *) Change the ZLIB compression method to be stateful, and make it 3245160814Ssimon available to TLS with the number defined in 3246160814Ssimon draft-ietf-tls-compression-04.txt. 3247160814Ssimon [Richard Levitte] 3248160814Ssimon 3249160814Ssimon *) Add the ASN.1 structures and functions for CertificatePair, which 3250160814Ssimon is defined as follows (according to X.509_4thEditionDraftV6.pdf): 3251160814Ssimon 3252160814Ssimon CertificatePair ::= SEQUENCE { 3253160814Ssimon forward [0] Certificate OPTIONAL, 3254160814Ssimon reverse [1] Certificate OPTIONAL, 3255160814Ssimon -- at least one of the pair shall be present -- } 3256160814Ssimon 3257160814Ssimon Also implement the PEM functions to read and write certificate 3258160814Ssimon pairs, and defined the PEM tag as "CERTIFICATE PAIR". 3259160814Ssimon 3260160814Ssimon This needed to be defined, mostly for the sake of the LDAP 3261160814Ssimon attribute crossCertificatePair, but may prove useful elsewhere as 3262160814Ssimon well. 3263160814Ssimon [Richard Levitte] 3264160814Ssimon 3265160814Ssimon *) Make it possible to inhibit symlinking of shared libraries in 3266160814Ssimon Makefile.shared, for Cygwin's sake. 3267160814Ssimon [Richard Levitte] 3268160814Ssimon 3269160814Ssimon *) Extend the BIGNUM API by creating a function 3270160814Ssimon void BN_set_negative(BIGNUM *a, int neg); 3271160814Ssimon and a macro that behave like 3272160814Ssimon int BN_is_negative(const BIGNUM *a); 3273160814Ssimon 3274160814Ssimon to avoid the need to access 'a->neg' directly in applications. 3275160814Ssimon [Nils Larsch] 3276160814Ssimon 3277160814Ssimon *) Implement fast modular reduction for pseudo-Mersenne primes 3278160814Ssimon used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 3279160814Ssimon EC_GROUP_new_curve_GFp() will now automatically use this 3280160814Ssimon if applicable. 3281160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3282160814Ssimon 3283160814Ssimon *) Add new lock type (CRYPTO_LOCK_BN). 3284160814Ssimon [Bodo Moeller] 3285160814Ssimon 3286160814Ssimon *) Change the ENGINE framework to automatically load engines 3287160814Ssimon dynamically from specific directories unless they could be 3288160814Ssimon found to already be built in or loaded. Move all the 3289160814Ssimon current engines except for the cryptodev one to a new 3290160814Ssimon directory engines/. 3291160814Ssimon The engines in engines/ are built as shared libraries if 3292160814Ssimon the "shared" options was given to ./Configure or ./config. 3293160814Ssimon Otherwise, they are inserted in libcrypto.a. 3294160814Ssimon /usr/local/ssl/engines is the default directory for dynamic 3295160814Ssimon engines, but that can be overriden at configure time through 3296160814Ssimon the usual use of --prefix and/or --openssldir, and at run 3297160814Ssimon time with the environment variable OPENSSL_ENGINES. 3298160814Ssimon [Geoff Thorpe and Richard Levitte] 3299160814Ssimon 3300160814Ssimon *) Add Makefile.shared, a helper makefile to build shared 3301160814Ssimon libraries. Addapt Makefile.org. 3302160814Ssimon [Richard Levitte] 3303160814Ssimon 3304160814Ssimon *) Add version info to Win32 DLLs. 3305160814Ssimon [Peter 'Luna' Runestig" <peter@runestig.com>] 3306160814Ssimon 3307160814Ssimon *) Add new 'medium level' PKCS#12 API. Certificates and keys 3308160814Ssimon can be added using this API to created arbitrary PKCS#12 3309160814Ssimon files while avoiding the low level API. 3310160814Ssimon 3311160814Ssimon New options to PKCS12_create(), key or cert can be NULL and 3312160814Ssimon will then be omitted from the output file. The encryption 3313160814Ssimon algorithm NIDs can be set to -1 for no encryption, the mac 3314160814Ssimon iteration count can be set to 0 to omit the mac. 3315160814Ssimon 3316160814Ssimon Enhance pkcs12 utility by making the -nokeys and -nocerts 3317160814Ssimon options work when creating a PKCS#12 file. New option -nomac 3318160814Ssimon to omit the mac, NONE can be set for an encryption algorithm. 3319160814Ssimon New code is modified to use the enhanced PKCS12_create() 3320160814Ssimon instead of the low level API. 3321160814Ssimon [Steve Henson] 3322160814Ssimon 3323160814Ssimon *) Extend ASN1 encoder to support indefinite length constructed 3324160814Ssimon encoding. This can output sequences tags and octet strings in 3325160814Ssimon this form. Modify pk7_asn1.c to support indefinite length 3326160814Ssimon encoding. This is experimental and needs additional code to 3327160814Ssimon be useful, such as an ASN1 bio and some enhanced streaming 3328160814Ssimon PKCS#7 code. 3329160814Ssimon 3330160814Ssimon Extend template encode functionality so that tagging is passed 3331160814Ssimon down to the template encoder. 3332160814Ssimon [Steve Henson] 3333160814Ssimon 3334160814Ssimon *) Let 'openssl req' fail if an argument to '-newkey' is not 3335160814Ssimon recognized instead of using RSA as a default. 3336160814Ssimon [Bodo Moeller] 3337160814Ssimon 3338160814Ssimon *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 3339160814Ssimon As these are not official, they are not included in "ALL"; 3340160814Ssimon the "ECCdraft" ciphersuite group alias can be used to select them. 3341160814Ssimon [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 3342160814Ssimon 3343160814Ssimon *) Add ECDH engine support. 3344160814Ssimon [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 3345160814Ssimon 3346160814Ssimon *) Add ECDH in new directory crypto/ecdh/. 3347160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 3348160814Ssimon 3349160814Ssimon *) Let BN_rand_range() abort with an error after 100 iterations 3350160814Ssimon without success (which indicates a broken PRNG). 3351160814Ssimon [Bodo Moeller] 3352160814Ssimon 3353160814Ssimon *) Change BN_mod_sqrt() so that it verifies that the input value 3354160814Ssimon is really the square of the return value. (Previously, 3355160814Ssimon BN_mod_sqrt would show GIGO behaviour.) 3356160814Ssimon [Bodo Moeller] 3357160814Ssimon 3358160814Ssimon *) Add named elliptic curves over binary fields from X9.62, SECG, 3359160814Ssimon and WAP/WTLS; add OIDs that were still missing. 3360160814Ssimon 3361160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3362160814Ssimon (Sun Microsystems Laboratories)] 3363160814Ssimon 3364160814Ssimon *) Extend the EC library for elliptic curves over binary fields 3365160814Ssimon (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 3366160814Ssimon New EC_METHOD: 3367160814Ssimon 3368160814Ssimon EC_GF2m_simple_method 3369160814Ssimon 3370160814Ssimon New API functions: 3371160814Ssimon 3372160814Ssimon EC_GROUP_new_curve_GF2m 3373160814Ssimon EC_GROUP_set_curve_GF2m 3374160814Ssimon EC_GROUP_get_curve_GF2m 3375160814Ssimon EC_POINT_set_affine_coordinates_GF2m 3376160814Ssimon EC_POINT_get_affine_coordinates_GF2m 3377160814Ssimon EC_POINT_set_compressed_coordinates_GF2m 3378160814Ssimon 3379160814Ssimon Point compression for binary fields is disabled by default for 3380160814Ssimon patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 3381160814Ssimon enable it). 3382160814Ssimon 3383160814Ssimon As binary polynomials are represented as BIGNUMs, various members 3384160814Ssimon of the EC_GROUP and EC_POINT data structures can be shared 3385160814Ssimon between the implementations for prime fields and binary fields; 3386160814Ssimon the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 3387160814Ssimon are essentially identical to their ..._GFp counterparts. 3388160814Ssimon (For simplicity, the '..._GFp' prefix has been dropped from 3389160814Ssimon various internal method names.) 3390160814Ssimon 3391160814Ssimon An internal 'field_div' method (similar to 'field_mul' and 3392160814Ssimon 'field_sqr') has been added; this is used only for binary fields. 3393160814Ssimon 3394160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3395160814Ssimon (Sun Microsystems Laboratories)] 3396160814Ssimon 3397160814Ssimon *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 3398160814Ssimon through methods ('mul', 'precompute_mult'). 3399160814Ssimon 3400160814Ssimon The generic implementations (now internally called 'ec_wNAF_mul' 3401160814Ssimon and 'ec_wNAF_precomputed_mult') remain the default if these 3402160814Ssimon methods are undefined. 3403160814Ssimon 3404160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3405160814Ssimon (Sun Microsystems Laboratories)] 3406160814Ssimon 3407160814Ssimon *) New function EC_GROUP_get_degree, which is defined through 3408160814Ssimon EC_METHOD. For curves over prime fields, this returns the bit 3409160814Ssimon length of the modulus. 3410160814Ssimon 3411160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3412160814Ssimon (Sun Microsystems Laboratories)] 3413160814Ssimon 3414160814Ssimon *) New functions EC_GROUP_dup, EC_POINT_dup. 3415160814Ssimon (These simply call ..._new and ..._copy). 3416160814Ssimon 3417160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3418160814Ssimon (Sun Microsystems Laboratories)] 3419160814Ssimon 3420160814Ssimon *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 3421160814Ssimon Polynomials are represented as BIGNUMs (where the sign bit is not 3422160814Ssimon used) in the following functions [macros]: 3423160814Ssimon 3424160814Ssimon BN_GF2m_add 3425160814Ssimon BN_GF2m_sub [= BN_GF2m_add] 3426160814Ssimon BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 3427160814Ssimon BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 3428160814Ssimon BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 3429160814Ssimon BN_GF2m_mod_inv 3430160814Ssimon BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 3431160814Ssimon BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 3432160814Ssimon BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 3433160814Ssimon BN_GF2m_cmp [= BN_ucmp] 3434160814Ssimon 3435160814Ssimon (Note that only the 'mod' functions are actually for fields GF(2^m). 3436160814Ssimon BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 3437160814Ssimon 3438160814Ssimon For some functions, an the irreducible polynomial defining a 3439160814Ssimon field can be given as an 'unsigned int[]' with strictly 3440160814Ssimon decreasing elements giving the indices of those bits that are set; 3441160814Ssimon i.e., p[] represents the polynomial 3442160814Ssimon f(t) = t^p[0] + t^p[1] + ... + t^p[k] 3443160814Ssimon where 3444160814Ssimon p[0] > p[1] > ... > p[k] = 0. 3445160814Ssimon This applies to the following functions: 3446160814Ssimon 3447160814Ssimon BN_GF2m_mod_arr 3448160814Ssimon BN_GF2m_mod_mul_arr 3449160814Ssimon BN_GF2m_mod_sqr_arr 3450160814Ssimon BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 3451160814Ssimon BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 3452160814Ssimon BN_GF2m_mod_exp_arr 3453160814Ssimon BN_GF2m_mod_sqrt_arr 3454160814Ssimon BN_GF2m_mod_solve_quad_arr 3455160814Ssimon BN_GF2m_poly2arr 3456160814Ssimon BN_GF2m_arr2poly 3457160814Ssimon 3458160814Ssimon Conversion can be performed by the following functions: 3459160814Ssimon 3460160814Ssimon BN_GF2m_poly2arr 3461160814Ssimon BN_GF2m_arr2poly 3462160814Ssimon 3463160814Ssimon bntest.c has additional tests for binary polynomial arithmetic. 3464160814Ssimon 3465160814Ssimon Two implementations for BN_GF2m_mod_div() are available. 3466160814Ssimon The default algorithm simply uses BN_GF2m_mod_inv() and 3467160814Ssimon BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 3468160814Ssimon if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 3469160814Ssimon copyright notice in crypto/bn/bn_gf2m.c before enabling it). 3470160814Ssimon 3471160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3472160814Ssimon (Sun Microsystems Laboratories)] 3473160814Ssimon 3474160814Ssimon *) Add new error code 'ERR_R_DISABLED' that can be used when some 3475160814Ssimon functionality is disabled at compile-time. 3476160814Ssimon [Douglas Stebila <douglas.stebila@sun.com>] 3477160814Ssimon 3478160814Ssimon *) Change default behaviour of 'openssl asn1parse' so that more 3479160814Ssimon information is visible when viewing, e.g., a certificate: 3480160814Ssimon 3481160814Ssimon Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 3482160814Ssimon mode the content of non-printable OCTET STRINGs is output in a 3483160814Ssimon style similar to INTEGERs, but with '[HEX DUMP]' prepended to 3484160814Ssimon avoid the appearance of a printable string. 3485160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3486160814Ssimon 3487160814Ssimon *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 3488160814Ssimon functions 3489160814Ssimon EC_GROUP_set_asn1_flag() 3490160814Ssimon EC_GROUP_get_asn1_flag() 3491160814Ssimon EC_GROUP_set_point_conversion_form() 3492160814Ssimon EC_GROUP_get_point_conversion_form() 3493160814Ssimon These control ASN1 encoding details: 3494160814Ssimon - Curves (i.e., groups) are encoded explicitly unless asn1_flag 3495160814Ssimon has been set to OPENSSL_EC_NAMED_CURVE. 3496160814Ssimon - Points are encoded in uncompressed form by default; options for 3497160814Ssimon asn1_for are as for point2oct, namely 3498160814Ssimon POINT_CONVERSION_COMPRESSED 3499160814Ssimon POINT_CONVERSION_UNCOMPRESSED 3500160814Ssimon POINT_CONVERSION_HYBRID 3501160814Ssimon 3502160814Ssimon Also add 'seed' and 'seed_len' members to EC_GROUP with access 3503160814Ssimon functions 3504160814Ssimon EC_GROUP_set_seed() 3505160814Ssimon EC_GROUP_get0_seed() 3506160814Ssimon EC_GROUP_get_seed_len() 3507160814Ssimon This is used only for ASN1 purposes (so far). 3508160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3509160814Ssimon 3510160814Ssimon *) Add 'field_type' member to EC_METHOD, which holds the NID 3511160814Ssimon of the appropriate field type OID. The new function 3512160814Ssimon EC_METHOD_get_field_type() returns this value. 3513160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3514160814Ssimon 3515160814Ssimon *) Add functions 3516160814Ssimon EC_POINT_point2bn() 3517160814Ssimon EC_POINT_bn2point() 3518160814Ssimon EC_POINT_point2hex() 3519160814Ssimon EC_POINT_hex2point() 3520160814Ssimon providing useful interfaces to EC_POINT_point2oct() and 3521160814Ssimon EC_POINT_oct2point(). 3522160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3523160814Ssimon 3524160814Ssimon *) Change internals of the EC library so that the functions 3525160814Ssimon EC_GROUP_set_generator() 3526160814Ssimon EC_GROUP_get_generator() 3527160814Ssimon EC_GROUP_get_order() 3528160814Ssimon EC_GROUP_get_cofactor() 3529160814Ssimon are implemented directly in crypto/ec/ec_lib.c and not dispatched 3530160814Ssimon to methods, which would lead to unnecessary code duplication when 3531160814Ssimon adding different types of curves. 3532160814Ssimon [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 3533160814Ssimon 3534160814Ssimon *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 3535160814Ssimon arithmetic, and such that modified wNAFs are generated 3536160814Ssimon (which avoid length expansion in many cases). 3537160814Ssimon [Bodo Moeller] 3538160814Ssimon 3539160814Ssimon *) Add a function EC_GROUP_check_discriminant() (defined via 3540160814Ssimon EC_METHOD) that verifies that the curve discriminant is non-zero. 3541160814Ssimon 3542160814Ssimon Add a function EC_GROUP_check() that makes some sanity tests 3543160814Ssimon on a EC_GROUP, its generator and order. This includes 3544160814Ssimon EC_GROUP_check_discriminant(). 3545160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3546160814Ssimon 3547160814Ssimon *) Add ECDSA in new directory crypto/ecdsa/. 3548160814Ssimon 3549160814Ssimon Add applications 'openssl ecparam' and 'openssl ecdsa' 3550160814Ssimon (these are based on 'openssl dsaparam' and 'openssl dsa'). 3551160814Ssimon 3552160814Ssimon ECDSA support is also included in various other files across the 3553160814Ssimon library. Most notably, 3554160814Ssimon - 'openssl req' now has a '-newkey ecdsa:file' option; 3555160814Ssimon - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 3556160814Ssimon - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 3557160814Ssimon d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 3558160814Ssimon them suitable for ECDSA where domain parameters must be 3559160814Ssimon extracted before the specific public key; 3560160814Ssimon - ECDSA engine support has been added. 3561160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3562160814Ssimon 3563160814Ssimon *) Include some named elliptic curves, and add OIDs from X9.62, 3564160814Ssimon SECG, and WAP/WTLS. Each curve can be obtained from the new 3565160814Ssimon function 3566160814Ssimon EC_GROUP_new_by_curve_name(), 3567160814Ssimon and the list of available named curves can be obtained with 3568160814Ssimon EC_get_builtin_curves(). 3569160814Ssimon Also add a 'curve_name' member to EC_GROUP objects, which can be 3570160814Ssimon accessed via 3571160814Ssimon EC_GROUP_set_curve_name() 3572160814Ssimon EC_GROUP_get_curve_name() 3573160814Ssimon [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 3574160814Ssimon 3575160814Ssimon *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 3576160814Ssimon was actually never needed) and in BN_mul(). The removal in BN_mul() 3577160814Ssimon required a small change in bn_mul_part_recursive() and the addition 3578160814Ssimon of the functions bn_cmp_part_words(), bn_sub_part_words() and 3579160814Ssimon bn_add_part_words(), which do the same thing as bn_cmp_words(), 3580160814Ssimon bn_sub_words() and bn_add_words() except they take arrays with 3581160814Ssimon differing sizes. 3582160814Ssimon [Richard Levitte] 3583160814Ssimon 3584194206Ssimon Changes between 0.9.7l and 0.9.7m [23 Feb 2007] 3585194206Ssimon 3586167612Ssimon *) Cleanse PEM buffers before freeing them since they may contain 3587167612Ssimon sensitive data. 3588167612Ssimon [Benjamin Bennett <ben@psc.edu>] 3589167612Ssimon 3590167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 3591167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 3592167612Ssimon authentication-only ciphersuites. 3593167612Ssimon [Bodo Moeller] 3594167612Ssimon 3595167612Ssimon *) Since AES128 and AES256 share a single mask bit in the logic of 3596167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 3597167612Ssimon kludge to work properly if AES128 is available and AES256 isn't. 3598167612Ssimon [Victor Duchovni] 3599167612Ssimon 3600194206Ssimon *) Expand security boundary to match 1.1.1 module. 3601194206Ssimon [Steve Henson] 3602194206Ssimon 3603194206Ssimon *) Remove redundant features: hash file source, editing of test vectors 3604194206Ssimon modify fipsld to use external fips_premain.c signature. 3605194206Ssimon [Steve Henson] 3606194206Ssimon 3607194206Ssimon *) New perl script mkfipsscr.pl to create shell scripts or batch files to 3608194206Ssimon run algorithm test programs. 3609194206Ssimon [Steve Henson] 3610194206Ssimon 3611194206Ssimon *) Make algorithm test programs more tolerant of whitespace. 3612194206Ssimon [Steve Henson] 3613194206Ssimon 3614167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 3615167612Ssimon protocol version while receiving ClientHello even if the 3616167612Ssimon ClientHello is fragmented. (The server can't insist on the 3617167612Ssimon particular protocol version it has chosen before the ServerHello 3618167612Ssimon message has informed the client about his choice.) 3619167612Ssimon [Bodo Moeller] 3620167612Ssimon 3621167612Ssimon *) Load error codes if they are not already present instead of using a 3622167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 3623167612Ssimon [Steve Henson] 3624167612Ssimon 3625167612Ssimon Changes between 0.9.7k and 0.9.7l [28 Sep 2006] 3626167612Ssimon 3627167612Ssimon *) Introduce limits to prevent malicious keys being able to 3628167612Ssimon cause a denial of service. (CVE-2006-2940) 3629167612Ssimon [Steve Henson, Bodo Moeller] 3630167612Ssimon 3631167612Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 3632167612Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 3633167612Ssimon 3634167612Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 3635167612Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 3636167612Ssimon 3637167612Ssimon *) Fix SSL client code which could crash if connecting to a 3638167612Ssimon malicious SSLv2 server. (CVE-2006-4343) 3639167612Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 3640167612Ssimon 3641162911Ssimon *) Change ciphersuite string processing so that an explicit 3642162911Ssimon ciphersuite selects this one ciphersuite (so that "AES256-SHA" 3643162911Ssimon will no longer include "AES128-SHA"), and any other similar 3644162911Ssimon ciphersuite (same bitmap) from *other* protocol versions (so that 3645162911Ssimon "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 3646162911Ssimon SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 3647162911Ssimon changes from 0.9.8b and 0.9.8d. 3648162911Ssimon [Bodo Moeller] 3649162911Ssimon 3650162911Ssimon Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 3651162911Ssimon 3652162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 3653162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 3654162911Ssimon 3655162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 3656162911Ssimon possible instead of select(), since the latter has some 3657162911Ssimon undesirable limitations. 3658162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 3659162911Ssimon 3660162911Ssimon *) Disable rogue ciphersuites: 3661162911Ssimon 3662162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 3663162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 3664162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 3665162911Ssimon 3666162911Ssimon The latter two were purportedly from 3667162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 3668162911Ssimon appear there. 3669162911Ssimon 3670162911Ssimon Also deactive the remaining ciphersuites from 3671162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 3672162911Ssimon unofficial, and the ID has long expired. 3673162911Ssimon [Bodo Moeller] 3674162911Ssimon 3675162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 3676162911Ssimon dual-core machines) and other potential thread-safety issues. 3677162911Ssimon [Bodo Moeller] 3678162911Ssimon 3679162911Ssimon Changes between 0.9.7i and 0.9.7j [04 May 2006] 3680162911Ssimon 3681162911Ssimon *) Adapt fipsld and the build system to link against the validated FIPS 3682162911Ssimon module in FIPS mode. 3683162911Ssimon [Steve Henson] 3684162911Ssimon 3685162911Ssimon *) Fixes for VC++ 2005 build under Windows. 3686162911Ssimon [Steve Henson] 3687162911Ssimon 3688162911Ssimon *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 3689162911Ssimon from a Windows bash shell such as MSYS. It is autodetected from the 3690162911Ssimon "config" script when run from a VC++ environment. Modify standard VC++ 3691162911Ssimon build to use fipscanister.o from the GNU make build. 3692162911Ssimon [Steve Henson] 3693162911Ssimon 3694160814Ssimon Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 3695160814Ssimon 3696160814Ssimon *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 3697160814Ssimon The value now differs depending on if you build for FIPS or not. 3698160814Ssimon BEWARE! A program linked with a shared FIPSed libcrypto can't be 3699160814Ssimon safely run with a non-FIPSed libcrypto, as it may crash because of 3700160814Ssimon the difference induced by this change. 3701160814Ssimon [Andy Polyakov] 3702160814Ssimon 3703160814Ssimon Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 3704160814Ssimon 3705160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 3706160814Ssimon (part of SSL_OP_ALL). This option used to disable the 3707160814Ssimon countermeasure against man-in-the-middle protocol-version 3708160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 3709160814Ssimon idea. (CVE-2005-2969) 3710160814Ssimon 3711160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 3712160814Ssimon for Information Security, National Institute of Advanced Industrial 3713160814Ssimon Science and Technology [AIST], Japan)] 3714160814Ssimon 3715160814Ssimon *) Minimal support for X9.31 signatures and PSS padding modes. This is 3716160814Ssimon mainly for FIPS compliance and not fully integrated at this stage. 3717160814Ssimon [Steve Henson] 3718160814Ssimon 3719160814Ssimon *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 3720160814Ssimon the exponentiation using a fixed-length exponent. (Otherwise, 3721160814Ssimon the information leaked through timing could expose the secret key 3722160814Ssimon after many signatures; cf. Bleichenbacher's attack on DSA with 3723160814Ssimon biased k.) 3724160814Ssimon [Bodo Moeller] 3725160814Ssimon 3726160814Ssimon *) Make a new fixed-window mod_exp implementation the default for 3727160814Ssimon RSA, DSA, and DH private-key operations so that the sequence of 3728160814Ssimon squares and multiplies and the memory access pattern are 3729160814Ssimon independent of the particular secret key. This will mitigate 3730160814Ssimon cache-timing and potential related attacks. 3731160814Ssimon 3732160814Ssimon BN_mod_exp_mont_consttime() is the new exponentiation implementation, 3733160814Ssimon and this is automatically used by BN_mod_exp_mont() if the new flag 3734160814Ssimon BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 3735160814Ssimon will use this BN flag for private exponents unless the flag 3736160814Ssimon RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 3737160814Ssimon DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 3738160814Ssimon 3739160814Ssimon [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 3740160814Ssimon 3741160814Ssimon *) Change the client implementation for SSLv23_method() and 3742160814Ssimon SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 3743160814Ssimon Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 3744160814Ssimon (Previously, the SSL 2.0 backwards compatible Client Hello 3745160814Ssimon message format would be used even with SSL_OP_NO_SSLv2.) 3746160814Ssimon [Bodo Moeller] 3747160814Ssimon 3748160814Ssimon *) Add support for smime-type MIME parameter in S/MIME messages which some 3749160814Ssimon clients need. 3750160814Ssimon [Steve Henson] 3751160814Ssimon 3752160814Ssimon *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 3753160814Ssimon a threadsafe manner. Modify rsa code to use new function and add calls 3754160814Ssimon to dsa and dh code (which had race conditions before). 3755160814Ssimon [Steve Henson] 3756160814Ssimon 3757160814Ssimon *) Include the fixed error library code in the C error file definitions 3758160814Ssimon instead of fixing them up at runtime. This keeps the error code 3759160814Ssimon structures constant. 3760160814Ssimon [Steve Henson] 3761160814Ssimon 3762160814Ssimon Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 3763160814Ssimon 3764160814Ssimon [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 3765160814Ssimon OpenSSL 0.9.8.] 3766160814Ssimon 3767160814Ssimon *) Fixes for newer kerberos headers. NB: the casts are needed because 3768160814Ssimon the 'length' field is signed on one version and unsigned on another 3769160814Ssimon with no (?) obvious way to tell the difference, without these VC++ 3770160814Ssimon complains. Also the "definition" of FAR (blank) is no longer included 3771160814Ssimon nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 3772160814Ssimon some needed definitions. 3773160814Ssimon [Steve Henson] 3774160814Ssimon 3775160814Ssimon *) Undo Cygwin change. 3776160814Ssimon [Ulf M�ller] 3777160814Ssimon 3778160814Ssimon *) Added support for proxy certificates according to RFC 3820. 3779160814Ssimon Because they may be a security thread to unaware applications, 3780160814Ssimon they must be explicitely allowed in run-time. See 3781160814Ssimon docs/HOWTO/proxy_certificates.txt for further information. 3782160814Ssimon [Richard Levitte] 3783160814Ssimon 3784160814Ssimon Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 3785160814Ssimon 3786160814Ssimon *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 3787160814Ssimon server and client random values. Previously 3788160814Ssimon (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 3789160814Ssimon less random data when sizeof(time_t) > 4 (some 64 bit platforms). 3790160814Ssimon 3791160814Ssimon This change has negligible security impact because: 3792160814Ssimon 3793160814Ssimon 1. Server and client random values still have 24 bytes of pseudo random 3794160814Ssimon data. 3795160814Ssimon 3796160814Ssimon 2. Server and client random values are sent in the clear in the initial 3797160814Ssimon handshake. 3798160814Ssimon 3799160814Ssimon 3. The master secret is derived using the premaster secret (48 bytes in 3800160814Ssimon size for static RSA ciphersuites) as well as client server and random 3801160814Ssimon values. 3802160814Ssimon 3803160814Ssimon The OpenSSL team would like to thank the UK NISCC for bringing this issue 3804160814Ssimon to our attention. 3805160814Ssimon 3806160814Ssimon [Stephen Henson, reported by UK NISCC] 3807160814Ssimon 3808160814Ssimon *) Use Windows randomness collection on Cygwin. 3809160814Ssimon [Ulf M�ller] 3810160814Ssimon 3811160814Ssimon *) Fix hang in EGD/PRNGD query when communication socket is closed 3812160814Ssimon prematurely by EGD/PRNGD. 3813160814Ssimon [Darren Tucker <dtucker@zip.com.au> via Lutz J�nicke, resolves #1014] 3814160814Ssimon 3815160814Ssimon *) Prompt for pass phrases when appropriate for PKCS12 input format. 3816160814Ssimon [Steve Henson] 3817160814Ssimon 3818160814Ssimon *) Back-port of selected performance improvements from development 3819160814Ssimon branch, as well as improved support for PowerPC platforms. 3820160814Ssimon [Andy Polyakov] 3821160814Ssimon 3822160814Ssimon *) Add lots of checks for memory allocation failure, error codes to indicate 3823160814Ssimon failure and freeing up memory if a failure occurs. 3824160814Ssimon [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 3825160814Ssimon 3826160814Ssimon *) Add new -passin argument to dgst. 3827160814Ssimon [Steve Henson] 3828160814Ssimon 3829160814Ssimon *) Perform some character comparisons of different types in X509_NAME_cmp: 3830160814Ssimon this is needed for some certificates that reencode DNs into UTF8Strings 3831160814Ssimon (in violation of RFC3280) and can't or wont issue name rollover 3832160814Ssimon certificates. 3833160814Ssimon [Steve Henson] 3834160814Ssimon 3835160814Ssimon *) Make an explicit check during certificate validation to see that 3836160814Ssimon the CA setting in each certificate on the chain is correct. As a 3837160814Ssimon side effect always do the following basic checks on extensions, 3838160814Ssimon not just when there's an associated purpose to the check: 3839160814Ssimon 3840160814Ssimon - if there is an unhandled critical extension (unless the user 3841160814Ssimon has chosen to ignore this fault) 3842160814Ssimon - if the path length has been exceeded (if one is set at all) 3843160814Ssimon - that certain extensions fit the associated purpose (if one has 3844160814Ssimon been given) 3845160814Ssimon [Richard Levitte] 3846160814Ssimon 3847142425Snectar Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 3848142425Snectar 3849142425Snectar *) Avoid a race condition when CRLs are checked in a multi threaded 3850142425Snectar environment. This would happen due to the reordering of the revoked 3851142425Snectar entries during signature checking and serial number lookup. Now the 3852142425Snectar encoding is cached and the serial number sort performed under a lock. 3853142425Snectar Add new STACK function sk_is_sorted(). 3854142425Snectar [Steve Henson] 3855142425Snectar 3856142425Snectar *) Add Delta CRL to the extension code. 3857142425Snectar [Steve Henson] 3858142425Snectar 3859142425Snectar *) Various fixes to s3_pkt.c so alerts are sent properly. 3860142425Snectar [David Holmes <d.holmes@f5.com>] 3861142425Snectar 3862142425Snectar *) Reduce the chances of duplicate issuer name and serial numbers (in 3863142425Snectar violation of RFC3280) using the OpenSSL certificate creation utilities. 3864142425Snectar This is done by creating a random 64 bit value for the initial serial 3865142425Snectar number when a serial number file is created or when a self signed 3866142425Snectar certificate is created using 'openssl req -x509'. The initial serial 3867142425Snectar number file is created using 'openssl x509 -next_serial' in CA.pl 3868142425Snectar rather than being initialized to 1. 3869142425Snectar [Steve Henson] 3870142425Snectar 3871127128Snectar Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 3872127128Snectar 3873127128Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 3874160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 3875127128Snectar [Joe Orton, Steve Henson] 3876127128Snectar 3877127128Snectar *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 3878160814Ssimon (CVE-2004-0112) 3879127128Snectar [Joe Orton, Steve Henson] 3880127128Snectar 3881127128Snectar *) Make it possible to have multiple active certificates with the same 3882127128Snectar subject in the CA index file. This is done only if the keyword 3883127128Snectar 'unique_subject' is set to 'no' in the main CA section (default 3884127128Snectar if 'CA_default') of the configuration file. The value is saved 3885127128Snectar with the database itself in a separate index attribute file, 3886127128Snectar named like the index file with '.attr' appended to the name. 3887127128Snectar [Richard Levitte] 3888127128Snectar 3889127128Snectar *) X509 verify fixes. Disable broken certificate workarounds when 3890127128Snectar X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 3891127128Snectar keyUsage extension present. Don't accept CRLs with unhandled critical 3892127128Snectar extensions: since verify currently doesn't process CRL extensions this 3893127128Snectar rejects a CRL with *any* critical extensions. Add new verify error codes 3894127128Snectar for these cases. 3895127128Snectar [Steve Henson] 3896127128Snectar 3897127128Snectar *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 3898127128Snectar A clarification of RFC2560 will require the use of OCTET STRINGs and 3899127128Snectar some implementations cannot handle the current raw format. Since OpenSSL 3900127128Snectar copies and compares OCSP nonces as opaque blobs without any attempt at 3901127128Snectar parsing them this should not create any compatibility issues. 3902127128Snectar [Steve Henson] 3903127128Snectar 3904127128Snectar *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 3905127128Snectar calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 3906127128Snectar this HMAC (and other) operations are several times slower than OpenSSL 3907127128Snectar < 0.9.7. 3908127128Snectar [Steve Henson] 3909127128Snectar 3910127128Snectar *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 3911127128Snectar [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 3912127128Snectar 3913127128Snectar *) Use the correct content when signing type "other". 3914127128Snectar [Steve Henson] 3915127128Snectar 3916120631Snectar Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 3917120631Snectar 3918120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 3919120631Snectar 3920120631Snectar Stop out of bounds reads in the ASN1 code when presented with 3921160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 3922120631Snectar 3923160814Ssimon Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 3924120631Snectar 3925120631Snectar If verify callback ignores invalid public key errors don't try to check 3926120631Snectar certificate signature with the NULL public key. 3927120631Snectar 3928120631Snectar [Steve Henson] 3929120631Snectar 3930120631Snectar *) New -ignore_err option in ocsp application to stop the server 3931120631Snectar exiting on the first error in a request. 3932120631Snectar [Steve Henson] 3933120631Snectar 3934120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 3935120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 3936120631Snectar specifications. 3937120631Snectar [Steve Henson] 3938120631Snectar 3939120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 3940120631Snectar extra data after the compression methods not only for TLS 1.0 3941120631Snectar but also for SSL 3.0 (as required by the specification). 3942120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 3943120631Snectar 3944120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 3945120631Snectar when it's 512 *bits* long, not 512 bytes. 3946120631Snectar [Richard Levitte] 3947120631Snectar 3948120631Snectar *) Change AES_cbc_encrypt() so it outputs exact multiple of 3949120631Snectar blocks during encryption. 3950120631Snectar [Richard Levitte] 3951120631Snectar 3952120631Snectar *) Various fixes to base64 BIO and non blocking I/O. On write 3953120631Snectar flushes were not handled properly if the BIO retried. On read 3954120631Snectar data was not being buffered properly and had various logic bugs. 3955120631Snectar This also affects blocking I/O when the data being decoded is a 3956120631Snectar certain size. 3957120631Snectar [Steve Henson] 3958120631Snectar 3959120631Snectar *) Various S/MIME bugfixes and compatibility changes: 3960120631Snectar output correct application/pkcs7 MIME type if 3961120631Snectar PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 3962120631Snectar Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 3963120631Snectar of files as .eml work). Correctly handle very long lines in MIME 3964120631Snectar parser. 3965120631Snectar [Steve Henson] 3966120631Snectar 3967120631Snectar Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 3968120631Snectar 3969120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 3970120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 3971120631Snectar a protocol version number mismatch like a decryption error 3972120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 3973120631Snectar [Bodo Moeller] 3974120631Snectar 3975120631Snectar *) Turn on RSA blinding by default in the default implementation 3976120631Snectar to avoid a timing attack. Applications that don't want it can call 3977120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 3978120631Snectar They would be ill-advised to do so in most cases. 3979120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 3980120631Snectar 3981120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 3982120631Snectar seeded (in this case, the secret RSA exponent is abused as 3983120631Snectar an unpredictable seed -- if it is not unpredictable, there 3984120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 3985120631Snectar by remembering the creator's thread ID in rsa->blinding and 3986120631Snectar having all other threads use local one-time blinding factors 3987120631Snectar (this requires more computation than sharing rsa->blinding, but 3988120631Snectar avoids excessive locking; and if an RSA object is not shared 3989120631Snectar between threads, blinding will still be very fast). 3990120631Snectar [Bodo Moeller] 3991120631Snectar 3992120631Snectar *) Fixed a typo bug that would cause ENGINE_set_default() to set an 3993120631Snectar ENGINE as defaults for all supported algorithms irrespective of 3994120631Snectar the 'flags' parameter. 'flags' is now honoured, so applications 3995120631Snectar should make sure they are passing it correctly. 3996120631Snectar [Geoff Thorpe] 3997120631Snectar 3998120631Snectar *) Target "mingw" now allows native Windows code to be generated in 3999120631Snectar the Cygwin environment as well as with the MinGW compiler. 4000120631Snectar [Ulf Moeller] 4001120631Snectar 4002111147Snectar Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 4003111147Snectar 4004111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 4005111147Snectar via timing by performing a MAC computation even if incorrrect 4006111147Snectar block cipher padding has been found. This is a countermeasure 4007111147Snectar against active attacks where the attacker has to distinguish 4008160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 4009111147Snectar 4010111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 4011111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 4012111147Snectar Martin Vuagnoux (EPFL, Ilion)] 4013111147Snectar 4014111147Snectar *) Make the no-err option work as intended. The intention with no-err 4015111147Snectar is not to have the whole error stack handling routines removed from 4016111147Snectar libcrypto, it's only intended to remove all the function name and 4017111147Snectar reason texts, thereby removing some of the footprint that may not 4018111147Snectar be interesting if those errors aren't displayed anyway. 4019111147Snectar 4020111147Snectar NOTE: it's still possible for any application or module to have it's 4021111147Snectar own set of error texts inserted. The routines are there, just not 4022111147Snectar used by default when no-err is given. 4023111147Snectar [Richard Levitte] 4024111147Snectar 4025111147Snectar *) Add support for FreeBSD on IA64. 4026111147Snectar [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 4027111147Snectar 4028111147Snectar *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 4029111147Snectar Kerberos function mit_des_cbc_cksum(). Before this change, 4030111147Snectar the value returned by DES_cbc_cksum() was like the one from 4031111147Snectar mit_des_cbc_cksum(), except the bytes were swapped. 4032111147Snectar [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 4033111147Snectar 4034111147Snectar *) Allow an application to disable the automatic SSL chain building. 4035111147Snectar Before this a rather primitive chain build was always performed in 4036111147Snectar ssl3_output_cert_chain(): an application had no way to send the 4037111147Snectar correct chain if the automatic operation produced an incorrect result. 4038111147Snectar 4039111147Snectar Now the chain builder is disabled if either: 4040111147Snectar 4041111147Snectar 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 4042111147Snectar 4043111147Snectar 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 4044111147Snectar 4045111147Snectar The reasoning behind this is that an application would not want the 4046111147Snectar auto chain building to take place if extra chain certificates are 4047111147Snectar present and it might also want a means of sending no additional 4048111147Snectar certificates (for example the chain has two certificates and the 4049111147Snectar root is omitted). 4050111147Snectar [Steve Henson] 4051111147Snectar 4052111147Snectar *) Add the possibility to build without the ENGINE framework. 4053111147Snectar [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 4054111147Snectar 4055111147Snectar *) Under Win32 gmtime() can return NULL: check return value in 4056111147Snectar OPENSSL_gmtime(). Add error code for case where gmtime() fails. 4057111147Snectar [Steve Henson] 4058111147Snectar 4059111147Snectar *) DSA routines: under certain error conditions uninitialized BN objects 4060111147Snectar could be freed. Solution: make sure initialization is performed early 4061111147Snectar enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 4062111147Snectar Nils Larsch <nla@trustcenter.de> via PR#459) 4063111147Snectar [Lutz Jaenicke] 4064111147Snectar 4065111147Snectar *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 4066111147Snectar checked on reconnect on the client side, therefore session resumption 4067111147Snectar could still fail with a "ssl session id is different" error. This 4068111147Snectar behaviour is masked when SSL_OP_ALL is used due to 4069111147Snectar SSL_OP_MICROSOFT_SESS_ID_BUG being set. 4070111147Snectar Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4071111147Snectar followup to PR #377. 4072111147Snectar [Lutz Jaenicke] 4073111147Snectar 4074111147Snectar *) IA-32 assembler support enhancements: unified ELF targets, support 4075111147Snectar for SCO/Caldera platforms, fix for Cygwin shared build. 4076111147Snectar [Andy Polyakov] 4077111147Snectar 4078111147Snectar *) Add support for FreeBSD on sparc64. As a consequence, support for 4079111147Snectar FreeBSD on non-x86 processors is separate from x86 processors on 4080111147Snectar the config script, much like the NetBSD support. 4081111147Snectar [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 4082111147Snectar 4083109998Smarkm Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 4084109998Smarkm 4085120631Snectar [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 4086120631Snectar OpenSSL 0.9.7.] 4087120631Snectar 4088109998Smarkm *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 4089109998Smarkm code (06) was taken as the first octet of the session ID and the last 4090109998Smarkm octet was ignored consequently. As a result SSLv2 client side session 4091109998Smarkm caching could not have worked due to the session ID mismatch between 4092109998Smarkm client and server. 4093109998Smarkm Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4094109998Smarkm PR #377. 4095109998Smarkm [Lutz Jaenicke] 4096109998Smarkm 4097109998Smarkm *) Change the declaration of needed Kerberos libraries to use EX_LIBS 4098109998Smarkm instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 4099109998Smarkm removed entirely. 4100109998Smarkm [Richard Levitte] 4101109998Smarkm 4102109998Smarkm *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 4103109998Smarkm seems that in spite of existing for more than a year, many application 4104109998Smarkm author have done nothing to provide the necessary callbacks, which 4105109998Smarkm means that this particular engine will not work properly anywhere. 4106109998Smarkm This is a very unfortunate situation which forces us, in the name 4107109998Smarkm of usability, to give the hw_ncipher.c a static lock, which is part 4108109998Smarkm of libcrypto. 4109109998Smarkm NOTE: This is for the 0.9.7 series ONLY. This hack will never 4110109998Smarkm appear in 0.9.8 or later. We EXPECT application authors to have 4111109998Smarkm dealt properly with this when 0.9.8 is released (unless we actually 4112109998Smarkm make such changes in the libcrypto locking code that changes will 4113109998Smarkm have to be made anyway). 4114109998Smarkm [Richard Levitte] 4115109998Smarkm 4116109998Smarkm *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 4117109998Smarkm octets have been read, EOF or an error occurs. Without this change 4118109998Smarkm some truncated ASN1 structures will not produce an error. 4119109998Smarkm [Steve Henson] 4120109998Smarkm 4121109998Smarkm *) Disable Heimdal support, since it hasn't been fully implemented. 4122109998Smarkm Still give the possibility to force the use of Heimdal, but with 4123109998Smarkm warnings and a request that patches get sent to openssl-dev. 4124109998Smarkm [Richard Levitte] 4125109998Smarkm 4126109998Smarkm *) Add the VC-CE target, introduce the WINCE sysname, and add 4127109998Smarkm INSTALL.WCE and appropriate conditionals to make it build. 4128109998Smarkm [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 4129109998Smarkm 4130109998Smarkm *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 4131109998Smarkm cygssl-x.y.z.dll, where x, y and z are the major, minor and 4132109998Smarkm edit numbers of the version. 4133109998Smarkm [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 4134109998Smarkm 4135109998Smarkm *) Introduce safe string copy and catenation functions 4136109998Smarkm (BUF_strlcpy() and BUF_strlcat()). 4137109998Smarkm [Ben Laurie (CHATS) and Richard Levitte] 4138109998Smarkm 4139109998Smarkm *) Avoid using fixed-size buffers for one-line DNs. 4140109998Smarkm [Ben Laurie (CHATS)] 4141109998Smarkm 4142109998Smarkm *) Add BUF_MEM_grow_clean() to avoid information leakage when 4143109998Smarkm resizing buffers containing secrets, and use where appropriate. 4144109998Smarkm [Ben Laurie (CHATS)] 4145109998Smarkm 4146109998Smarkm *) Avoid using fixed size buffers for configuration file location. 4147109998Smarkm [Ben Laurie (CHATS)] 4148109998Smarkm 4149109998Smarkm *) Avoid filename truncation for various CA files. 4150109998Smarkm [Ben Laurie (CHATS)] 4151109998Smarkm 4152109998Smarkm *) Use sizeof in preference to magic numbers. 4153109998Smarkm [Ben Laurie (CHATS)] 4154109998Smarkm 4155109998Smarkm *) Avoid filename truncation in cert requests. 4156109998Smarkm [Ben Laurie (CHATS)] 4157109998Smarkm 4158109998Smarkm *) Add assertions to check for (supposedly impossible) buffer 4159109998Smarkm overflows. 4160109998Smarkm [Ben Laurie (CHATS)] 4161109998Smarkm 4162109998Smarkm *) Don't cache truncated DNS entries in the local cache (this could 4163109998Smarkm potentially lead to a spoofing attack). 4164109998Smarkm [Ben Laurie (CHATS)] 4165109998Smarkm 4166109998Smarkm *) Fix various buffers to be large enough for hex/decimal 4167109998Smarkm representations in a platform independent manner. 4168109998Smarkm [Ben Laurie (CHATS)] 4169109998Smarkm 4170109998Smarkm *) Add CRYPTO_realloc_clean() to avoid information leakage when 4171109998Smarkm resizing buffers containing secrets, and use where appropriate. 4172109998Smarkm [Ben Laurie (CHATS)] 4173109998Smarkm 4174109998Smarkm *) Add BIO_indent() to avoid much slightly worrying code to do 4175109998Smarkm indents. 4176109998Smarkm [Ben Laurie (CHATS)] 4177109998Smarkm 4178109998Smarkm *) Convert sprintf()/BIO_puts() to BIO_printf(). 4179109998Smarkm [Ben Laurie (CHATS)] 4180109998Smarkm 4181109998Smarkm *) buffer_gets() could terminate with the buffer only half 4182109998Smarkm full. Fixed. 4183109998Smarkm [Ben Laurie (CHATS)] 4184109998Smarkm 4185109998Smarkm *) Add assertions to prevent user-supplied crypto functions from 4186109998Smarkm overflowing internal buffers by having large block sizes, etc. 4187109998Smarkm [Ben Laurie (CHATS)] 4188109998Smarkm 4189109998Smarkm *) New OPENSSL_assert() macro (similar to assert(), but enabled 4190109998Smarkm unconditionally). 4191109998Smarkm [Ben Laurie (CHATS)] 4192109998Smarkm 4193109998Smarkm *) Eliminate unused copy of key in RC4. 4194109998Smarkm [Ben Laurie (CHATS)] 4195109998Smarkm 4196109998Smarkm *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 4197109998Smarkm [Ben Laurie (CHATS)] 4198109998Smarkm 4199109998Smarkm *) Fix off-by-one error in EGD path. 4200109998Smarkm [Ben Laurie (CHATS)] 4201109998Smarkm 4202109998Smarkm *) If RANDFILE path is too long, ignore instead of truncating. 4203109998Smarkm [Ben Laurie (CHATS)] 4204109998Smarkm 4205109998Smarkm *) Eliminate unused and incorrectly sized X.509 structure 4206109998Smarkm CBCParameter. 4207109998Smarkm [Ben Laurie (CHATS)] 4208109998Smarkm 4209109998Smarkm *) Eliminate unused and dangerous function knumber(). 4210109998Smarkm [Ben Laurie (CHATS)] 4211109998Smarkm 4212109998Smarkm *) Eliminate unused and dangerous structure, KSSL_ERR. 4213109998Smarkm [Ben Laurie (CHATS)] 4214109998Smarkm 4215109998Smarkm *) Protect against overlong session ID context length in an encoded 4216109998Smarkm session object. Since these are local, this does not appear to be 4217109998Smarkm exploitable. 4218109998Smarkm [Ben Laurie (CHATS)] 4219109998Smarkm 4220109998Smarkm *) Change from security patch (see 0.9.6e below) that did not affect 4221109998Smarkm the 0.9.6 release series: 4222109998Smarkm 4223109998Smarkm Remote buffer overflow in SSL3 protocol - an attacker could 4224109998Smarkm supply an oversized master key in Kerberos-enabled versions. 4225160814Ssimon (CVE-2002-0657) 4226109998Smarkm [Ben Laurie (CHATS)] 4227109998Smarkm 4228109998Smarkm *) Change the SSL kerb5 codes to match RFC 2712. 4229109998Smarkm [Richard Levitte] 4230109998Smarkm 4231109998Smarkm *) Make -nameopt work fully for req and add -reqopt switch. 4232109998Smarkm [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 4233109998Smarkm 4234109998Smarkm *) The "block size" for block ciphers in CFB and OFB mode should be 1. 4235109998Smarkm [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 4236109998Smarkm 4237109998Smarkm *) Make sure tests can be performed even if the corresponding algorithms 4238109998Smarkm have been removed entirely. This was also the last step to make 4239109998Smarkm OpenSSL compilable with DJGPP under all reasonable conditions. 4240109998Smarkm [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 4241109998Smarkm 4242109998Smarkm *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 4243109998Smarkm to allow version independent disabling of normally unselected ciphers, 4244109998Smarkm which may be activated as a side-effect of selecting a single cipher. 4245109998Smarkm 4246109998Smarkm (E.g., cipher list string "RSA" enables ciphersuites that are left 4247109998Smarkm out of "ALL" because they do not provide symmetric encryption. 4248109998Smarkm "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 4249109998Smarkm [Lutz Jaenicke, Bodo Moeller] 4250109998Smarkm 4251109998Smarkm *) Add appropriate support for separate platform-dependent build 4252109998Smarkm directories. The recommended way to make a platform-dependent 4253109998Smarkm build directory is the following (tested on Linux), maybe with 4254109998Smarkm some local tweaks: 4255109998Smarkm 4256109998Smarkm # Place yourself outside of the OpenSSL source tree. In 4257109998Smarkm # this example, the environment variable OPENSSL_SOURCE 4258109998Smarkm # is assumed to contain the absolute OpenSSL source directory. 4259109998Smarkm mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 4260109998Smarkm cd objtree/"`uname -s`-`uname -r`-`uname -m`" 4261111147Snectar (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 4262109998Smarkm mkdir -p `dirname $F` 4263109998Smarkm ln -s $OPENSSL_SOURCE/$F $F 4264109998Smarkm done 4265109998Smarkm 4266109998Smarkm To be absolutely sure not to disturb the source tree, a "make clean" 4267109998Smarkm is a good thing. If it isn't successfull, don't worry about it, 4268109998Smarkm it probably means the source directory is very clean. 4269109998Smarkm [Richard Levitte] 4270109998Smarkm 4271109998Smarkm *) Make sure any ENGINE control commands make local copies of string 4272109998Smarkm pointers passed to them whenever necessary. Otherwise it is possible 4273109998Smarkm the caller may have overwritten (or deallocated) the original string 4274109998Smarkm data when a later ENGINE operation tries to use the stored values. 4275109998Smarkm [G�tz Babin-Ebell <babinebell@trustcenter.de>] 4276109998Smarkm 4277109998Smarkm *) Improve diagnostics in file reading and command-line digests. 4278109998Smarkm [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 4279109998Smarkm 4280109998Smarkm *) Add AES modes CFB and OFB to the object database. Correct an 4281109998Smarkm error in AES-CFB decryption. 4282109998Smarkm [Richard Levitte] 4283109998Smarkm 4284109998Smarkm *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 4285109998Smarkm allows existing EVP_CIPHER_CTX structures to be reused after 4286109998Smarkm calling EVP_*Final(). This behaviour is used by encryption 4287109998Smarkm BIOs and some applications. This has the side effect that 4288109998Smarkm applications must explicitly clean up cipher contexts with 4289109998Smarkm EVP_CIPHER_CTX_cleanup() or they will leak memory. 4290109998Smarkm [Steve Henson] 4291109998Smarkm 4292109998Smarkm *) Check the values of dna and dnb in bn_mul_recursive before calling 4293109998Smarkm bn_mul_comba (a non zero value means the a or b arrays do not contain 4294109998Smarkm n2 elements) and fallback to bn_mul_normal if either is not zero. 4295109998Smarkm [Steve Henson] 4296109998Smarkm 4297109998Smarkm *) Fix escaping of non-ASCII characters when using the -subj option 4298109998Smarkm of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 4299109998Smarkm [Lutz Jaenicke] 4300109998Smarkm 4301109998Smarkm *) Make object definitions compliant to LDAP (RFC2256): SN is the short 4302109998Smarkm form for "surname", serialNumber has no short form. 4303109998Smarkm Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 4304109998Smarkm therefore remove "mail" short name for "internet 7". 4305109998Smarkm The OID for unique identifiers in X509 certificates is 4306109998Smarkm x500UniqueIdentifier, not uniqueIdentifier. 4307109998Smarkm Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 4308109998Smarkm [Lutz Jaenicke] 4309109998Smarkm 4310109998Smarkm *) Add an "init" command to the ENGINE config module and auto initialize 4311109998Smarkm ENGINEs. Without any "init" command the ENGINE will be initialized 4312109998Smarkm after all ctrl commands have been executed on it. If init=1 the 4313109998Smarkm ENGINE is initailized at that point (ctrls before that point are run 4314109998Smarkm on the uninitialized ENGINE and after on the initialized one). If 4315109998Smarkm init=0 then the ENGINE will not be iniatialized at all. 4316109998Smarkm [Steve Henson] 4317109998Smarkm 4318109998Smarkm *) Fix the 'app_verify_callback' interface so that the user-defined 4319109998Smarkm argument is actually passed to the callback: In the 4320109998Smarkm SSL_CTX_set_cert_verify_callback() prototype, the callback 4321109998Smarkm declaration has been changed from 4322109998Smarkm int (*cb)() 4323109998Smarkm into 4324109998Smarkm int (*cb)(X509_STORE_CTX *,void *); 4325109998Smarkm in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 4326109998Smarkm i=s->ctx->app_verify_callback(&ctx) 4327109998Smarkm has been changed into 4328109998Smarkm i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 4329109998Smarkm 4330109998Smarkm To update applications using SSL_CTX_set_cert_verify_callback(), 4331109998Smarkm a dummy argument can be added to their callback functions. 4332109998Smarkm [D. K. Smetters <smetters@parc.xerox.com>] 4333109998Smarkm 4334109998Smarkm *) Added the '4758cca' ENGINE to support IBM 4758 cards. 4335109998Smarkm [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 4336109998Smarkm 4337109998Smarkm *) Add and OPENSSL_LOAD_CONF define which will cause 4338109998Smarkm OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 4339109998Smarkm This allows older applications to transparently support certain 4340109998Smarkm OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 4341109998Smarkm Two new functions OPENSSL_add_all_algorithms_noconf() which will never 4342109998Smarkm load the config file and OPENSSL_add_all_algorithms_conf() which will 4343109998Smarkm always load it have also been added. 4344109998Smarkm [Steve Henson] 4345109998Smarkm 4346109998Smarkm *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 4347109998Smarkm Adjust NIDs and EVP layer. 4348109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4349109998Smarkm 4350109998Smarkm *) Config modules support in openssl utility. 4351109998Smarkm 4352109998Smarkm Most commands now load modules from the config file, 4353109998Smarkm though in a few (such as version) this isn't done 4354109998Smarkm because it couldn't be used for anything. 4355109998Smarkm 4356109998Smarkm In the case of ca and req the config file used is 4357109998Smarkm the same as the utility itself: that is the -config 4358109998Smarkm command line option can be used to specify an 4359109998Smarkm alternative file. 4360109998Smarkm [Steve Henson] 4361109998Smarkm 4362109998Smarkm *) Move default behaviour from OPENSSL_config(). If appname is NULL 4363109998Smarkm use "openssl_conf" if filename is NULL use default openssl config file. 4364109998Smarkm [Steve Henson] 4365109998Smarkm 4366109998Smarkm *) Add an argument to OPENSSL_config() to allow the use of an alternative 4367109998Smarkm config section name. Add a new flag to tolerate a missing config file 4368109998Smarkm and move code to CONF_modules_load_file(). 4369109998Smarkm [Steve Henson] 4370109998Smarkm 4371109998Smarkm *) Support for crypto accelerator cards from Accelerated Encryption 4372109998Smarkm Processing, www.aep.ie. (Use engine 'aep') 4373109998Smarkm The support was copied from 0.9.6c [engine] and adapted/corrected 4374109998Smarkm to work with the new engine framework. 4375109998Smarkm [AEP Inc. and Richard Levitte] 4376109998Smarkm 4377109998Smarkm *) Support for SureWare crypto accelerator cards from Baltimore 4378109998Smarkm Technologies. (Use engine 'sureware') 4379109998Smarkm The support was copied from 0.9.6c [engine] and adapted 4380109998Smarkm to work with the new engine framework. 4381109998Smarkm [Richard Levitte] 4382109998Smarkm 4383109998Smarkm *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 4384109998Smarkm make the newer ENGINE framework commands for the CHIL engine work. 4385109998Smarkm [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 4386109998Smarkm 4387109998Smarkm *) Make it possible to produce shared libraries on ReliantUNIX. 4388109998Smarkm [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 4389109998Smarkm 4390109998Smarkm *) Add the configuration target debug-linux-ppro. 4391109998Smarkm Make 'openssl rsa' use the general key loading routines 4392109998Smarkm implemented in apps.c, and make those routines able to 4393109998Smarkm handle the key format FORMAT_NETSCAPE and the variant 4394109998Smarkm FORMAT_IISSGC. 4395109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4396109998Smarkm 4397109998Smarkm *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 4398109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4399109998Smarkm 4400109998Smarkm *) Add -keyform to rsautl, and document -engine. 4401109998Smarkm [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 4402109998Smarkm 4403109998Smarkm *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 4404109998Smarkm BIO_R_NO_SUCH_FILE error code rather than the generic 4405109998Smarkm ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 4406109998Smarkm [Ben Laurie] 4407109998Smarkm 4408109998Smarkm *) Add new functions 4409109998Smarkm ERR_peek_last_error 4410109998Smarkm ERR_peek_last_error_line 4411109998Smarkm ERR_peek_last_error_line_data. 4412109998Smarkm These are similar to 4413109998Smarkm ERR_peek_error 4414109998Smarkm ERR_peek_error_line 4415109998Smarkm ERR_peek_error_line_data, 4416109998Smarkm but report on the latest error recorded rather than the first one 4417109998Smarkm still in the error queue. 4418109998Smarkm [Ben Laurie, Bodo Moeller] 4419109998Smarkm 4420109998Smarkm *) default_algorithms option in ENGINE config module. This allows things 4421109998Smarkm like: 4422109998Smarkm default_algorithms = ALL 4423109998Smarkm default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 4424109998Smarkm [Steve Henson] 4425109998Smarkm 4426109998Smarkm *) Prelminary ENGINE config module. 4427109998Smarkm [Steve Henson] 4428109998Smarkm 4429109998Smarkm *) New experimental application configuration code. 4430109998Smarkm [Steve Henson] 4431109998Smarkm 4432109998Smarkm *) Change the AES code to follow the same name structure as all other 4433109998Smarkm symmetric ciphers, and behave the same way. Move everything to 4434109998Smarkm the directory crypto/aes, thereby obsoleting crypto/rijndael. 4435109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4436109998Smarkm 4437109998Smarkm *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 4438109998Smarkm [Ben Laurie and Theo de Raadt] 4439109998Smarkm 4440109998Smarkm *) Add option to output public keys in req command. 4441109998Smarkm [Massimiliano Pala madwolf@openca.org] 4442109998Smarkm 4443109998Smarkm *) Use wNAFs in EC_POINTs_mul() for improved efficiency 4444109998Smarkm (up to about 10% better than before for P-192 and P-224). 4445109998Smarkm [Bodo Moeller] 4446109998Smarkm 4447109998Smarkm *) New functions/macros 4448109998Smarkm 4449109998Smarkm SSL_CTX_set_msg_callback(ctx, cb) 4450109998Smarkm SSL_CTX_set_msg_callback_arg(ctx, arg) 4451109998Smarkm SSL_set_msg_callback(ssl, cb) 4452109998Smarkm SSL_set_msg_callback_arg(ssl, arg) 4453109998Smarkm 4454109998Smarkm to request calling a callback function 4455109998Smarkm 4456109998Smarkm void cb(int write_p, int version, int content_type, 4457109998Smarkm const void *buf, size_t len, SSL *ssl, void *arg) 4458109998Smarkm 4459109998Smarkm whenever a protocol message has been completely received 4460109998Smarkm (write_p == 0) or sent (write_p == 1). Here 'version' is the 4461109998Smarkm protocol version according to which the SSL library interprets 4462109998Smarkm the current protocol message (SSL2_VERSION, SSL3_VERSION, or 4463109998Smarkm TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 4464109998Smarkm the content type as defined in the SSL 3.0/TLS 1.0 protocol 4465109998Smarkm specification (change_cipher_spec(20), alert(21), handshake(22)). 4466109998Smarkm 'buf' and 'len' point to the actual message, 'ssl' to the 4467109998Smarkm SSL object, and 'arg' is the application-defined value set by 4468109998Smarkm SSL[_CTX]_set_msg_callback_arg(). 4469109998Smarkm 4470109998Smarkm 'openssl s_client' and 'openssl s_server' have new '-msg' options 4471109998Smarkm to enable a callback that displays all protocol messages. 4472109998Smarkm [Bodo Moeller] 4473109998Smarkm 4474109998Smarkm *) Change the shared library support so shared libraries are built as 4475109998Smarkm soon as the corresponding static library is finished, and thereby get 4476109998Smarkm openssl and the test programs linked against the shared library. 4477109998Smarkm This still only happens when the keyword "shard" has been given to 4478109998Smarkm the configuration scripts. 4479109998Smarkm 4480109998Smarkm NOTE: shared library support is still an experimental thing, and 4481109998Smarkm backward binary compatibility is still not guaranteed. 4482109998Smarkm ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 4483109998Smarkm 4484109998Smarkm *) Add support for Subject Information Access extension. 4485109998Smarkm [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 4486109998Smarkm 4487109998Smarkm *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 4488109998Smarkm additional bytes when new memory had to be allocated, not just 4489109998Smarkm when reusing an existing buffer. 4490109998Smarkm [Bodo Moeller] 4491109998Smarkm 4492109998Smarkm *) New command line and configuration option 'utf8' for the req command. 4493109998Smarkm This allows field values to be specified as UTF8 strings. 4494109998Smarkm [Steve Henson] 4495109998Smarkm 4496109998Smarkm *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 4497109998Smarkm runs for the former and machine-readable output for the latter. 4498109998Smarkm [Ben Laurie] 4499109998Smarkm 4500109998Smarkm *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 4501109998Smarkm of the e-mail address in the DN (i.e., it will go into a certificate 4502109998Smarkm extension only). The new configuration file option 'email_in_dn = no' 4503109998Smarkm has the same effect. 4504109998Smarkm [Massimiliano Pala madwolf@openca.org] 4505109998Smarkm 4506109998Smarkm *) Change all functions with names starting with des_ to be starting 4507109998Smarkm with DES_ instead. Add wrappers that are compatible with libdes, 4508109998Smarkm but are named _ossl_old_des_*. Finally, add macros that map the 4509109998Smarkm des_* symbols to the corresponding _ossl_old_des_* if libdes 4510109998Smarkm compatibility is desired. If OpenSSL 0.9.6c compatibility is 4511109998Smarkm desired, the des_* symbols will be mapped to DES_*, with one 4512109998Smarkm exception. 4513109998Smarkm 4514109998Smarkm Since we provide two compatibility mappings, the user needs to 4515109998Smarkm define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 4516109998Smarkm compatibility is desired. The default (i.e., when that macro 4517109998Smarkm isn't defined) is OpenSSL 0.9.6c compatibility. 4518109998Smarkm 4519109998Smarkm There are also macros that enable and disable the support of old 4520109998Smarkm des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 4521109998Smarkm and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 4522109998Smarkm are defined, the default will apply: to support the old des routines. 4523109998Smarkm 4524109998Smarkm In either case, one must include openssl/des.h to get the correct 4525109998Smarkm definitions. Do not try to just include openssl/des_old.h, that 4526109998Smarkm won't work. 4527109998Smarkm 4528109998Smarkm NOTE: This is a major break of an old API into a new one. Software 4529109998Smarkm authors are encouraged to switch to the DES_ style functions. Some 4530109998Smarkm time in the future, des_old.h and the libdes compatibility functions 4531109998Smarkm will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 4532109998Smarkm default), and then completely removed. 4533109998Smarkm [Richard Levitte] 4534109998Smarkm 4535109998Smarkm *) Test for certificates which contain unsupported critical extensions. 4536109998Smarkm If such a certificate is found during a verify operation it is 4537109998Smarkm rejected by default: this behaviour can be overridden by either 4538109998Smarkm handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 4539109998Smarkm by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 4540109998Smarkm X509_supported_extension() has also been added which returns 1 if a 4541109998Smarkm particular extension is supported. 4542109998Smarkm [Steve Henson] 4543109998Smarkm 4544109998Smarkm *) Modify the behaviour of EVP cipher functions in similar way to digests 4545109998Smarkm to retain compatibility with existing code. 4546109998Smarkm [Steve Henson] 4547109998Smarkm 4548109998Smarkm *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 4549109998Smarkm compatibility with existing code. In particular the 'ctx' parameter does 4550109998Smarkm not have to be to be initialized before the call to EVP_DigestInit() and 4551109998Smarkm it is tidied up after a call to EVP_DigestFinal(). New function 4552109998Smarkm EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 4553109998Smarkm EVP_MD_CTX_copy() changed to not require the destination to be 4554109998Smarkm initialized valid and new function EVP_MD_CTX_copy_ex() added which 4555109998Smarkm requires the destination to be valid. 4556109998Smarkm 4557109998Smarkm Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 4558109998Smarkm EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 4559109998Smarkm [Steve Henson] 4560109998Smarkm 4561109998Smarkm *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 4562109998Smarkm so that complete 'Handshake' protocol structures are kept in memory 4563109998Smarkm instead of overwriting 'msg_type' and 'length' with 'body' data. 4564109998Smarkm [Bodo Moeller] 4565109998Smarkm 4566109998Smarkm *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 4567109998Smarkm [Massimo Santin via Richard Levitte] 4568109998Smarkm 4569109998Smarkm *) Major restructuring to the underlying ENGINE code. This includes 4570109998Smarkm reduction of linker bloat, separation of pure "ENGINE" manipulation 4571109998Smarkm (initialisation, etc) from functionality dealing with implementations 4572109998Smarkm of specific crypto iterfaces. This change also introduces integrated 4573109998Smarkm support for symmetric ciphers and digest implementations - so ENGINEs 4574109998Smarkm can now accelerate these by providing EVP_CIPHER and EVP_MD 4575109998Smarkm implementations of their own. This is detailed in crypto/engine/README 4576109998Smarkm as it couldn't be adequately described here. However, there are a few 4577109998Smarkm API changes worth noting - some RSA, DSA, DH, and RAND functions that 4578109998Smarkm were changed in the original introduction of ENGINE code have now 4579109998Smarkm reverted back - the hooking from this code to ENGINE is now a good 4580109998Smarkm deal more passive and at run-time, operations deal directly with 4581109998Smarkm RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 4582109998Smarkm dereferencing through an ENGINE pointer any more. Also, the ENGINE 4583109998Smarkm functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 4584109998Smarkm they were not being used by the framework as there is no concept of a 4585109998Smarkm BIGNUM_METHOD and they could not be generalised to the new 4586109998Smarkm 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 4587109998Smarkm ENGINE_cpy() has been removed as it cannot be consistently defined in 4588109998Smarkm the new code. 4589109998Smarkm [Geoff Thorpe] 4590109998Smarkm 4591109998Smarkm *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 4592109998Smarkm [Steve Henson] 4593109998Smarkm 4594109998Smarkm *) Change mkdef.pl to sort symbols that get the same entry number, 4595109998Smarkm and make sure the automatically generated functions ERR_load_* 4596109998Smarkm become part of libeay.num as well. 4597109998Smarkm [Richard Levitte] 4598109998Smarkm 4599109998Smarkm *) New function SSL_renegotiate_pending(). This returns true once 4600109998Smarkm renegotiation has been requested (either SSL_renegotiate() call 4601109998Smarkm or HelloRequest/ClientHello receveived from the peer) and becomes 4602109998Smarkm false once a handshake has been completed. 4603109998Smarkm (For servers, SSL_renegotiate() followed by SSL_do_handshake() 4604109998Smarkm sends a HelloRequest, but does not ensure that a handshake takes 4605109998Smarkm place. SSL_renegotiate_pending() is useful for checking if the 4606109998Smarkm client has followed the request.) 4607109998Smarkm [Bodo Moeller] 4608109998Smarkm 4609109998Smarkm *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 4610109998Smarkm By default, clients may request session resumption even during 4611109998Smarkm renegotiation (if session ID contexts permit); with this option, 4612109998Smarkm session resumption is possible only in the first handshake. 4613109998Smarkm 4614109998Smarkm SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 4615109998Smarkm more bits available for options that should not be part of 4616109998Smarkm SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 4617109998Smarkm [Bodo Moeller] 4618109998Smarkm 4619109998Smarkm *) Add some demos for certificate and certificate request creation. 4620109998Smarkm [Steve Henson] 4621109998Smarkm 4622109998Smarkm *) Make maximum certificate chain size accepted from the peer application 4623109998Smarkm settable (SSL*_get/set_max_cert_list()), as proposed by 4624109998Smarkm "Douglas E. Engert" <deengert@anl.gov>. 4625109998Smarkm [Lutz Jaenicke] 4626109998Smarkm 4627109998Smarkm *) Add support for shared libraries for Unixware-7 4628109998Smarkm (Boyd Lynn Gerber <gerberb@zenez.com>). 4629109998Smarkm [Lutz Jaenicke] 4630109998Smarkm 4631109998Smarkm *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 4632109998Smarkm be done prior to destruction. Use this to unload error strings from 4633109998Smarkm ENGINEs that load their own error strings. NB: This adds two new API 4634109998Smarkm functions to "get" and "set" this destroy handler in an ENGINE. 4635109998Smarkm [Geoff Thorpe] 4636109998Smarkm 4637109998Smarkm *) Alter all existing ENGINE implementations (except "openssl" and 4638109998Smarkm "openbsd") to dynamically instantiate their own error strings. This 4639109998Smarkm makes them more flexible to be built both as statically-linked ENGINEs 4640109998Smarkm and self-contained shared-libraries loadable via the "dynamic" ENGINE. 4641109998Smarkm Also, add stub code to each that makes building them as self-contained 4642109998Smarkm shared-libraries easier (see README.ENGINE). 4643109998Smarkm [Geoff Thorpe] 4644109998Smarkm 4645109998Smarkm *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 4646109998Smarkm implementations into applications that are completely implemented in 4647109998Smarkm self-contained shared-libraries. The "dynamic" ENGINE exposes control 4648109998Smarkm commands that can be used to configure what shared-library to load and 4649109998Smarkm to control aspects of the way it is handled. Also, made an update to 4650109998Smarkm the README.ENGINE file that brings its information up-to-date and 4651109998Smarkm provides some information and instructions on the "dynamic" ENGINE 4652109998Smarkm (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 4653109998Smarkm [Geoff Thorpe] 4654109998Smarkm 4655109998Smarkm *) Make it possible to unload ranges of ERR strings with a new 4656109998Smarkm "ERR_unload_strings" function. 4657109998Smarkm [Geoff Thorpe] 4658109998Smarkm 4659109998Smarkm *) Add a copy() function to EVP_MD. 4660109998Smarkm [Ben Laurie] 4661109998Smarkm 4662109998Smarkm *) Make EVP_MD routines take a context pointer instead of just the 4663109998Smarkm md_data void pointer. 4664109998Smarkm [Ben Laurie] 4665109998Smarkm 4666109998Smarkm *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 4667109998Smarkm that the digest can only process a single chunk of data 4668109998Smarkm (typically because it is provided by a piece of 4669109998Smarkm hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 4670109998Smarkm is only going to provide a single chunk of data, and hence the 4671109998Smarkm framework needn't accumulate the data for oneshot drivers. 4672109998Smarkm [Ben Laurie] 4673109998Smarkm 4674109998Smarkm *) As with "ERR", make it possible to replace the underlying "ex_data" 4675109998Smarkm functions. This change also alters the storage and management of global 4676109998Smarkm ex_data state - it's now all inside ex_data.c and all "class" code (eg. 4677109998Smarkm RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 4678109998Smarkm index counters. The API functions that use this state have been changed 4679109998Smarkm to take a "class_index" rather than pointers to the class's local STACK 4680109998Smarkm and counter, and there is now an API function to dynamically create new 4681109998Smarkm classes. This centralisation allows us to (a) plug a lot of the 4682109998Smarkm thread-safety problems that existed, and (b) makes it possible to clean 4683109998Smarkm up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 4684109998Smarkm such data would previously have always leaked in application code and 4685109998Smarkm workarounds were in place to make the memory debugging turn a blind eye 4686109998Smarkm to it. Application code that doesn't use this new function will still 4687109998Smarkm leak as before, but their memory debugging output will announce it now 4688109998Smarkm rather than letting it slide. 4689109998Smarkm 4690109998Smarkm Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 4691109998Smarkm induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 4692109998Smarkm has a return value to indicate success or failure. 4693109998Smarkm [Geoff Thorpe] 4694109998Smarkm 4695109998Smarkm *) Make it possible to replace the underlying "ERR" functions such that the 4696109998Smarkm global state (2 LHASH tables and 2 locks) is only used by the "default" 4697109998Smarkm implementation. This change also adds two functions to "get" and "set" 4698109998Smarkm the implementation prior to it being automatically set the first time 4699109998Smarkm any other ERR function takes place. Ie. an application can call "get", 4700109998Smarkm pass the return value to a module it has just loaded, and that module 4701109998Smarkm can call its own "set" function using that value. This means the 4702109998Smarkm module's "ERR" operations will use (and modify) the error state in the 4703109998Smarkm application and not in its own statically linked copy of OpenSSL code. 4704109998Smarkm [Geoff Thorpe] 4705109998Smarkm 4706109998Smarkm *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 4707109998Smarkm reference counts. This performs normal REF_PRINT/REF_CHECK macros on 4708109998Smarkm the operation, and provides a more encapsulated way for external code 4709109998Smarkm (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 4710109998Smarkm to use these functions rather than manually incrementing the counts. 4711109998Smarkm 4712109998Smarkm Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 4713109998Smarkm [Geoff Thorpe] 4714109998Smarkm 4715109998Smarkm *) Add EVP test program. 4716109998Smarkm [Ben Laurie] 4717109998Smarkm 4718109998Smarkm *) Add symmetric cipher support to ENGINE. Expect the API to change! 4719109998Smarkm [Ben Laurie] 4720109998Smarkm 4721109998Smarkm *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 4722109998Smarkm X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 4723109998Smarkm X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 4724109998Smarkm These allow a CRL to be built without having to access X509_CRL fields 4725109998Smarkm directly. Modify 'ca' application to use new functions. 4726109998Smarkm [Steve Henson] 4727109998Smarkm 4728109998Smarkm *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 4729109998Smarkm bug workarounds. Rollback attack detection is a security feature. 4730109998Smarkm The problem will only arise on OpenSSL servers when TLSv1 is not 4731109998Smarkm available (sslv3_server_method() or SSL_OP_NO_TLSv1). 4732109998Smarkm Software authors not wanting to support TLSv1 will have special reasons 4733109998Smarkm for their choice and can explicitly enable this option. 4734109998Smarkm [Bodo Moeller, Lutz Jaenicke] 4735109998Smarkm 4736109998Smarkm *) Rationalise EVP so it can be extended: don't include a union of 4737109998Smarkm cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 4738109998Smarkm (similar to those existing for EVP_CIPHER_CTX). 4739109998Smarkm Usage example: 4740109998Smarkm 4741109998Smarkm EVP_MD_CTX md; 4742109998Smarkm 4743109998Smarkm EVP_MD_CTX_init(&md); /* new function call */ 4744109998Smarkm EVP_DigestInit(&md, EVP_sha1()); 4745109998Smarkm EVP_DigestUpdate(&md, in, len); 4746109998Smarkm EVP_DigestFinal(&md, out, NULL); 4747109998Smarkm EVP_MD_CTX_cleanup(&md); /* new function call */ 4748109998Smarkm 4749109998Smarkm [Ben Laurie] 4750109998Smarkm 4751109998Smarkm *) Make DES key schedule conform to the usual scheme, as well as 4752109998Smarkm correcting its structure. This means that calls to DES functions 4753109998Smarkm now have to pass a pointer to a des_key_schedule instead of a 4754109998Smarkm plain des_key_schedule (which was actually always a pointer 4755109998Smarkm anyway): E.g., 4756109998Smarkm 4757109998Smarkm des_key_schedule ks; 4758109998Smarkm 4759109998Smarkm des_set_key_checked(..., &ks); 4760109998Smarkm des_ncbc_encrypt(..., &ks, ...); 4761109998Smarkm 4762109998Smarkm (Note that a later change renames 'des_...' into 'DES_...'.) 4763109998Smarkm [Ben Laurie] 4764109998Smarkm 4765109998Smarkm *) Initial reduction of linker bloat: the use of some functions, such as 4766109998Smarkm PEM causes large amounts of unused functions to be linked in due to 4767109998Smarkm poor organisation. For example pem_all.c contains every PEM function 4768109998Smarkm which has a knock on effect of linking in large amounts of (unused) 4769109998Smarkm ASN1 code. Grouping together similar functions and splitting unrelated 4770109998Smarkm functions prevents this. 4771109998Smarkm [Steve Henson] 4772109998Smarkm 4773109998Smarkm *) Cleanup of EVP macros. 4774109998Smarkm [Ben Laurie] 4775109998Smarkm 4776109998Smarkm *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 4777109998Smarkm correct _ecb suffix. 4778109998Smarkm [Ben Laurie] 4779109998Smarkm 4780109998Smarkm *) Add initial OCSP responder support to ocsp application. The 4781109998Smarkm revocation information is handled using the text based index 4782109998Smarkm use by the ca application. The responder can either handle 4783109998Smarkm requests generated internally, supplied in files (for example 4784109998Smarkm via a CGI script) or using an internal minimal server. 4785109998Smarkm [Steve Henson] 4786109998Smarkm 4787109998Smarkm *) Add configuration choices to get zlib compression for TLS. 4788109998Smarkm [Richard Levitte] 4789109998Smarkm 4790109998Smarkm *) Changes to Kerberos SSL for RFC 2712 compliance: 4791109998Smarkm 1. Implemented real KerberosWrapper, instead of just using 4792109998Smarkm KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 4793109998Smarkm 2. Implemented optional authenticator field of KerberosWrapper. 4794109998Smarkm 4795109998Smarkm Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 4796109998Smarkm and authenticator structs; see crypto/krb5/. 4797109998Smarkm 4798109998Smarkm Generalized Kerberos calls to support multiple Kerberos libraries. 4799109998Smarkm [Vern Staats <staatsvr@asc.hpc.mil>, 4800109998Smarkm Jeffrey Altman <jaltman@columbia.edu> 4801109998Smarkm via Richard Levitte] 4802109998Smarkm 4803109998Smarkm *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 4804109998Smarkm already does with RSA. testdsa.h now has 'priv_key/pub_key' 4805109998Smarkm values for each of the key sizes rather than having just 4806109998Smarkm parameters (and 'speed' generating keys each time). 4807109998Smarkm [Geoff Thorpe] 4808109998Smarkm 4809109998Smarkm *) Speed up EVP routines. 4810109998Smarkm Before: 4811109998Smarkmencrypt 4812109998Smarkmtype 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 4813109998Smarkmdes-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 4814109998Smarkmdes-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 4815109998Smarkmdes-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 4816109998Smarkmdecrypt 4817109998Smarkmdes-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 4818109998Smarkmdes-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 4819109998Smarkmdes-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 4820109998Smarkm After: 4821109998Smarkmencrypt 4822109998Smarkmdes-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 4823109998Smarkmdecrypt 4824109998Smarkmdes-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 4825109998Smarkm [Ben Laurie] 4826109998Smarkm 4827109998Smarkm *) Added the OS2-EMX target. 4828109998Smarkm ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 4829109998Smarkm 4830109998Smarkm *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 4831109998Smarkm to support NCONF routines in extension code. New function CONF_set_nconf() 4832109998Smarkm to allow functions which take an NCONF to also handle the old LHASH 4833109998Smarkm structure: this means that the old CONF compatible routines can be 4834109998Smarkm retained (in particular wrt extensions) without having to duplicate the 4835109998Smarkm code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 4836109998Smarkm [Steve Henson] 4837109998Smarkm 4838109998Smarkm *) Enhance the general user interface with mechanisms for inner control 4839109998Smarkm and with possibilities to have yes/no kind of prompts. 4840109998Smarkm [Richard Levitte] 4841109998Smarkm 4842109998Smarkm *) Change all calls to low level digest routines in the library and 4843109998Smarkm applications to use EVP. Add missing calls to HMAC_cleanup() and 4844109998Smarkm don't assume HMAC_CTX can be copied using memcpy(). 4845109998Smarkm [Verdon Walker <VWalker@novell.com>, Steve Henson] 4846109998Smarkm 4847109998Smarkm *) Add the possibility to control engines through control names but with 4848109998Smarkm arbitrary arguments instead of just a string. 4849109998Smarkm Change the key loaders to take a UI_METHOD instead of a callback 4850109998Smarkm function pointer. NOTE: this breaks binary compatibility with earlier 4851109998Smarkm versions of OpenSSL [engine]. 4852109998Smarkm Adapt the nCipher code for these new conditions and add a card insertion 4853109998Smarkm callback. 4854109998Smarkm [Richard Levitte] 4855109998Smarkm 4856109998Smarkm *) Enhance the general user interface with mechanisms to better support 4857109998Smarkm dialog box interfaces, application-defined prompts, the possibility 4858109998Smarkm to use defaults (for example default passwords from somewhere else) 4859109998Smarkm and interrupts/cancellations. 4860109998Smarkm [Richard Levitte] 4861109998Smarkm 4862109998Smarkm *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 4863109998Smarkm attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 4864109998Smarkm [Steve Henson] 4865109998Smarkm 4866109998Smarkm *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 4867109998Smarkm tidy up some unnecessarily weird code in 'sk_new()'). 4868109998Smarkm [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 4869109998Smarkm 4870109998Smarkm *) Change the key loading routines for ENGINEs to use the same kind 4871109998Smarkm callback (pem_password_cb) as all other routines that need this 4872109998Smarkm kind of callback. 4873109998Smarkm [Richard Levitte] 4874109998Smarkm 4875109998Smarkm *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 4876109998Smarkm 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 4877109998Smarkm than this minimum value is recommended. 4878109998Smarkm [Lutz Jaenicke] 4879109998Smarkm 4880109998Smarkm *) New random seeder for OpenVMS, using the system process statistics 4881109998Smarkm that are easily reachable. 4882109998Smarkm [Richard Levitte] 4883109998Smarkm 4884109998Smarkm *) Windows apparently can't transparently handle global 4885109998Smarkm variables defined in DLLs. Initialisations such as: 4886109998Smarkm 4887109998Smarkm const ASN1_ITEM *it = &ASN1_INTEGER_it; 4888109998Smarkm 4889109998Smarkm wont compile. This is used by the any applications that need to 4890109998Smarkm declare their own ASN1 modules. This was fixed by adding the option 4891109998Smarkm EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 4892109998Smarkm needed for static libraries under Win32. 4893109998Smarkm [Steve Henson] 4894109998Smarkm 4895109998Smarkm *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 4896109998Smarkm setting of purpose and trust fields. New X509_STORE trust and 4897109998Smarkm purpose functions and tidy up setting in other SSL functions. 4898109998Smarkm [Steve Henson] 4899109998Smarkm 4900109998Smarkm *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 4901109998Smarkm structure. These are inherited by X509_STORE_CTX when it is 4902109998Smarkm initialised. This allows various defaults to be set in the 4903109998Smarkm X509_STORE structure (such as flags for CRL checking and custom 4904109998Smarkm purpose or trust settings) for functions which only use X509_STORE_CTX 4905109998Smarkm internally such as S/MIME. 4906109998Smarkm 4907109998Smarkm Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 4908109998Smarkm trust settings if they are not set in X509_STORE. This allows X509_STORE 4909109998Smarkm purposes and trust (in S/MIME for example) to override any set by default. 4910109998Smarkm 4911109998Smarkm Add command line options for CRL checking to smime, s_client and s_server 4912109998Smarkm applications. 4913109998Smarkm [Steve Henson] 4914109998Smarkm 4915109998Smarkm *) Initial CRL based revocation checking. If the CRL checking flag(s) 4916109998Smarkm are set then the CRL is looked up in the X509_STORE structure and 4917109998Smarkm its validity and signature checked, then if the certificate is found 4918109998Smarkm in the CRL the verify fails with a revoked error. 4919109998Smarkm 4920109998Smarkm Various new CRL related callbacks added to X509_STORE_CTX structure. 4921109998Smarkm 4922109998Smarkm Command line options added to 'verify' application to support this. 4923109998Smarkm 4924109998Smarkm This needs some additional work, such as being able to handle multiple 4925109998Smarkm CRLs with different times, extension based lookup (rather than just 4926109998Smarkm by subject name) and ultimately more complete V2 CRL extension 4927109998Smarkm handling. 4928109998Smarkm [Steve Henson] 4929109998Smarkm 4930109998Smarkm *) Add a general user interface API (crypto/ui/). This is designed 4931109998Smarkm to replace things like des_read_password and friends (backward 4932109998Smarkm compatibility functions using this new API are provided). 4933109998Smarkm The purpose is to remove prompting functions from the DES code 4934109998Smarkm section as well as provide for prompting through dialog boxes in 4935109998Smarkm a window system and the like. 4936109998Smarkm [Richard Levitte] 4937109998Smarkm 4938109998Smarkm *) Add "ex_data" support to ENGINE so implementations can add state at a 4939109998Smarkm per-structure level rather than having to store it globally. 4940109998Smarkm [Geoff] 4941109998Smarkm 4942109998Smarkm *) Make it possible for ENGINE structures to be copied when retrieved by 4943109998Smarkm ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 4944109998Smarkm This causes the "original" ENGINE structure to act like a template, 4945109998Smarkm analogous to the RSA vs. RSA_METHOD type of separation. Because of this 4946109998Smarkm operational state can be localised to each ENGINE structure, despite the 4947109998Smarkm fact they all share the same "methods". New ENGINE structures returned in 4948109998Smarkm this case have no functional references and the return value is the single 4949109998Smarkm structural reference. This matches the single structural reference returned 4950109998Smarkm by ENGINE_by_id() normally, when it is incremented on the pre-existing 4951109998Smarkm ENGINE structure. 4952109998Smarkm [Geoff] 4953109998Smarkm 4954109998Smarkm *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 4955109998Smarkm needs to match any other type at all we need to manually clear the 4956109998Smarkm tag cache. 4957109998Smarkm [Steve Henson] 4958109998Smarkm 4959109998Smarkm *) Changes to the "openssl engine" utility to include; 4960109998Smarkm - verbosity levels ('-v', '-vv', and '-vvv') that provide information 4961109998Smarkm about an ENGINE's available control commands. 4962109998Smarkm - executing control commands from command line arguments using the 4963109998Smarkm '-pre' and '-post' switches. '-post' is only used if '-t' is 4964109998Smarkm specified and the ENGINE is successfully initialised. The syntax for 4965109998Smarkm the individual commands are colon-separated, for example; 4966109998Smarkm openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 4967109998Smarkm [Geoff] 4968109998Smarkm 4969109998Smarkm *) New dynamic control command support for ENGINEs. ENGINEs can now 4970109998Smarkm declare their own commands (numbers), names (strings), descriptions, 4971109998Smarkm and input types for run-time discovery by calling applications. A 4972109998Smarkm subset of these commands are implicitly classed as "executable" 4973109998Smarkm depending on their input type, and only these can be invoked through 4974109998Smarkm the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 4975109998Smarkm can be based on user input, config files, etc). The distinction is 4976109998Smarkm that "executable" commands cannot return anything other than a boolean 4977109998Smarkm result and can only support numeric or string input, whereas some 4978109998Smarkm discoverable commands may only be for direct use through 4979109998Smarkm ENGINE_ctrl(), eg. supporting the exchange of binary data, function 4980109998Smarkm pointers, or other custom uses. The "executable" commands are to 4981109998Smarkm support parameterisations of ENGINE behaviour that can be 4982109998Smarkm unambiguously defined by ENGINEs and used consistently across any 4983109998Smarkm OpenSSL-based application. Commands have been added to all the 4984109998Smarkm existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 4985109998Smarkm control over shared-library paths without source code alterations. 4986109998Smarkm [Geoff] 4987109998Smarkm 4988109998Smarkm *) Changed all ENGINE implementations to dynamically allocate their 4989109998Smarkm ENGINEs rather than declaring them statically. Apart from this being 4990109998Smarkm necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 4991109998Smarkm this also allows the implementations to compile without using the 4992109998Smarkm internal engine_int.h header. 4993109998Smarkm [Geoff] 4994109998Smarkm 4995109998Smarkm *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 4996109998Smarkm 'const' value. Any code that should be able to modify a RAND_METHOD 4997109998Smarkm should already have non-const pointers to it (ie. they should only 4998109998Smarkm modify their own ones). 4999109998Smarkm [Geoff] 5000109998Smarkm 5001109998Smarkm *) Made a variety of little tweaks to the ENGINE code. 5002109998Smarkm - "atalla" and "ubsec" string definitions were moved from header files 5003109998Smarkm to C code. "nuron" string definitions were placed in variables 5004109998Smarkm rather than hard-coded - allowing parameterisation of these values 5005109998Smarkm later on via ctrl() commands. 5006109998Smarkm - Removed unused "#if 0"'d code. 5007109998Smarkm - Fixed engine list iteration code so it uses ENGINE_free() to release 5008109998Smarkm structural references. 5009109998Smarkm - Constified the RAND_METHOD element of ENGINE structures. 5010109998Smarkm - Constified various get/set functions as appropriate and added 5011109998Smarkm missing functions (including a catch-all ENGINE_cpy that duplicates 5012109998Smarkm all ENGINE values onto a new ENGINE except reference counts/state). 5013109998Smarkm - Removed NULL parameter checks in get/set functions. Setting a method 5014109998Smarkm or function to NULL is a way of cancelling out a previously set 5015109998Smarkm value. Passing a NULL ENGINE parameter is just plain stupid anyway 5016109998Smarkm and doesn't justify the extra error symbols and code. 5017109998Smarkm - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 5018109998Smarkm flags from engine_int.h to engine.h. 5019109998Smarkm - Changed prototypes for ENGINE handler functions (init(), finish(), 5020109998Smarkm ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 5021109998Smarkm [Geoff] 5022109998Smarkm 5023109998Smarkm *) Implement binary inversion algorithm for BN_mod_inverse in addition 5024109998Smarkm to the algorithm using long division. The binary algorithm can be 5025109998Smarkm used only if the modulus is odd. On 32-bit systems, it is faster 5026109998Smarkm only for relatively small moduli (roughly 20-30% for 128-bit moduli, 5027109998Smarkm roughly 5-15% for 256-bit moduli), so we use it only for moduli 5028109998Smarkm up to 450 bits. In 64-bit environments, the binary algorithm 5029109998Smarkm appears to be advantageous for much longer moduli; here we use it 5030109998Smarkm for moduli up to 2048 bits. 5031109998Smarkm [Bodo Moeller] 5032109998Smarkm 5033109998Smarkm *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 5034109998Smarkm could not support the combine flag in choice fields. 5035109998Smarkm [Steve Henson] 5036109998Smarkm 5037109998Smarkm *) Add a 'copy_extensions' option to the 'ca' utility. This copies 5038109998Smarkm extensions from a certificate request to the certificate. 5039109998Smarkm [Steve Henson] 5040109998Smarkm 5041109998Smarkm *) Allow multiple 'certopt' and 'nameopt' options to be separated 5042109998Smarkm by commas. Add 'namopt' and 'certopt' options to the 'ca' config 5043109998Smarkm file: this allows the display of the certificate about to be 5044109998Smarkm signed to be customised, to allow certain fields to be included 5045109998Smarkm or excluded and extension details. The old system didn't display 5046109998Smarkm multicharacter strings properly, omitted fields not in the policy 5047109998Smarkm and couldn't display additional details such as extensions. 5048109998Smarkm [Steve Henson] 5049109998Smarkm 5050109998Smarkm *) Function EC_POINTs_mul for multiple scalar multiplication 5051109998Smarkm of an arbitrary number of elliptic curve points 5052109998Smarkm \sum scalars[i]*points[i], 5053109998Smarkm optionally including the generator defined for the EC_GROUP: 5054109998Smarkm scalar*generator + \sum scalars[i]*points[i]. 5055109998Smarkm 5056109998Smarkm EC_POINT_mul is a simple wrapper function for the typical case 5057109998Smarkm that the point list has just one item (besides the optional 5058109998Smarkm generator). 5059109998Smarkm [Bodo Moeller] 5060109998Smarkm 5061109998Smarkm *) First EC_METHODs for curves over GF(p): 5062109998Smarkm 5063109998Smarkm EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 5064109998Smarkm operations and provides various method functions that can also 5065109998Smarkm operate with faster implementations of modular arithmetic. 5066109998Smarkm 5067109998Smarkm EC_GFp_mont_method() reuses most functions that are part of 5068109998Smarkm EC_GFp_simple_method, but uses Montgomery arithmetic. 5069109998Smarkm 5070109998Smarkm [Bodo Moeller; point addition and point doubling 5071109998Smarkm implementation directly derived from source code provided by 5072109998Smarkm Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 5073109998Smarkm 5074109998Smarkm *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 5075109998Smarkm crypto/ec/ec_lib.c): 5076109998Smarkm 5077109998Smarkm Curves are EC_GROUP objects (with an optional group generator) 5078109998Smarkm based on EC_METHODs that are built into the library. 5079109998Smarkm 5080109998Smarkm Points are EC_POINT objects based on EC_GROUP objects. 5081109998Smarkm 5082109998Smarkm Most of the framework would be able to handle curves over arbitrary 5083109998Smarkm finite fields, but as there are no obvious types for fields other 5084109998Smarkm than GF(p), some functions are limited to that for now. 5085109998Smarkm [Bodo Moeller] 5086109998Smarkm 5087109998Smarkm *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 5088109998Smarkm that the file contains a complete HTTP response. 5089109998Smarkm [Richard Levitte] 5090109998Smarkm 5091109998Smarkm *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 5092109998Smarkm change the def and num file printf format specifier from "%-40sXXX" 5093109998Smarkm to "%-39s XXX". The latter will always guarantee a space after the 5094109998Smarkm field while the former will cause them to run together if the field 5095109998Smarkm is 40 of more characters long. 5096109998Smarkm [Steve Henson] 5097109998Smarkm 5098109998Smarkm *) Constify the cipher and digest 'method' functions and structures 5099109998Smarkm and modify related functions to take constant EVP_MD and EVP_CIPHER 5100109998Smarkm pointers. 5101109998Smarkm [Steve Henson] 5102109998Smarkm 5103109998Smarkm *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 5104109998Smarkm in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 5105109998Smarkm [Bodo Moeller] 5106109998Smarkm 5107109998Smarkm *) Modify EVP_Digest*() routines so they now return values. Although the 5108109998Smarkm internal software routines can never fail additional hardware versions 5109109998Smarkm might. 5110109998Smarkm [Steve Henson] 5111109998Smarkm 5112109998Smarkm *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 5113109998Smarkm 5114109998Smarkm Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 5115109998Smarkm (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 5116109998Smarkm 5117109998Smarkm ASN1 error codes 5118109998Smarkm ERR_R_NESTED_ASN1_ERROR 5119109998Smarkm ... 5120109998Smarkm ERR_R_MISSING_ASN1_EOS 5121109998Smarkm were 4 .. 9, conflicting with 5122109998Smarkm ERR_LIB_RSA (= ERR_R_RSA_LIB) 5123109998Smarkm ... 5124109998Smarkm ERR_LIB_PEM (= ERR_R_PEM_LIB). 5125109998Smarkm They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 5126109998Smarkm 5127109998Smarkm Add new error code 'ERR_R_INTERNAL_ERROR'. 5128109998Smarkm [Bodo Moeller] 5129109998Smarkm 5130109998Smarkm *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 5131109998Smarkm suffices. 5132109998Smarkm [Bodo Moeller] 5133109998Smarkm 5134109998Smarkm *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 5135109998Smarkm sets the subject name for a new request or supersedes the 5136109998Smarkm subject name in a given request. Formats that can be parsed are 5137109998Smarkm 'CN=Some Name, OU=myOU, C=IT' 5138109998Smarkm and 5139109998Smarkm 'CN=Some Name/OU=myOU/C=IT'. 5140109998Smarkm 5141109998Smarkm Add options '-batch' and '-verbose' to 'openssl req'. 5142109998Smarkm [Massimiliano Pala <madwolf@hackmasters.net>] 5143109998Smarkm 5144109998Smarkm *) Introduce the possibility to access global variables through 5145109998Smarkm functions on platform were that's the best way to handle exporting 5146109998Smarkm global variables in shared libraries. To enable this functionality, 5147109998Smarkm one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 5148109998Smarkm "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 5149109998Smarkm is normally done by Configure or something similar). 5150109998Smarkm 5151109998Smarkm To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 5152109998Smarkm in the source file (foo.c) like this: 5153109998Smarkm 5154109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 5155109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(double,bar); 5156109998Smarkm 5157109998Smarkm To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 5158109998Smarkm and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 5159109998Smarkm 5160109998Smarkm OPENSSL_DECLARE_GLOBAL(int,foo); 5161109998Smarkm #define foo OPENSSL_GLOBAL_REF(foo) 5162109998Smarkm OPENSSL_DECLARE_GLOBAL(double,bar); 5163109998Smarkm #define bar OPENSSL_GLOBAL_REF(bar) 5164109998Smarkm 5165109998Smarkm The #defines are very important, and therefore so is including the 5166109998Smarkm header file everywhere where the defined globals are used. 5167109998Smarkm 5168109998Smarkm The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 5169109998Smarkm of ASN.1 items, but that structure is a bit different. 5170109998Smarkm 5171109998Smarkm The largest change is in util/mkdef.pl which has been enhanced with 5172109998Smarkm better and easier to understand logic to choose which symbols should 5173109998Smarkm go into the Windows .def files as well as a number of fixes and code 5174109998Smarkm cleanup (among others, algorithm keywords are now sorted 5175109998Smarkm lexicographically to avoid constant rewrites). 5176109998Smarkm [Richard Levitte] 5177109998Smarkm 5178109998Smarkm *) In BN_div() keep a copy of the sign of 'num' before writing the 5179109998Smarkm result to 'rm' because if rm==num the value will be overwritten 5180109998Smarkm and produce the wrong result if 'num' is negative: this caused 5181109998Smarkm problems with BN_mod() and BN_nnmod(). 5182109998Smarkm [Steve Henson] 5183109998Smarkm 5184109998Smarkm *) Function OCSP_request_verify(). This checks the signature on an 5185109998Smarkm OCSP request and verifies the signer certificate. The signer 5186109998Smarkm certificate is just checked for a generic purpose and OCSP request 5187109998Smarkm trust settings. 5188109998Smarkm [Steve Henson] 5189109998Smarkm 5190109998Smarkm *) Add OCSP_check_validity() function to check the validity of OCSP 5191109998Smarkm responses. OCSP responses are prepared in real time and may only 5192109998Smarkm be a few seconds old. Simply checking that the current time lies 5193109998Smarkm between thisUpdate and nextUpdate max reject otherwise valid responses 5194109998Smarkm caused by either OCSP responder or client clock inaccuracy. Instead 5195109998Smarkm we allow thisUpdate and nextUpdate to fall within a certain period of 5196109998Smarkm the current time. The age of the response can also optionally be 5197109998Smarkm checked. Two new options -validity_period and -status_age added to 5198109998Smarkm ocsp utility. 5199109998Smarkm [Steve Henson] 5200109998Smarkm 5201109998Smarkm *) If signature or public key algorithm is unrecognized print out its 5202109998Smarkm OID rather that just UNKNOWN. 5203109998Smarkm [Steve Henson] 5204109998Smarkm 5205109998Smarkm *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 5206109998Smarkm OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 5207109998Smarkm ID to be generated from the issuer certificate alone which can then be 5208109998Smarkm passed to OCSP_id_issuer_cmp(). 5209109998Smarkm [Steve Henson] 5210109998Smarkm 5211109998Smarkm *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 5212109998Smarkm ASN1 modules to export functions returning ASN1_ITEM pointers 5213109998Smarkm instead of the ASN1_ITEM structures themselves. This adds several 5214109998Smarkm new macros which allow the underlying ASN1 function/structure to 5215109998Smarkm be accessed transparently. As a result code should not use ASN1_ITEM 5216109998Smarkm references directly (such as &X509_it) but instead use the relevant 5217109998Smarkm macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 5218109998Smarkm use of the new ASN1 code on platforms where exporting structures 5219109998Smarkm is problematical (for example in shared libraries) but exporting 5220109998Smarkm functions returning pointers to structures is not. 5221109998Smarkm [Steve Henson] 5222109998Smarkm 5223109998Smarkm *) Add support for overriding the generation of SSL/TLS session IDs. 5224109998Smarkm These callbacks can be registered either in an SSL_CTX or per SSL. 5225109998Smarkm The purpose of this is to allow applications to control, if they wish, 5226109998Smarkm the arbitrary values chosen for use as session IDs, particularly as it 5227109998Smarkm can be useful for session caching in multiple-server environments. A 5228109998Smarkm command-line switch for testing this (and any client code that wishes 5229109998Smarkm to use such a feature) has been added to "s_server". 5230109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 5231109998Smarkm 5232109998Smarkm *) Modify mkdef.pl to recognise and parse preprocessor conditionals 5233109998Smarkm of the form '#if defined(...) || defined(...) || ...' and 5234109998Smarkm '#if !defined(...) && !defined(...) && ...'. This also avoids 5235109998Smarkm the growing number of special cases it was previously handling. 5236109998Smarkm [Richard Levitte] 5237109998Smarkm 5238109998Smarkm *) Make all configuration macros available for application by making 5239109998Smarkm sure they are available in opensslconf.h, by giving them names starting 5240109998Smarkm with "OPENSSL_" to avoid conflicts with other packages and by making 5241109998Smarkm sure e_os2.h will cover all platform-specific cases together with 5242109998Smarkm opensslconf.h. 5243109998Smarkm Additionally, it is now possible to define configuration/platform- 5244109998Smarkm specific names (called "system identities"). In the C code, these 5245109998Smarkm are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 5246109998Smarkm macro with the name beginning with "OPENSSL_SYS_", which is determined 5247109998Smarkm from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 5248109998Smarkm what is available. 5249109998Smarkm [Richard Levitte] 5250109998Smarkm 5251109998Smarkm *) New option -set_serial to 'req' and 'x509' this allows the serial 5252109998Smarkm number to use to be specified on the command line. Previously self 5253109998Smarkm signed certificates were hard coded with serial number 0 and the 5254109998Smarkm CA options of 'x509' had to use a serial number in a file which was 5255109998Smarkm auto incremented. 5256109998Smarkm [Steve Henson] 5257109998Smarkm 5258109998Smarkm *) New options to 'ca' utility to support V2 CRL entry extensions. 5259109998Smarkm Currently CRL reason, invalidity date and hold instruction are 5260109998Smarkm supported. Add new CRL extensions to V3 code and some new objects. 5261109998Smarkm [Steve Henson] 5262109998Smarkm 5263109998Smarkm *) New function EVP_CIPHER_CTX_set_padding() this is used to 5264109998Smarkm disable standard block padding (aka PKCS#5 padding) in the EVP 5265109998Smarkm API, which was previously mandatory. This means that the data is 5266109998Smarkm not padded in any way and so the total length much be a multiple 5267109998Smarkm of the block size, otherwise an error occurs. 5268109998Smarkm [Steve Henson] 5269109998Smarkm 5270109998Smarkm *) Initial (incomplete) OCSP SSL support. 5271109998Smarkm [Steve Henson] 5272109998Smarkm 5273109998Smarkm *) New function OCSP_parse_url(). This splits up a URL into its host, 5274109998Smarkm port and path components: primarily to parse OCSP URLs. New -url 5275109998Smarkm option to ocsp utility. 5276109998Smarkm [Steve Henson] 5277109998Smarkm 5278109998Smarkm *) New nonce behavior. The return value of OCSP_check_nonce() now 5279109998Smarkm reflects the various checks performed. Applications can decide 5280109998Smarkm whether to tolerate certain situations such as an absent nonce 5281109998Smarkm in a response when one was present in a request: the ocsp application 5282109998Smarkm just prints out a warning. New function OCSP_add1_basic_nonce() 5283109998Smarkm this is to allow responders to include a nonce in a response even if 5284109998Smarkm the request is nonce-less. 5285109998Smarkm [Steve Henson] 5286109998Smarkm 5287109998Smarkm *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 5288109998Smarkm skipped when using openssl x509 multiple times on a single input file, 5289109998Smarkm e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 5290109998Smarkm [Bodo Moeller] 5291109998Smarkm 5292109998Smarkm *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 5293109998Smarkm set string type: to handle setting ASN1_TIME structures. Fix ca 5294109998Smarkm utility to correctly initialize revocation date of CRLs. 5295109998Smarkm [Steve Henson] 5296109998Smarkm 5297109998Smarkm *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 5298109998Smarkm the clients preferred ciphersuites and rather use its own preferences. 5299109998Smarkm Should help to work around M$ SGC (Server Gated Cryptography) bug in 5300109998Smarkm Internet Explorer by ensuring unchanged hash method during stepup. 5301109998Smarkm (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 5302109998Smarkm [Lutz Jaenicke] 5303109998Smarkm 5304109998Smarkm *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 5305109998Smarkm to aes and add a new 'exist' option to print out symbols that don't 5306109998Smarkm appear to exist. 5307109998Smarkm [Steve Henson] 5308109998Smarkm 5309109998Smarkm *) Additional options to ocsp utility to allow flags to be set and 5310109998Smarkm additional certificates supplied. 5311109998Smarkm [Steve Henson] 5312109998Smarkm 5313109998Smarkm *) Add the option -VAfile to 'openssl ocsp', so the user can give the 5314109998Smarkm OCSP client a number of certificate to only verify the response 5315109998Smarkm signature against. 5316109998Smarkm [Richard Levitte] 5317109998Smarkm 5318109998Smarkm *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 5319109998Smarkm handle the new API. Currently only ECB, CBC modes supported. Add new 5320109998Smarkm AES OIDs. 5321109998Smarkm 5322109998Smarkm Add TLS AES ciphersuites as described in RFC3268, "Advanced 5323109998Smarkm Encryption Standard (AES) Ciphersuites for Transport Layer 5324109998Smarkm Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 5325109998Smarkm not enabled by default and were not part of the "ALL" ciphersuite 5326109998Smarkm alias because they were not yet official; they could be 5327109998Smarkm explicitly requested by specifying the "AESdraft" ciphersuite 5328109998Smarkm group alias. In the final release of OpenSSL 0.9.7, the group 5329109998Smarkm alias is called "AES" and is part of "ALL".) 5330109998Smarkm [Ben Laurie, Steve Henson, Bodo Moeller] 5331109998Smarkm 5332109998Smarkm *) New function OCSP_copy_nonce() to copy nonce value (if present) from 5333109998Smarkm request to response. 5334109998Smarkm [Steve Henson] 5335109998Smarkm 5336109998Smarkm *) Functions for OCSP responders. OCSP_request_onereq_count(), 5337109998Smarkm OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 5338109998Smarkm extract information from a certificate request. OCSP_response_create() 5339109998Smarkm creates a response and optionally adds a basic response structure. 5340109998Smarkm OCSP_basic_add1_status() adds a complete single response to a basic 5341109998Smarkm response and returns the OCSP_SINGLERESP structure just added (to allow 5342109998Smarkm extensions to be included for example). OCSP_basic_add1_cert() adds a 5343109998Smarkm certificate to a basic response and OCSP_basic_sign() signs a basic 5344109998Smarkm response with various flags. New helper functions ASN1_TIME_check() 5345109998Smarkm (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 5346109998Smarkm (converts ASN1_TIME to GeneralizedTime). 5347109998Smarkm [Steve Henson] 5348109998Smarkm 5349109998Smarkm *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 5350109998Smarkm in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 5351109998Smarkm structure from a certificate. X509_pubkey_digest() digests the public_key 5352109998Smarkm contents: this is used in various key identifiers. 5353109998Smarkm [Steve Henson] 5354109998Smarkm 5355109998Smarkm *) Make sk_sort() tolerate a NULL argument. 5356109998Smarkm [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 5357109998Smarkm 5358109998Smarkm *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 5359109998Smarkm passed by the function are trusted implicitly. If any of them signed the 5360109998Smarkm response then it is assumed to be valid and is not verified. 5361109998Smarkm [Steve Henson] 5362109998Smarkm 5363109998Smarkm *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 5364109998Smarkm to data. This was previously part of the PKCS7 ASN1 code. This 5365109998Smarkm was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 5366109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 5367109998Smarkm <support@securenetterm.com>] 5368109998Smarkm 5369109998Smarkm *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 5370109998Smarkm routines: without these tracing memory leaks is very painful. 5371109998Smarkm Fix leaks in PKCS12 and PKCS7 routines. 5372109998Smarkm [Steve Henson] 5373109998Smarkm 5374109998Smarkm *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 5375109998Smarkm Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 5376109998Smarkm effectively meant GeneralizedTime would never be used. Now it 5377109998Smarkm is initialised to -1 but X509_time_adj() now has to check the value 5378109998Smarkm and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 5379109998Smarkm V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 5380109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 5381109998Smarkm <support@securenetterm.com>] 5382109998Smarkm 5383109998Smarkm *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 5384109998Smarkm result in a zero length in the ASN1_INTEGER structure which was 5385109998Smarkm not consistent with the structure when d2i_ASN1_INTEGER() was used 5386109998Smarkm and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 5387109998Smarkm to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 5388109998Smarkm where it did not print out a minus for negative ASN1_INTEGER. 5389109998Smarkm [Steve Henson] 5390109998Smarkm 5391109998Smarkm *) Add summary printout to ocsp utility. The various functions which 5392109998Smarkm convert status values to strings have been renamed to: 5393109998Smarkm OCSP_response_status_str(), OCSP_cert_status_str() and 5394109998Smarkm OCSP_crl_reason_str() and are no longer static. New options 5395109998Smarkm to verify nonce values and to disable verification. OCSP response 5396109998Smarkm printout format cleaned up. 5397109998Smarkm [Steve Henson] 5398109998Smarkm 5399109998Smarkm *) Add additional OCSP certificate checks. These are those specified 5400109998Smarkm in RFC2560. This consists of two separate checks: the CA of the 5401109998Smarkm certificate being checked must either be the OCSP signer certificate 5402109998Smarkm or the issuer of the OCSP signer certificate. In the latter case the 5403109998Smarkm OCSP signer certificate must contain the OCSP signing extended key 5404109998Smarkm usage. This check is performed by attempting to match the OCSP 5405109998Smarkm signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 5406109998Smarkm in the OCSP_CERTID structures of the response. 5407109998Smarkm [Steve Henson] 5408109998Smarkm 5409109998Smarkm *) Initial OCSP certificate verification added to OCSP_basic_verify() 5410109998Smarkm and related routines. This uses the standard OpenSSL certificate 5411109998Smarkm verify routines to perform initial checks (just CA validity) and 5412109998Smarkm to obtain the certificate chain. Then additional checks will be 5413109998Smarkm performed on the chain. Currently the root CA is checked to see 5414109998Smarkm if it is explicitly trusted for OCSP signing. This is used to set 5415109998Smarkm a root CA as a global signing root: that is any certificate that 5416109998Smarkm chains to that CA is an acceptable OCSP signing certificate. 5417109998Smarkm [Steve Henson] 5418109998Smarkm 5419109998Smarkm *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 5420109998Smarkm extensions from a separate configuration file. 5421109998Smarkm As when reading extensions from the main configuration file, 5422109998Smarkm the '-extensions ...' option may be used for specifying the 5423109998Smarkm section to use. 5424109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 5425109998Smarkm 5426109998Smarkm *) New OCSP utility. Allows OCSP requests to be generated or 5427109998Smarkm read. The request can be sent to a responder and the output 5428109998Smarkm parsed, outputed or printed in text form. Not complete yet: 5429109998Smarkm still needs to check the OCSP response validity. 5430109998Smarkm [Steve Henson] 5431109998Smarkm 5432109998Smarkm *) New subcommands for 'openssl ca': 5433109998Smarkm 'openssl ca -status <serial>' prints the status of the cert with 5434109998Smarkm the given serial number (according to the index file). 5435109998Smarkm 'openssl ca -updatedb' updates the expiry status of certificates 5436109998Smarkm in the index file. 5437109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 5438109998Smarkm 5439109998Smarkm *) New '-newreq-nodes' command option to CA.pl. This is like 5440109998Smarkm '-newreq', but calls 'openssl req' with the '-nodes' option 5441109998Smarkm so that the resulting key is not encrypted. 5442109998Smarkm [Damien Miller <djm@mindrot.org>] 5443109998Smarkm 5444109998Smarkm *) New configuration for the GNU Hurd. 5445109998Smarkm [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 5446109998Smarkm 5447109998Smarkm *) Initial code to implement OCSP basic response verify. This 5448109998Smarkm is currently incomplete. Currently just finds the signer's 5449109998Smarkm certificate and verifies the signature on the response. 5450109998Smarkm [Steve Henson] 5451109998Smarkm 5452109998Smarkm *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 5453109998Smarkm value of OPENSSLDIR. This is available via the new '-d' option 5454109998Smarkm to 'openssl version', and is also included in 'openssl version -a'. 5455109998Smarkm [Bodo Moeller] 5456109998Smarkm 5457109998Smarkm *) Allowing defining memory allocation callbacks that will be given 5458109998Smarkm file name and line number information in additional arguments 5459109998Smarkm (a const char* and an int). The basic functionality remains, as 5460109998Smarkm well as the original possibility to just replace malloc(), 5461109998Smarkm realloc() and free() by functions that do not know about these 5462109998Smarkm additional arguments. To register and find out the current 5463109998Smarkm settings for extended allocation functions, the following 5464109998Smarkm functions are provided: 5465109998Smarkm 5466109998Smarkm CRYPTO_set_mem_ex_functions 5467109998Smarkm CRYPTO_set_locked_mem_ex_functions 5468109998Smarkm CRYPTO_get_mem_ex_functions 5469109998Smarkm CRYPTO_get_locked_mem_ex_functions 5470109998Smarkm 5471109998Smarkm These work the same way as CRYPTO_set_mem_functions and friends. 5472109998Smarkm CRYPTO_get_[locked_]mem_functions now writes 0 where such an 5473109998Smarkm extended allocation function is enabled. 5474109998Smarkm Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 5475109998Smarkm a conventional allocation function is enabled. 5476109998Smarkm [Richard Levitte, Bodo Moeller] 5477109998Smarkm 5478109998Smarkm *) Finish off removing the remaining LHASH function pointer casts. 5479109998Smarkm There should no longer be any prototype-casting required when using 5480109998Smarkm the LHASH abstraction, and any casts that remain are "bugs". See 5481109998Smarkm the callback types and macros at the head of lhash.h for details 5482109998Smarkm (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 5483109998Smarkm [Geoff Thorpe] 5484109998Smarkm 5485109998Smarkm *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 5486109998Smarkm If /dev/[u]random devices are not available or do not return enough 5487109998Smarkm entropy, EGD style sockets (served by EGD or PRNGD) will automatically 5488109998Smarkm be queried. 5489109998Smarkm The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 5490109998Smarkm /etc/entropy will be queried once each in this sequence, quering stops 5491109998Smarkm when enough entropy was collected without querying more sockets. 5492109998Smarkm [Lutz Jaenicke] 5493109998Smarkm 5494109998Smarkm *) Change the Unix RAND_poll() variant to be able to poll several 5495109998Smarkm random devices, as specified by DEVRANDOM, until a sufficient amount 5496109998Smarkm of data has been collected. We spend at most 10 ms on each file 5497109998Smarkm (select timeout) and read in non-blocking mode. DEVRANDOM now 5498109998Smarkm defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 5499109998Smarkm (previously it was just the string "/dev/urandom"), so on typical 5500109998Smarkm platforms the 10 ms delay will never occur. 5501109998Smarkm Also separate out the Unix variant to its own file, rand_unix.c. 5502109998Smarkm For VMS, there's a currently-empty rand_vms.c. 5503109998Smarkm [Richard Levitte] 5504109998Smarkm 5505109998Smarkm *) Move OCSP client related routines to ocsp_cl.c. These 5506109998Smarkm provide utility functions which an application needing 5507109998Smarkm to issue a request to an OCSP responder and analyse the 5508109998Smarkm response will typically need: as opposed to those which an 5509109998Smarkm OCSP responder itself would need which will be added later. 5510109998Smarkm 5511109998Smarkm OCSP_request_sign() signs an OCSP request with an API similar 5512109998Smarkm to PKCS7_sign(). OCSP_response_status() returns status of OCSP 5513109998Smarkm response. OCSP_response_get1_basic() extracts basic response 5514109998Smarkm from response. OCSP_resp_find_status(): finds and extracts status 5515109998Smarkm information from an OCSP_CERTID structure (which will be created 5516109998Smarkm when the request structure is built). These are built from lower 5517109998Smarkm level functions which work on OCSP_SINGLERESP structures but 5518109998Smarkm wont normally be used unless the application wishes to examine 5519109998Smarkm extensions in the OCSP response for example. 5520109998Smarkm 5521109998Smarkm Replace nonce routines with a pair of functions. 5522109998Smarkm OCSP_request_add1_nonce() adds a nonce value and optionally 5523109998Smarkm generates a random value. OCSP_check_nonce() checks the 5524109998Smarkm validity of the nonce in an OCSP response. 5525109998Smarkm [Steve Henson] 5526109998Smarkm 5527109998Smarkm *) Change function OCSP_request_add() to OCSP_request_add0_id(). 5528109998Smarkm This doesn't copy the supplied OCSP_CERTID and avoids the 5529109998Smarkm need to free up the newly created id. Change return type 5530109998Smarkm to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 5531109998Smarkm This can then be used to add extensions to the request. 5532109998Smarkm Deleted OCSP_request_new(), since most of its functionality 5533109998Smarkm is now in OCSP_REQUEST_new() (and the case insensitive name 5534109998Smarkm clash) apart from the ability to set the request name which 5535109998Smarkm will be added elsewhere. 5536109998Smarkm [Steve Henson] 5537109998Smarkm 5538109998Smarkm *) Update OCSP API. Remove obsolete extensions argument from 5539109998Smarkm various functions. Extensions are now handled using the new 5540109998Smarkm OCSP extension code. New simple OCSP HTTP function which 5541109998Smarkm can be used to send requests and parse the response. 5542109998Smarkm [Steve Henson] 5543109998Smarkm 5544109998Smarkm *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 5545109998Smarkm ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 5546109998Smarkm uses the special reorder version of SET OF to sort the attributes 5547109998Smarkm and reorder them to match the encoded order. This resolves a long 5548109998Smarkm standing problem: a verify on a PKCS7 structure just after signing 5549109998Smarkm it used to fail because the attribute order did not match the 5550109998Smarkm encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 5551109998Smarkm it uses the received order. This is necessary to tolerate some broken 5552109998Smarkm software that does not order SET OF. This is handled by encoding 5553109998Smarkm as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 5554109998Smarkm to produce the required SET OF. 5555109998Smarkm [Steve Henson] 5556109998Smarkm 5557109998Smarkm *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 5558109998Smarkm OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 5559109998Smarkm files to get correct declarations of the ASN.1 item variables. 5560109998Smarkm [Richard Levitte] 5561109998Smarkm 5562109998Smarkm *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 5563109998Smarkm PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 5564109998Smarkm asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 5565109998Smarkm NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 5566109998Smarkm New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 5567109998Smarkm ASN1_ITEM and no wrapper functions. 5568109998Smarkm [Steve Henson] 5569109998Smarkm 5570109998Smarkm *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 5571109998Smarkm replace the old function pointer based I/O routines. Change most of 5572109998Smarkm the *_d2i_bio() and *_d2i_fp() functions to use these. 5573109998Smarkm [Steve Henson] 5574109998Smarkm 5575109998Smarkm *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 5576109998Smarkm lines, recognice more "algorithms" that can be deselected, and make 5577109998Smarkm it complain about algorithm deselection that isn't recognised. 5578109998Smarkm [Richard Levitte] 5579109998Smarkm 5580109998Smarkm *) New ASN1 functions to handle dup, sign, verify, digest, pack and 5581109998Smarkm unpack operations in terms of ASN1_ITEM. Modify existing wrappers 5582109998Smarkm to use new functions. Add NO_ASN1_OLD which can be set to remove 5583109998Smarkm some old style ASN1 functions: this can be used to determine if old 5584109998Smarkm code will still work when these eventually go away. 5585109998Smarkm [Steve Henson] 5586109998Smarkm 5587109998Smarkm *) New extension functions for OCSP structures, these follow the 5588109998Smarkm same conventions as certificates and CRLs. 5589109998Smarkm [Steve Henson] 5590109998Smarkm 5591109998Smarkm *) New function X509V3_add1_i2d(). This automatically encodes and 5592109998Smarkm adds an extension. Its behaviour can be customised with various 5593109998Smarkm flags to append, replace or delete. Various wrappers added for 5594109998Smarkm certifcates and CRLs. 5595109998Smarkm [Steve Henson] 5596109998Smarkm 5597109998Smarkm *) Fix to avoid calling the underlying ASN1 print routine when 5598109998Smarkm an extension cannot be parsed. Correct a typo in the 5599109998Smarkm OCSP_SERVICELOC extension. Tidy up print OCSP format. 5600109998Smarkm [Steve Henson] 5601109998Smarkm 5602109998Smarkm *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 5603109998Smarkm entries for variables. 5604109998Smarkm [Steve Henson] 5605109998Smarkm 5606109998Smarkm *) Add functionality to apps/openssl.c for detecting locking 5607109998Smarkm problems: As the program is single-threaded, all we have 5608109998Smarkm to do is register a locking callback using an array for 5609109998Smarkm storing which locks are currently held by the program. 5610109998Smarkm [Bodo Moeller] 5611109998Smarkm 5612109998Smarkm *) Use a lock around the call to CRYPTO_get_ex_new_index() in 5613109998Smarkm SSL_get_ex_data_X509_STORE_idx(), which is used in 5614109998Smarkm ssl_verify_cert_chain() and thus can be called at any time 5615109998Smarkm during TLS/SSL handshakes so that thread-safety is essential. 5616109998Smarkm Unfortunately, the ex_data design is not at all suited 5617109998Smarkm for multi-threaded use, so it probably should be abolished. 5618109998Smarkm [Bodo Moeller] 5619109998Smarkm 5620109998Smarkm *) Added Broadcom "ubsec" ENGINE to OpenSSL. 5621109998Smarkm [Broadcom, tweaked and integrated by Geoff Thorpe] 5622109998Smarkm 5623109998Smarkm *) Move common extension printing code to new function 5624109998Smarkm X509V3_print_extensions(). Reorganise OCSP print routines and 5625109998Smarkm implement some needed OCSP ASN1 functions. Add OCSP extensions. 5626109998Smarkm [Steve Henson] 5627109998Smarkm 5628109998Smarkm *) New function X509_signature_print() to remove duplication in some 5629109998Smarkm print routines. 5630109998Smarkm [Steve Henson] 5631109998Smarkm 5632109998Smarkm *) Add a special meaning when SET OF and SEQUENCE OF flags are both 5633109998Smarkm set (this was treated exactly the same as SET OF previously). This 5634109998Smarkm is used to reorder the STACK representing the structure to match the 5635109998Smarkm encoding. This will be used to get round a problem where a PKCS7 5636109998Smarkm structure which was signed could not be verified because the STACK 5637109998Smarkm order did not reflect the encoded order. 5638109998Smarkm [Steve Henson] 5639109998Smarkm 5640109998Smarkm *) Reimplement the OCSP ASN1 module using the new code. 5641109998Smarkm [Steve Henson] 5642109998Smarkm 5643109998Smarkm *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 5644109998Smarkm for its ASN1 operations. The old style function pointers still exist 5645109998Smarkm for now but they will eventually go away. 5646109998Smarkm [Steve Henson] 5647109998Smarkm 5648109998Smarkm *) Merge in replacement ASN1 code from the ASN1 branch. This almost 5649109998Smarkm completely replaces the old ASN1 functionality with a table driven 5650109998Smarkm encoder and decoder which interprets an ASN1_ITEM structure describing 5651109998Smarkm the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 5652109998Smarkm largely maintained. Almost all of the old asn1_mac.h macro based ASN1 5653109998Smarkm has also been converted to the new form. 5654109998Smarkm [Steve Henson] 5655109998Smarkm 5656109998Smarkm *) Change BN_mod_exp_recp so that negative moduli are tolerated 5657109998Smarkm (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 5658109998Smarkm so that BN_mod_exp_mont and BN_mod_exp_mont_word work 5659109998Smarkm for negative moduli. 5660109998Smarkm [Bodo Moeller] 5661109998Smarkm 5662109998Smarkm *) Fix BN_uadd and BN_usub: Always return non-negative results instead 5663109998Smarkm of not touching the result's sign bit. 5664109998Smarkm [Bodo Moeller] 5665109998Smarkm 5666109998Smarkm *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 5667109998Smarkm set. 5668109998Smarkm [Bodo Moeller] 5669109998Smarkm 5670109998Smarkm *) Changed the LHASH code to use prototypes for callbacks, and created 5671109998Smarkm macros to declare and implement thin (optionally static) functions 5672109998Smarkm that provide type-safety and avoid function pointer casting for the 5673109998Smarkm type-specific callbacks. 5674109998Smarkm [Geoff Thorpe] 5675109998Smarkm 5676109998Smarkm *) Added Kerberos Cipher Suites to be used with TLS, as written in 5677109998Smarkm RFC 2712. 5678109998Smarkm [Veers Staats <staatsvr@asc.hpc.mil>, 5679109998Smarkm Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 5680109998Smarkm 5681109998Smarkm *) Reformat the FAQ so the different questions and answers can be divided 5682109998Smarkm in sections depending on the subject. 5683109998Smarkm [Richard Levitte] 5684109998Smarkm 5685109998Smarkm *) Have the zlib compression code load ZLIB.DLL dynamically under 5686109998Smarkm Windows. 5687109998Smarkm [Richard Levitte] 5688109998Smarkm 5689109998Smarkm *) New function BN_mod_sqrt for computing square roots modulo a prime 5690109998Smarkm (using the probabilistic Tonelli-Shanks algorithm unless 5691109998Smarkm p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 5692109998Smarkm be handled deterministically). 5693109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5694109998Smarkm 5695109998Smarkm *) Make BN_mod_inverse faster by explicitly handling small quotients 5696109998Smarkm in the Euclid loop. (Speed gain about 20% for small moduli [256 or 5697109998Smarkm 512 bits], about 30% for larger ones [1024 or 2048 bits].) 5698109998Smarkm [Bodo Moeller] 5699109998Smarkm 5700109998Smarkm *) New function BN_kronecker. 5701109998Smarkm [Bodo Moeller] 5702109998Smarkm 5703109998Smarkm *) Fix BN_gcd so that it works on negative inputs; the result is 5704109998Smarkm positive unless both parameters are zero. 5705109998Smarkm Previously something reasonably close to an infinite loop was 5706109998Smarkm possible because numbers could be growing instead of shrinking 5707109998Smarkm in the implementation of Euclid's algorithm. 5708109998Smarkm [Bodo Moeller] 5709109998Smarkm 5710109998Smarkm *) Fix BN_is_word() and BN_is_one() macros to take into account the 5711109998Smarkm sign of the number in question. 5712109998Smarkm 5713109998Smarkm Fix BN_is_word(a,w) to work correctly for w == 0. 5714109998Smarkm 5715109998Smarkm The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 5716109998Smarkm because its test if the absolute value of 'a' equals 'w'. 5717109998Smarkm Note that BN_abs_is_word does *not* handle w == 0 reliably; 5718109998Smarkm it exists mostly for use in the implementations of BN_is_zero(), 5719109998Smarkm BN_is_one(), and BN_is_word(). 5720109998Smarkm [Bodo Moeller] 5721109998Smarkm 5722109998Smarkm *) New function BN_swap. 5723109998Smarkm [Bodo Moeller] 5724109998Smarkm 5725109998Smarkm *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 5726109998Smarkm the exponentiation functions are more likely to produce reasonable 5727109998Smarkm results on negative inputs. 5728109998Smarkm [Bodo Moeller] 5729109998Smarkm 5730109998Smarkm *) Change BN_mod_mul so that the result is always non-negative. 5731109998Smarkm Previously, it could be negative if one of the factors was negative; 5732109998Smarkm I don't think anyone really wanted that behaviour. 5733109998Smarkm [Bodo Moeller] 5734109998Smarkm 5735109998Smarkm *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 5736109998Smarkm (except for exponentiation, which stays in crypto/bn/bn_exp.c, 5737109998Smarkm and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 5738109998Smarkm and add new functions: 5739109998Smarkm 5740109998Smarkm BN_nnmod 5741109998Smarkm BN_mod_sqr 5742109998Smarkm BN_mod_add 5743109998Smarkm BN_mod_add_quick 5744109998Smarkm BN_mod_sub 5745109998Smarkm BN_mod_sub_quick 5746109998Smarkm BN_mod_lshift1 5747109998Smarkm BN_mod_lshift1_quick 5748109998Smarkm BN_mod_lshift 5749109998Smarkm BN_mod_lshift_quick 5750109998Smarkm 5751109998Smarkm These functions always generate non-negative results. 5752109998Smarkm 5753109998Smarkm BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 5754109998Smarkm such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 5755109998Smarkm 5756109998Smarkm BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 5757109998Smarkm BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 5758109998Smarkm be reduced modulo m. 5759109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5760109998Smarkm 5761111147Snectar#if 0 5762111147Snectar The following entry accidentily appeared in the CHANGES file 5763111147Snectar distributed with OpenSSL 0.9.7. The modifications described in 5764111147Snectar it do *not* apply to OpenSSL 0.9.7. 5765111147Snectar 5766109998Smarkm *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 5767109998Smarkm was actually never needed) and in BN_mul(). The removal in BN_mul() 5768109998Smarkm required a small change in bn_mul_part_recursive() and the addition 5769109998Smarkm of the functions bn_cmp_part_words(), bn_sub_part_words() and 5770109998Smarkm bn_add_part_words(), which do the same thing as bn_cmp_words(), 5771109998Smarkm bn_sub_words() and bn_add_words() except they take arrays with 5772109998Smarkm differing sizes. 5773109998Smarkm [Richard Levitte] 5774111147Snectar#endif 5775109998Smarkm 5776109998Smarkm *) In 'openssl passwd', verify passwords read from the terminal 5777109998Smarkm unless the '-salt' option is used (which usually means that 5778109998Smarkm verification would just waste user's time since the resulting 5779109998Smarkm hash is going to be compared with some given password hash) 5780109998Smarkm or the new '-noverify' option is used. 5781109998Smarkm 5782109998Smarkm This is an incompatible change, but it does not affect 5783109998Smarkm non-interactive use of 'openssl passwd' (passwords on the command 5784109998Smarkm line, '-stdin' option, '-in ...' option) and thus should not 5785109998Smarkm cause any problems. 5786109998Smarkm [Bodo Moeller] 5787109998Smarkm 5788109998Smarkm *) Remove all references to RSAref, since there's no more need for it. 5789109998Smarkm [Richard Levitte] 5790109998Smarkm 5791109998Smarkm *) Make DSO load along a path given through an environment variable 5792109998Smarkm (SHLIB_PATH) with shl_load(). 5793109998Smarkm [Richard Levitte] 5794109998Smarkm 5795109998Smarkm *) Constify the ENGINE code as a result of BIGNUM constification. 5796109998Smarkm Also constify the RSA code and most things related to it. In a 5797109998Smarkm few places, most notable in the depth of the ASN.1 code, ugly 5798109998Smarkm casts back to non-const were required (to be solved at a later 5799109998Smarkm time) 5800109998Smarkm [Richard Levitte] 5801109998Smarkm 5802109998Smarkm *) Make it so the openssl application has all engines loaded by default. 5803109998Smarkm [Richard Levitte] 5804109998Smarkm 5805109998Smarkm *) Constify the BIGNUM routines a little more. 5806109998Smarkm [Richard Levitte] 5807109998Smarkm 5808109998Smarkm *) Add the following functions: 5809109998Smarkm 5810109998Smarkm ENGINE_load_cswift() 5811109998Smarkm ENGINE_load_chil() 5812109998Smarkm ENGINE_load_atalla() 5813109998Smarkm ENGINE_load_nuron() 5814109998Smarkm ENGINE_load_builtin_engines() 5815109998Smarkm 5816109998Smarkm That way, an application can itself choose if external engines that 5817109998Smarkm are built-in in OpenSSL shall ever be used or not. The benefit is 5818109998Smarkm that applications won't have to be linked with libdl or other dso 5819109998Smarkm libraries unless it's really needed. 5820109998Smarkm 5821109998Smarkm Changed 'openssl engine' to load all engines on demand. 5822109998Smarkm Changed the engine header files to avoid the duplication of some 5823109998Smarkm declarations (they differed!). 5824109998Smarkm [Richard Levitte] 5825109998Smarkm 5826109998Smarkm *) 'openssl engine' can now list capabilities. 5827109998Smarkm [Richard Levitte] 5828109998Smarkm 5829109998Smarkm *) Better error reporting in 'openssl engine'. 5830109998Smarkm [Richard Levitte] 5831109998Smarkm 5832109998Smarkm *) Never call load_dh_param(NULL) in s_server. 5833109998Smarkm [Bodo Moeller] 5834109998Smarkm 5835109998Smarkm *) Add engine application. It can currently list engines by name and 5836109998Smarkm identity, and test if they are actually available. 5837109998Smarkm [Richard Levitte] 5838109998Smarkm 5839109998Smarkm *) Improve RPM specification file by forcing symbolic linking and making 5840109998Smarkm sure the installed documentation is also owned by root.root. 5841109998Smarkm [Damien Miller <djm@mindrot.org>] 5842109998Smarkm 5843109998Smarkm *) Give the OpenSSL applications more possibilities to make use of 5844109998Smarkm keys (public as well as private) handled by engines. 5845109998Smarkm [Richard Levitte] 5846109998Smarkm 5847109998Smarkm *) Add OCSP code that comes from CertCo. 5848109998Smarkm [Richard Levitte] 5849109998Smarkm 5850109998Smarkm *) Add VMS support for the Rijndael code. 5851109998Smarkm [Richard Levitte] 5852109998Smarkm 5853109998Smarkm *) Added untested support for Nuron crypto accelerator. 5854109998Smarkm [Ben Laurie] 5855109998Smarkm 5856109998Smarkm *) Add support for external cryptographic devices. This code was 5857109998Smarkm previously distributed separately as the "engine" branch. 5858109998Smarkm [Geoff Thorpe, Richard Levitte] 5859109998Smarkm 5860109998Smarkm *) Rework the filename-translation in the DSO code. It is now possible to 5861109998Smarkm have far greater control over how a "name" is turned into a filename 5862109998Smarkm depending on the operating environment and any oddities about the 5863109998Smarkm different shared library filenames on each system. 5864109998Smarkm [Geoff Thorpe] 5865109998Smarkm 5866109998Smarkm *) Support threads on FreeBSD-elf in Configure. 5867109998Smarkm [Richard Levitte] 5868109998Smarkm 5869109998Smarkm *) Fix for SHA1 assembly problem with MASM: it produces 5870109998Smarkm warnings about corrupt line number information when assembling 5871109998Smarkm with debugging information. This is caused by the overlapping 5872109998Smarkm of two sections. 5873109998Smarkm [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 5874109998Smarkm 5875109998Smarkm *) NCONF changes. 5876109998Smarkm NCONF_get_number() has no error checking at all. As a replacement, 5877109998Smarkm NCONF_get_number_e() is defined (_e for "error checking") and is 5878109998Smarkm promoted strongly. The old NCONF_get_number is kept around for 5879109998Smarkm binary backward compatibility. 5880109998Smarkm Make it possible for methods to load from something other than a BIO, 5881109998Smarkm by providing a function pointer that is given a name instead of a BIO. 5882109998Smarkm For example, this could be used to load configuration data from an 5883109998Smarkm LDAP server. 5884109998Smarkm [Richard Levitte] 5885109998Smarkm 5886109998Smarkm *) Fix for non blocking accept BIOs. Added new I/O special reason 5887109998Smarkm BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 5888109998Smarkm with non blocking I/O was not possible because no retry code was 5889109998Smarkm implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 5890109998Smarkm this case. 5891109998Smarkm [Steve Henson] 5892109998Smarkm 5893109998Smarkm *) Added the beginnings of Rijndael support. 5894109998Smarkm [Ben Laurie] 5895109998Smarkm 5896109998Smarkm *) Fix for bug in DirectoryString mask setting. Add support for 5897109998Smarkm X509_NAME_print_ex() in 'req' and X509_print_ex() function 5898109998Smarkm to allow certificate printing to more controllable, additional 5899109998Smarkm 'certopt' option to 'x509' to allow new printing options to be 5900109998Smarkm set. 5901109998Smarkm [Steve Henson] 5902109998Smarkm 5903109998Smarkm *) Clean old EAY MD5 hack from e_os.h. 5904109998Smarkm [Richard Levitte] 5905109998Smarkm 5906142425Snectar Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 5907142425Snectar 5908142425Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 5909160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 5910142425Snectar [Joe Orton, Steve Henson] 5911142425Snectar 5912142425Snectar Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 5913142425Snectar 5914142425Snectar *) Fix additional bug revealed by the NISCC test suite: 5915142425Snectar 5916142425Snectar Stop bug triggering large recursion when presented with 5917160814Ssimon certain ASN.1 tags (CVE-2003-0851) 5918142425Snectar [Steve Henson] 5919142425Snectar 5920120631Snectar Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 5921120631Snectar 5922120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 5923120631Snectar 5924120631Snectar Stop out of bounds reads in the ASN1 code when presented with 5925160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 5926120631Snectar 5927120631Snectar If verify callback ignores invalid public key errors don't try to check 5928120631Snectar certificate signature with the NULL public key. 5929120631Snectar 5930120631Snectar [Steve Henson] 5931120631Snectar 5932120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 5933120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 5934120631Snectar specifications. 5935120631Snectar [Steve Henson] 5936120631Snectar 5937120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 5938120631Snectar extra data after the compression methods not only for TLS 1.0 5939120631Snectar but also for SSL 3.0 (as required by the specification). 5940120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 5941120631Snectar 5942120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 5943120631Snectar when it's 512 *bits* long, not 512 bytes. 5944120631Snectar [Richard Levitte] 5945120631Snectar 5946120631Snectar Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 5947120631Snectar 5948120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 5949120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 5950120631Snectar a protocol version number mismatch like a decryption error 5951120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 5952120631Snectar [Bodo Moeller] 5953120631Snectar 5954120631Snectar *) Turn on RSA blinding by default in the default implementation 5955120631Snectar to avoid a timing attack. Applications that don't want it can call 5956120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 5957120631Snectar They would be ill-advised to do so in most cases. 5958120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 5959120631Snectar 5960120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 5961120631Snectar seeded (in this case, the secret RSA exponent is abused as 5962120631Snectar an unpredictable seed -- if it is not unpredictable, there 5963120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 5964120631Snectar by remembering the creator's thread ID in rsa->blinding and 5965120631Snectar having all other threads use local one-time blinding factors 5966120631Snectar (this requires more computation than sharing rsa->blinding, but 5967120631Snectar avoids excessive locking; and if an RSA object is not shared 5968120631Snectar between threads, blinding will still be very fast). 5969120631Snectar [Bodo Moeller] 5970120631Snectar 5971111147Snectar Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 5972111147Snectar 5973111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 5974111147Snectar via timing by performing a MAC computation even if incorrrect 5975111147Snectar block cipher padding has been found. This is a countermeasure 5976111147Snectar against active attacks where the attacker has to distinguish 5977160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 5978111147Snectar 5979111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 5980111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 5981111147Snectar Martin Vuagnoux (EPFL, Ilion)] 5982111147Snectar 5983109998Smarkm Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 5984109998Smarkm 5985109998Smarkm *) New function OPENSSL_cleanse(), which is used to cleanse a section of 5986109998Smarkm memory from it's contents. This is done with a counter that will 5987109998Smarkm place alternating values in each byte. This can be used to solve 5988109998Smarkm two issues: 1) the removal of calls to memset() by highly optimizing 5989109998Smarkm compilers, and 2) cleansing with other values than 0, since those can 5990109998Smarkm be read through on certain media, for example a swap space on disk. 5991109998Smarkm [Geoff Thorpe] 5992109998Smarkm 5993109998Smarkm *) Bugfix: client side session caching did not work with external caching, 5994109998Smarkm because the session->cipher setting was not restored when reloading 5995109998Smarkm from the external cache. This problem was masked, when 5996109998Smarkm SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 5997109998Smarkm (Found by Steve Haslam <steve@araqnid.ddts.net>.) 5998109998Smarkm [Lutz Jaenicke] 5999109998Smarkm 6000109998Smarkm *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 6001109998Smarkm length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 6002109998Smarkm [Zeev Lieber <zeev-l@yahoo.com>] 6003109998Smarkm 6004109998Smarkm *) Undo an undocumented change introduced in 0.9.6e which caused 6005109998Smarkm repeated calls to OpenSSL_add_all_ciphers() and 6006109998Smarkm OpenSSL_add_all_digests() to be ignored, even after calling 6007109998Smarkm EVP_cleanup(). 6008109998Smarkm [Richard Levitte] 6009109998Smarkm 6010109998Smarkm *) Change the default configuration reader to deal with last line not 6011109998Smarkm being properly terminated. 6012109998Smarkm [Richard Levitte] 6013109998Smarkm 6014109998Smarkm *) Change X509_NAME_cmp() so it applies the special rules on handling 6015109998Smarkm DN values that are of type PrintableString, as well as RDNs of type 6016109998Smarkm emailAddress where the value has the type ia5String. 6017109998Smarkm [stefank@valicert.com via Richard Levitte] 6018109998Smarkm 6019109998Smarkm *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 6020109998Smarkm the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 6021109998Smarkm doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 6022109998Smarkm the bitwise-OR of the two for use by the majority of applications 6023109998Smarkm wanting this behaviour, and update the docs. The documented 6024109998Smarkm behaviour and actual behaviour were inconsistent and had been 6025109998Smarkm changing anyway, so this is more a bug-fix than a behavioural 6026109998Smarkm change. 6027109998Smarkm [Geoff Thorpe, diagnosed by Nadav Har'El] 6028109998Smarkm 6029109998Smarkm *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 6030109998Smarkm (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 6031109998Smarkm [Bodo Moeller] 6032109998Smarkm 6033109998Smarkm *) Fix initialization code race conditions in 6034109998Smarkm SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 6035109998Smarkm SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 6036109998Smarkm SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 6037109998Smarkm TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 6038109998Smarkm ssl2_get_cipher_by_char(), 6039109998Smarkm ssl3_get_cipher_by_char(). 6040109998Smarkm [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 6041109998Smarkm 6042109998Smarkm *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 6043109998Smarkm the cached sessions are flushed, as the remove_cb() might use ex_data 6044109998Smarkm contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 6045109998Smarkm (see [openssl.org #212]). 6046109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 6047109998Smarkm 6048109998Smarkm *) Fix typo in OBJ_txt2obj which incorrectly passed the content 6049109998Smarkm length, instead of the encoding length to d2i_ASN1_OBJECT. 6050109998Smarkm [Steve Henson] 6051109998Smarkm 6052101618Snectar Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 6053101618Snectar 6054101618Snectar *) [In 0.9.6g-engine release:] 6055101618Snectar Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 6056101618Snectar [Lynn Gazis <lgazis@rainbow.com>] 6057101618Snectar 6058101613Snectar Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 6059101613Snectar 6060101613Snectar *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 6061101613Snectar and get fix the header length calculation. 6062101613Snectar [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 6063101613Snectar Alon Kantor <alonk@checkpoint.com> (and others), 6064101613Snectar Steve Henson] 6065101613Snectar 6066101613Snectar *) Use proper error handling instead of 'assertions' in buffer 6067101613Snectar overflow checks added in 0.9.6e. This prevents DoS (the 6068101613Snectar assertions could call abort()). 6069101613Snectar [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 6070101613Snectar 6071100936Snectar Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 6072100936Snectar 6073109998Smarkm *) Add various sanity checks to asn1_get_length() to reject 6074109998Smarkm the ASN1 length bytes if they exceed sizeof(long), will appear 6075109998Smarkm negative or the content length exceeds the length of the 6076109998Smarkm supplied buffer. 6077109998Smarkm [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6078109998Smarkm 6079100936Snectar *) Fix cipher selection routines: ciphers without encryption had no flags 6080100936Snectar for the cipher strength set and where therefore not handled correctly 6081100936Snectar by the selection routines (PR #130). 6082100936Snectar [Lutz Jaenicke] 6083100936Snectar 6084100936Snectar *) Fix EVP_dsa_sha macro. 6085100936Snectar [Nils Larsch] 6086100936Snectar 6087100936Snectar *) New option 6088100936Snectar SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 6089100936Snectar for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 6090100936Snectar that was added in OpenSSL 0.9.6d. 6091100936Snectar 6092100936Snectar As the countermeasure turned out to be incompatible with some 6093100936Snectar broken SSL implementations, the new option is part of SSL_OP_ALL. 6094100936Snectar SSL_OP_ALL is usually employed when compatibility with weird SSL 6095100936Snectar implementations is desired (e.g. '-bugs' option to 's_client' and 6096100936Snectar 's_server'), so the new option is automatically set in many 6097100936Snectar applications. 6098100936Snectar [Bodo Moeller] 6099100936Snectar 6100100936Snectar *) Changes in security patch: 6101100936Snectar 6102100936Snectar Changes marked "(CHATS)" were sponsored by the Defense Advanced 6103100936Snectar Research Projects Agency (DARPA) and Air Force Research Laboratory, 6104100936Snectar Air Force Materiel Command, USAF, under agreement number 6105100936Snectar F30602-01-2-0537. 6106100936Snectar 6107100936Snectar *) Add various sanity checks to asn1_get_length() to reject 6108100936Snectar the ASN1 length bytes if they exceed sizeof(long), will appear 6109100936Snectar negative or the content length exceeds the length of the 6110160814Ssimon supplied buffer. (CVE-2002-0659) 6111100936Snectar [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6112100936Snectar 6113100936Snectar *) Assertions for various potential buffer overflows, not known to 6114100936Snectar happen in practice. 6115100936Snectar [Ben Laurie (CHATS)] 6116100936Snectar 6117100936Snectar *) Various temporary buffers to hold ASCII versions of integers were 6118160814Ssimon too small for 64 bit platforms. (CVE-2002-0655) 6119100936Snectar [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 6120100936Snectar 6121100936Snectar *) Remote buffer overflow in SSL3 protocol - an attacker could 6122160814Ssimon supply an oversized session ID to a client. (CVE-2002-0656) 6123100936Snectar [Ben Laurie (CHATS)] 6124100936Snectar 6125100936Snectar *) Remote buffer overflow in SSL2 protocol - an attacker could 6126160814Ssimon supply an oversized client master key. (CVE-2002-0656) 6127100936Snectar [Ben Laurie (CHATS)] 6128100936Snectar 6129100928Snectar Changes between 0.9.6c and 0.9.6d [9 May 2002] 6130100928Snectar 6131100928Snectar *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 6132100928Snectar encoded as NULL) with id-dsa-with-sha1. 6133100928Snectar [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 6134100928Snectar 6135100928Snectar *) Check various X509_...() return values in apps/req.c. 6136100928Snectar [Nils Larsch <nla@trustcenter.de>] 6137100928Snectar 6138100928Snectar *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 6139100928Snectar an end-of-file condition would erronously be flagged, when the CRLF 6140100928Snectar was just at the end of a processed block. The bug was discovered when 6141100928Snectar processing data through a buffering memory BIO handing the data to a 6142100928Snectar BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 6143100928Snectar <ptsekov@syntrex.com> and Nedelcho Stanev. 6144100928Snectar [Lutz Jaenicke] 6145100928Snectar 6146100928Snectar *) Implement a countermeasure against a vulnerability recently found 6147100928Snectar in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 6148100928Snectar before application data chunks to avoid the use of known IVs 6149100928Snectar with data potentially chosen by the attacker. 6150100928Snectar [Bodo Moeller] 6151100928Snectar 6152100928Snectar *) Fix length checks in ssl3_get_client_hello(). 6153100928Snectar [Bodo Moeller] 6154100928Snectar 6155100928Snectar *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 6156100928Snectar to prevent ssl3_read_internal() from incorrectly assuming that 6157100928Snectar ssl3_read_bytes() found application data while handshake 6158100928Snectar processing was enabled when in fact s->s3->in_read_app_data was 6159100928Snectar merely automatically cleared during the initial handshake. 6160100928Snectar [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 6161100928Snectar 6162100928Snectar *) Fix object definitions for Private and Enterprise: they were not 6163100928Snectar recognized in their shortname (=lowercase) representation. Extend 6164100928Snectar obj_dat.pl to issue an error when using undefined keywords instead 6165100928Snectar of silently ignoring the problem (Svenning Sorensen 6166100928Snectar <sss@sss.dnsalias.net>). 6167100928Snectar [Lutz Jaenicke] 6168100928Snectar 6169100928Snectar *) Fix DH_generate_parameters() so that it works for 'non-standard' 6170100928Snectar generators, i.e. generators other than 2 and 5. (Previously, the 6171100928Snectar code did not properly initialise the 'add' and 'rem' values to 6172100928Snectar BN_generate_prime().) 6173100928Snectar 6174100928Snectar In the new general case, we do not insist that 'generator' is 6175100928Snectar actually a primitive root: This requirement is rather pointless; 6176100928Snectar a generator of the order-q subgroup is just as good, if not 6177100928Snectar better. 6178100928Snectar [Bodo Moeller] 6179100928Snectar 6180100928Snectar *) Map new X509 verification errors to alerts. Discovered and submitted by 6181100928Snectar Tom Wu <tom@arcot.com>. 6182100928Snectar [Lutz Jaenicke] 6183100928Snectar 6184100928Snectar *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 6185100928Snectar returning non-zero before the data has been completely received 6186100928Snectar when using non-blocking I/O. 6187100928Snectar [Bodo Moeller; problem pointed out by John Hughes] 6188100928Snectar 6189100928Snectar *) Some of the ciphers missed the strength entry (SSL_LOW etc). 6190100928Snectar [Ben Laurie, Lutz Jaenicke] 6191100928Snectar 6192100928Snectar *) Fix bug in SSL_clear(): bad sessions were not removed (found by 6193100928Snectar Yoram Zahavi <YoramZ@gilian.com>). 6194100928Snectar [Lutz Jaenicke] 6195100928Snectar 6196100928Snectar *) Add information about CygWin 1.3 and on, and preserve proper 6197100928Snectar configuration for the versions before that. 6198100928Snectar [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 6199100928Snectar 6200100928Snectar *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 6201100928Snectar check whether we deal with a copy of a session and do not delete from 6202100928Snectar the cache in this case. Problem reported by "Izhar Shoshani Levi" 6203100928Snectar <izhar@checkpoint.com>. 6204100928Snectar [Lutz Jaenicke] 6205100928Snectar 6206100928Snectar *) Do not store session data into the internal session cache, if it 6207100928Snectar is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 6208100928Snectar flag is set). Proposed by Aslam <aslam@funk.com>. 6209100928Snectar [Lutz Jaenicke] 6210100928Snectar 6211100928Snectar *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 6212100928Snectar value is 0. 6213100928Snectar [Richard Levitte] 6214100928Snectar 6215109998Smarkm *) [In 0.9.6d-engine release:] 6216109998Smarkm Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 6217100928Snectar [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 6218100928Snectar 6219100928Snectar *) Add the configuration target linux-s390x. 6220100928Snectar [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 6221100928Snectar 6222100928Snectar *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 6223100928Snectar ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 6224100928Snectar variable as an indication that a ClientHello message has been 6225100928Snectar received. As the flag value will be lost between multiple 6226100928Snectar invocations of ssl3_accept when using non-blocking I/O, the 6227100928Snectar function may not be aware that a handshake has actually taken 6228100928Snectar place, thus preventing a new session from being added to the 6229100928Snectar session cache. 6230100928Snectar 6231100928Snectar To avoid this problem, we now set s->new_session to 2 instead of 6232100928Snectar using a local variable. 6233100928Snectar [Lutz Jaenicke, Bodo Moeller] 6234100928Snectar 6235100928Snectar *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 6236100928Snectar if the SSL_R_LENGTH_MISMATCH error is detected. 6237100928Snectar [Geoff Thorpe, Bodo Moeller] 6238100928Snectar 6239100928Snectar *) New 'shared_ldflag' column in Configure platform table. 6240100928Snectar [Richard Levitte] 6241100928Snectar 6242100928Snectar *) Fix EVP_CIPHER_mode macro. 6243100928Snectar ["Dan S. Camper" <dan@bti.net>] 6244100928Snectar 6245100928Snectar *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 6246100928Snectar type, we must throw them away by setting rr->length to 0. 6247100928Snectar [D P Chang <dpc@qualys.com>] 6248100928Snectar 624989837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 625089837Skris 625189837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 625289837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 625389837Skris worked incorrectly for those cases where range = 10..._2 and 625489837Skris 3*range is two bits longer than range.) 625589837Skris [Bodo Moeller] 625689837Skris 625789837Skris *) Only add signing time to PKCS7 structures if it is not already 625889837Skris present. 625989837Skris [Steve Henson] 626089837Skris 626189837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 626289837Skris OBJ_ld_ce should be OBJ_id_ce. 626389837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 626489837Skris incorrect (cf. RFC 3039). 626589837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 626689837Skris 626789837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 626889837Skris returns early because it has nothing to do. 626989837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 627089837Skris 627189837Skris *) [In 0.9.6c-engine release:] 627289837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 627389837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 627489837Skris 627589837Skris *) [In 0.9.6c-engine release:] 627689837Skris Add support for Cryptographic Appliance's keyserver technology. 627789837Skris (Use engine 'keyclient') 627889837Skris [Cryptographic Appliances and Geoff Thorpe] 627989837Skris 628089837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 628189837Skris is called via tools/c89.sh because arguments have to be 628289837Skris rearranged (all '-L' options must appear before the first object 628389837Skris modules). 628489837Skris [Richard Shapiro <rshapiro@abinitio.com>] 628589837Skris 628689837Skris *) [In 0.9.6c-engine release:] 628789837Skris Add support for Broadcom crypto accelerator cards, backported 628889837Skris from 0.9.7. 628989837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 629089837Skris 629189837Skris *) [In 0.9.6c-engine release:] 629289837Skris Add support for SureWare crypto accelerator cards from 629389837Skris Baltimore Technologies. (Use engine 'sureware') 629489837Skris [Baltimore Technologies and Mark Cox] 629589837Skris 629689837Skris *) [In 0.9.6c-engine release:] 629789837Skris Add support for crypto accelerator cards from Accelerated 629889837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 629989837Skris [AEP Inc. and Mark Cox] 630089837Skris 630189837Skris *) Add a configuration entry for gcc on UnixWare. 630289837Skris [Gary Benson <gbenson@redhat.com>] 630389837Skris 630489837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 630589837Skris messages are stored in a single piece (fixed-length part and 630689837Skris variable-length part combined) and fix various bugs found on the way. 630789837Skris [Bodo Moeller] 630889837Skris 630989837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 631089837Skris instead. BIO_gethostbyname() does not know what timeouts are 631189837Skris appropriate, so entries would stay in cache even when they have 631289837Skris become invalid. 631389837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 631489837Skris 631589837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 631689837Skris faced with a pathologically small ClientHello fragment that does 631789837Skris not contain client_version: Instead of aborting with an error, 631889837Skris simply choose the highest available protocol version (i.e., 631989837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 632089837Skris messages are never sent like this, but this change gives us 632189837Skris strictly correct behaviour at least for TLS. 632289837Skris [Bodo Moeller] 632389837Skris 632489837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 632589837Skris never resets s->method to s->ctx->method when called from within 632689837Skris one of the SSL handshake functions. 632789837Skris [Bodo Moeller; problem pointed out by Niko Baric] 632889837Skris 632989837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 633089837Skris (sent using the client's version number) if client_version is 633189837Skris smaller than the protocol version in use. Also change 633289837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 633389837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 633489837Skris the client will at least see that alert. 633589837Skris [Bodo Moeller] 633689837Skris 633789837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 633889837Skris correctly. 633989837Skris [Bodo Moeller] 634089837Skris 634189837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 634289837Skris client receives HelloRequest while in a handshake. 634389837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 634489837Skris 634589837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 634689837Skris should end in 'break', not 'goto end' which circuments various 634789837Skris cleanups done in state SSL_ST_OK. But session related stuff 634889837Skris must be disabled for SSL_ST_OK in the case that we just sent a 634989837Skris HelloRequest. 635089837Skris 635189837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 635289837Skris before just sending a HelloRequest. 635389837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 635489837Skris 635589837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 635689837Skris reveal whether illegal block cipher padding was found or a MAC 635789837Skris verification error occured. (Neither SSLerr() codes nor alerts 635889837Skris are directly visible to potential attackers, but the information 635989837Skris may leak via logfiles.) 636089837Skris 636189837Skris Similar changes are not required for the SSL 2.0 implementation 636289837Skris because the number of padding bytes is sent in clear for SSL 2.0, 636389837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 636489837Skris failed to verify that the purported number of padding bytes is in 636589837Skris the legal range. 636689837Skris [Bodo Moeller] 636789837Skris 636889837Skris *) Add OpenUNIX-8 support including shared libraries 636989837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 637089837Skris [Lutz Jaenicke] 637189837Skris 637289837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 637389837Skris 'wristwatch attack' using huge encoding parameters (cf. 637489837Skris James H. Manger's CRYPTO 2001 paper). Note that the 637589837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 637689837Skris encoding parameters and hence was not vulnerable. 637789837Skris [Bodo Moeller] 637889837Skris 637989837Skris *) BN_sqr() bug fix. 638089837Skris [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 638189837Skris 638289837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 638389837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 638489837Skris followed by modular reduction. 638589837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 638689837Skris 638789837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 638889837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 638989837Skris [Bodo Moeller] 639089837Skris 639189837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 639289837Skris This function was broken, as the check for a new client hello message 639389837Skris to handle SGC did not allow these large messages. 639489837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 639589837Skris [Lutz Jaenicke] 639689837Skris 639789837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 639889837Skris [Lutz Jaenicke] 639989837Skris 640089837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 640189837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 640289837Skris [Lutz Jaenicke] 640389837Skris 640489837Skris *) Rework the configuration and shared library support for Tru64 Unix. 640589837Skris The configuration part makes use of modern compiler features and 640689837Skris still retains old compiler behavior for those that run older versions 640789837Skris of the OS. The shared library support part includes a variant that 640889837Skris uses the RPATH feature, and is available through the special 640989837Skris configuration target "alpha-cc-rpath", which will never be selected 641089837Skris automatically. 641189837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 641289837Skris 641389837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 641489837Skris with the same message size as in ssl3_get_certificate_request(). 641589837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 641689837Skris messages might inadvertently be reject as too long. 641789837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 641889837Skris 641989837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 642089837Skris [Andy Polyakov] 642189837Skris 642289837Skris *) Modified SSL library such that the verify_callback that has been set 642389837Skris specificly for an SSL object with SSL_set_verify() is actually being 642489837Skris used. Before the change, a verify_callback set with this function was 642589837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 642689837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 642789837Skris to allow the necessary settings. 642889837Skris [Lutz Jaenicke] 642989837Skris 643089837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 643189837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 643289837Skris done automatically (in contradiction to the requirements of the C 643389837Skris standard). This made problems when used from OpenSSH. 643489837Skris [Lutz Jaenicke] 643589837Skris 643689837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 643789837Skris dh->length and always used 643889837Skris 643989837Skris BN_rand_range(priv_key, dh->p). 644089837Skris 644189837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 644289837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 644389837Skris dh->length (recommended exponent length) is much smaller than the 644489837Skris length of dh->p. We could use BN_rand_range() if the order of 644589837Skris the subgroup was stored in the DH structure, but we only have 644689837Skris dh->length. 644789837Skris 644889837Skris So switch back to 644989837Skris 645089837Skris BN_rand(priv_key, l, ...) 645189837Skris 645289837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 645389837Skris otherwise. 645489837Skris [Bodo Moeller] 645589837Skris 645689837Skris *) In 645789837Skris 645889837Skris RSA_eay_public_encrypt 645989837Skris RSA_eay_private_decrypt 646089837Skris RSA_eay_private_encrypt (signing) 646189837Skris RSA_eay_public_decrypt (signature verification) 646289837Skris 646389837Skris (default implementations for RSA_public_encrypt, 646489837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 646589837Skris always reject numbers >= n. 646689837Skris [Bodo Moeller] 646789837Skris 646889837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 646989837Skris to synchronize access to 'locking_thread'. This is necessary on 647089837Skris systems where access to 'locking_thread' (an 'unsigned long' 647189837Skris variable) is not atomic. 647289837Skris [Bodo Moeller] 647389837Skris 647489837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 647589837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 647689837Skris a race condition if 0 is a valid thread ID. 647789837Skris [Travis Vitek <vitek@roguewave.com>] 647889837Skris 647989837Skris *) Add support for shared libraries under Irix. 648089837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 648189837Skris 648289837Skris *) Add configuration option to build on Linux on both big-endian and 648389837Skris little-endian MIPS. 648489837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 648589837Skris 648689837Skris *) Add the possibility to create shared libraries on HP-UX. 648789837Skris [Richard Levitte] 648889837Skris 648979998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 649079998Skris 649179998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 649279998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 649379998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 649479998Skris PRNG state recovery was possible based on the output of 649579998Skris one PRNG request appropriately sized to gain knowledge on 649679998Skris 'md' followed by enough consecutive 1-byte PRNG requests 649779998Skris to traverse all of 'state'. 649879998Skris 649979998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 650079998Skris during PRNG output generation, hash all of the previous 650179998Skris 'md_local' value, not just the half used for PRNG output. 650279998Skris 650379998Skris 2. Make the number of bytes from 'state' included into the hash 650479998Skris independent from the number of PRNG bytes requested. 650579998Skris 650679998Skris The first measure alone would be sufficient to avoid 650779998Skris Markku-Juhani's attack. (Actually it had never occurred 650879998Skris to me that the half of 'md_local' used for chaining was the 650979998Skris half from which PRNG output bytes were taken -- I had always 651079998Skris assumed that the secret half would be used.) The second 651179998Skris measure makes sure that additional data from 'state' is never 651279998Skris mixed into 'md_local' in small portions; this heuristically 651379998Skris further strengthens the PRNG. 651479998Skris [Bodo Moeller] 651579998Skris 651679998Skris *) Fix crypto/bn/asm/mips3.s. 651779998Skris [Andy Polyakov] 651879998Skris 651979998Skris *) When only the key is given to "enc", the IV is undefined. Print out 652079998Skris an error message in this case. 652179998Skris [Lutz Jaenicke] 652279998Skris 652379998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 652479998Skris [Steve Henson] 652579998Skris 652679998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 652779998Skris positive and less than q. 652879998Skris [Bodo Moeller] 652979998Skris 653079998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 653179998Skris used: it isn't thread safe and the add_lock_callback should handle 653279998Skris that itself. 653379998Skris [Paul Rose <Paul.Rose@bridge.com>] 653479998Skris 653579998Skris *) Verify that incoming data obeys the block size in 653679998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 653779998Skris [Bodo Moeller] 653879998Skris 653979998Skris *) Fix OAEP check. 654079998Skris [Ulf M�ller, Bodo M�ller] 654179998Skris 654279998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 654389837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 654479998Skris when fixing the server behaviour for backwards-compatible 'client 654579998Skris hello' messages. (Note that the attack is impractical against 654679998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 654779998Skris means that the probability of guessing a valid ciphertext is 654879998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 654979998Skris paper.) 655079998Skris 655179998Skris Before 0.9.5, the countermeasure (hide the error by generating a 655279998Skris random 'decryption result') did not work properly because 655379998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 655479998Skris detect the supposedly ignored error. 655579998Skris 655679998Skris Both problems are now fixed. 655779998Skris [Bodo Moeller] 655879998Skris 655979998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 656079998Skris (previously it was 1024). 656179998Skris [Bodo Moeller] 656279998Skris 656379998Skris *) Fix for compatibility mode trust settings: ignore trust settings 656479998Skris unless some valid trust or reject settings are present. 656579998Skris [Steve Henson] 656679998Skris 656779998Skris *) Fix for blowfish EVP: its a variable length cipher. 656879998Skris [Steve Henson] 656979998Skris 657079998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 657179998Skris parameters in DSA public key structures and return an error in the 657279998Skris DSA routines if parameters are absent. 657379998Skris [Steve Henson] 657479998Skris 657579998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 657679998Skris in the current directory if neither $RANDFILE nor $HOME was set. 657779998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 657879998Skris caused some confusion to Windows users who haven't defined $HOME. 657979998Skris Thus RAND_file_name() is changed again: e_os.h can define a 658079998Skris DEFAULT_HOME, which will be used if $HOME is not set. 658179998Skris For Windows, we use "C:"; on other platforms, we still require 658279998Skris environment variables. 658379998Skris 658479998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 658579998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 658679998Skris having multiple threads call RAND_poll() concurrently. 658779998Skris [Bodo Moeller] 658879998Skris 658979998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 659079998Skris combination of a flag and a thread ID variable. 659179998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 659279998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 659389837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 659479998Skris that they do not hold after the first thread unsets add_do_not_lock). 659579998Skris [Bodo Moeller] 659679998Skris 659779998Skris *) Change bctest again: '-x' expressions are not available in all 659879998Skris versions of 'test'. 659979998Skris [Bodo Moeller] 660079998Skris 660176866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 660272613Skris 660376866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 660476866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 660576866Skris 660676866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 660776866Skris the default extension for executables, if any. Also, make the perl 660876866Skris scripts that use symlink() to test if it really exists and use "cp" 660976866Skris if it doesn't. All this made OpenSSL compilable and installable in 661076866Skris CygWin. 661176866Skris [Richard Levitte] 661276866Skris 661376866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 661476866Skris If SEQUENCE is length is indefinite just set c->slen to the total 661576866Skris amount of data available. 661676866Skris [Steve Henson, reported by shige@FreeBSD.org] 661776866Skris [This change does not apply to 0.9.7.] 661876866Skris 661976866Skris *) Change bctest to avoid here-documents inside command substitution 662076866Skris (workaround for FreeBSD /bin/sh bug). 662176866Skris For compatibility with Ultrix, avoid shell functions (introduced 662276866Skris in the bctest version that searches along $PATH). 662376866Skris [Bodo Moeller] 662476866Skris 662576866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 662676866Skris with des_encrypt() defined on some operating systems, like Solaris 662776866Skris and UnixWare. 662876866Skris [Richard Levitte] 662976866Skris 663076866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 663176866Skris On the Importance of Eliminating Errors in Cryptographic 663276866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 663376866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 663476866Skris [Ulf Moeller] 663576866Skris 663676866Skris *) MIPS assembler BIGNUM division bug fix. 663776866Skris [Andy Polyakov] 663876866Skris 663976866Skris *) Disabled incorrect Alpha assembler code. 664076866Skris [Richard Levitte] 664176866Skris 664276866Skris *) Fix PKCS#7 decode routines so they correctly update the length 664376866Skris after reading an EOC for the EXPLICIT tag. 664476866Skris [Steve Henson] 664576866Skris [This change does not apply to 0.9.7.] 664676866Skris 664776866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 664876866Skris if a 3DES key was generated with a 0 initial byte. Include 664976866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 665076866Skris (but broken) behaviour. 665176866Skris [Steve Henson] 665276866Skris 665376866Skris *) Enhance bctest to search for a working bc along $PATH and print 665476866Skris it when found. 665576866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 665676866Skris 665776866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 665876866Skris don't write to the wrong index in ERR_set_error_data. 665976866Skris [Bodo Moeller] 666076866Skris 666176866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 666276866Skris did not exist. 666376866Skris [Bodo Moeller] 666476866Skris 666576866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 666676866Skris [Jeremy Cooper <jeremy@baymoo.org>] 666776866Skris 666876866Skris *) Make it possible to reuse SSLv2 sessions. 666976866Skris [Richard Levitte] 667076866Skris 667176866Skris *) In copy_email() check for >= 0 as a return value for 667276866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 667376866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 667476866Skris 667576866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 667676866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 667776866Skris PKCS7_verify() fails with non detached data. 667876866Skris [Steve Henson] 667976866Skris 668076866Skris *) Don't use getenv in library functions when run as setuid/setgid. 668176866Skris New function OPENSSL_issetugid(). 668276866Skris [Ulf Moeller] 668376866Skris 668476866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 668576866Skris due to incorrect handling of multi-threading: 668676866Skris 668776866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 668876866Skris 668976866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 669076866Skris 669176866Skris 3. Count how many times MemCheck_off() has been called so that 669276866Skris nested use can be treated correctly. This also avoids 669376866Skris inband-signalling in the previous code (which relied on the 669476866Skris assumption that thread ID 0 is impossible). 669576866Skris [Bodo Moeller] 669676866Skris 669776866Skris *) Add "-rand" option also to s_client and s_server. 669876866Skris [Lutz Jaenicke] 669976866Skris 670076866Skris *) Fix CPU detection on Irix 6.x. 670176866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 670276866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 670376866Skris 670476866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 670576866Skris was empty. 670676866Skris [Steve Henson] 670776866Skris [This change does not apply to 0.9.7.] 670876866Skris 670976866Skris *) Use the cached encoding of an X509_NAME structure rather than 671076866Skris copying it. This is apparently the reason for the libsafe "errors" 671176866Skris but the code is actually correct. 671276866Skris [Steve Henson] 671376866Skris 671472613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 671572613Skris Bleichenbacher's DSA attack. 671676866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 671776866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 671876866Skris and leaves the highest bit random. 671976866Skris [Ulf Moeller, Bodo Moeller] 672072613Skris 672172613Skris *) In the NCONF_...-based implementations for CONF_... queries 672272613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 672372613Skris a temporary CONF structure with the data component set to NULL 672472613Skris (which gives segmentation faults in lh_retrieve). 672572613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 672672613Skris CONF_get_number (which may use environment variables) and directly 672772613Skris return NULL from CONF_get_section. 672872613Skris [Bodo Moeller] 672972613Skris 673072613Skris *) Fix potential buffer overrun for EBCDIC. 673172613Skris [Ulf Moeller] 673272613Skris 673372613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 673472613Skris keyUsage if basicConstraints absent for a CA. 673572613Skris [Steve Henson] 673672613Skris 673772613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 673872613Skris is more generally accepted (no spaces before the semicolon), since 673972613Skris some programs can't parse those values properly otherwise. Also make 674072613Skris sure BIO's that break lines after each write do not create invalid 674172613Skris headers. 674272613Skris [Richard Levitte] 674372613Skris 674472613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 674572613Skris macros previously used would not encode an empty SEQUENCE OF 674672613Skris and break the signature. 674772613Skris [Steve Henson] 674876866Skris [This change does not apply to 0.9.7.] 674972613Skris 675072613Skris *) Zero the premaster secret after deriving the master secret in 675172613Skris DH ciphersuites. 675272613Skris [Steve Henson] 675372613Skris 675472613Skris *) Add some EVP_add_digest_alias registrations (as found in 675572613Skris OpenSSL_add_all_digests()) to SSL_library_init() 675672613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 675772613Skris compatibility with peers using X.509 certificates 675872613Skris with unconventional AlgorithmIdentifier OIDs. 675972613Skris [Bodo Moeller] 676072613Skris 676172613Skris *) Fix for Irix with NO_ASM. 676272613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 676372613Skris 676472613Skris *) ./config script fixes. 676572613Skris [Ulf Moeller, Richard Levitte] 676672613Skris 676772613Skris *) Fix 'openssl passwd -1'. 676872613Skris [Bodo Moeller] 676972613Skris 677072613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 677172613Skris terminated strings whose length is passed in the passlen 677272613Skris parameter, for example from PEM callbacks. This was done 677372613Skris by adding an extra length parameter to asc2uni(). 677472613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 677572613Skris 677672613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 677772613Skris call failed, free the DSA structure. 677872613Skris [Bodo Moeller] 677972613Skris 678072613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 678172613Skris These are present in some PKCS#12 files. 678272613Skris [Steve Henson] 678372613Skris 678472613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 678572613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 678672613Skris when writing a 32767 byte record. 678772613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 678872613Skris 678972613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 679072613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 679172613Skris 679272613Skris (RSA objects have a reference count access to which is protected 679372613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 679472613Skris so they are meant to be shared between threads.) 679572613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 679672613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 679772613Skris 679872613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 679972613Skris [Bodo Moeller] 680072613Skris 680176866Skris *) Use better test patterns in bntest. 680276866Skris [Ulf M�ller] 680376866Skris 680472613Skris *) rand_win.c fix for Borland C. 680572613Skris [Ulf M�ller] 680672613Skris 680772613Skris *) BN_rshift bugfix for n == 0. 680872613Skris [Bodo Moeller] 680972613Skris 681076866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 681176866Skris so that 'make test' does not abort just because 'bc' is broken. 681276866Skris [Bodo Moeller] 681376866Skris 681472613Skris *) Store verify_result within SSL_SESSION also for client side to 681572613Skris avoid potential security hole. (Re-used sessions on the client side 681672613Skris always resulted in verify_result==X509_V_OK, not using the original 681772613Skris result of the server certificate verification.) 681872613Skris [Lutz Jaenicke] 681972613Skris 682072613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 682172613Skris SSL3_RT_APPLICATION_DATA, return 0. 682272613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 682372613Skris [Bodo Moeller] 682472613Skris 682572613Skris *) Fix SSL_peek: 682672613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 682772613Skris releases, have been re-implemented by renaming the previous 682872613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 682972613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 683072613Skris to them. The new ssl[23]_{read,peek} functions are calls to 683172613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 683272613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 683372613Skris does the actual work for ssl3_read_internal. 683472613Skris [Bodo Moeller] 683572613Skris 683676866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 683776866Skris the method-specific "init()" handler. Also clean up ex_data after 683876866Skris calling the method-specific "finish()" handler. Previously, this was 683976866Skris happening the other way round. 684076866Skris [Geoff Thorpe] 684176866Skris 684272613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 684372613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 684472613Skris [Bodo Moeller] 684572613Skris 684676866Skris *) Make sure that shared libraries get the internal name engine with 684776866Skris the full version number and not just 0. This should mark the 684876866Skris shared libraries as not backward compatible. Of course, this should 684976866Skris be changed again when we can guarantee backward binary compatibility. 685076866Skris [Richard Levitte] 685176866Skris 685272613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 685372613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 685472613Skris 685576866Skris *) Rework the system to generate shared libraries: 685676866Skris 685776866Skris - Make note of the expected extension for the shared libraries and 685876866Skris if there is a need for symbolic links from for example libcrypto.so.0 685976866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 686076866Skris that. 686176866Skris 686276866Skris - Make as few rebuilds of the shared libraries as possible. 686376866Skris 686476866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 686576866Skris 686676866Skris - When installing, install the shared libraries separately from the 686776866Skris static ones. 686876866Skris [Richard Levitte] 686976866Skris 687072613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 687172613Skris 687272613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 687372613Skris and not in SSL_clear because the latter is also used by the 687472613Skris accept/connect functions; previously, the settings made by 687572613Skris SSL_set_read_ahead would be lost during the handshake. 687672613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 687772613Skris 687872613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 687972613Skris Previously, it would create entries for disableed algorithms no 688072613Skris matter what. 688172613Skris [Richard Levitte] 688272613Skris 688372613Skris *) Added several new manual pages for SSL_* function. 688472613Skris [Lutz Jaenicke] 688572613Skris 688668651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 688768651Skris 688868651Skris *) In ssl23_get_client_hello, generate an error message when faced 688968651Skris with an initial SSL 3.0/TLS record that is too small to contain the 689068651Skris first two bytes of the ClientHello message, i.e. client_version. 689168651Skris (Note that this is a pathologic case that probably has never happened 689268651Skris in real life.) The previous approach was to use the version number 689368651Skris from the record header as a substitute; but our protocol choice 689468651Skris should not depend on that one because it is not authenticated 689568651Skris by the Finished messages. 689668651Skris [Bodo Moeller] 689768651Skris 689868651Skris *) More robust randomness gathering functions for Windows. 689968651Skris [Jeffrey Altman <jaltman@columbia.edu>] 690068651Skris 690168651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 690268651Skris not set then we don't setup the error code for issuer check errors 690368651Skris to avoid possibly overwriting other errors which the callback does 690468651Skris handle. If an application does set the flag then we assume it knows 690568651Skris what it is doing and can handle the new informational codes 690668651Skris appropriately. 690768651Skris [Steve Henson] 690868651Skris 690968651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 691068651Skris a general "ANY" type, as such it should be able to decode anything 691168651Skris including tagged types. However it didn't check the class so it would 691268651Skris wrongly interpret tagged types in the same way as their universal 691368651Skris counterpart and unknown types were just rejected. Changed so that the 691468651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 691568651Skris that is the encoding is stored intact. There is also a new type 691668651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 691768651Skris case we have no idea what the actual type is so we just lump them all 691868651Skris together. 691968651Skris [Steve Henson] 692068651Skris 692168651Skris *) On VMS, stdout may very well lead to a file that is written to 692268651Skris in a record-oriented fashion. That means that every write() will 692368651Skris write a separate record, which will be read separately by the 692468651Skris programs trying to read from it. This can be very confusing. 692568651Skris 692668651Skris The solution is to put a BIO filter in the way that will buffer 692768651Skris text until a linefeed is reached, and then write everything a 692868651Skris line at a time, so every record written will be an actual line, 692968651Skris not chunks of lines and not (usually doesn't happen, but I've 693068651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 693168651Skris the answer. 693268651Skris 693368651Skris Currently, it's a VMS-only method, because that's where it has 693468651Skris been tested well enough. 693568651Skris [Richard Levitte] 693668651Skris 693768651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 693868651Skris it can return incorrect results. 693968651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 694068651Skris but it was in 0.9.6-beta[12].) 694168651Skris [Bodo Moeller] 694268651Skris 694368651Skris *) Disable the check for content being present when verifying detached 694468651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 694568651Skris include zero length content when signing messages. 694668651Skris [Steve Henson] 694768651Skris 694868651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 694968651Skris BIO_ctrl (for BIO pairs). 695068651Skris [Bodo M�ller] 695168651Skris 695268651Skris *) Add DSO method for VMS. 695368651Skris [Richard Levitte] 695468651Skris 695568651Skris *) Bug fix: Montgomery multiplication could produce results with the 695668651Skris wrong sign. 695768651Skris [Ulf M�ller] 695868651Skris 695968651Skris *) Add RPM specification openssl.spec and modify it to build three 696068651Skris packages. The default package contains applications, application 696168651Skris documentation and run-time libraries. The devel package contains 696268651Skris include files, static libraries and function documentation. The 696368651Skris doc package contains the contents of the doc directory. The original 696468651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 696568651Skris [Richard Levitte] 696668651Skris 696768651Skris *) Add a large number of documentation files for many SSL routines. 696868651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 696968651Skris 697068651Skris *) Add a configuration entry for Sony News 4. 697168651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 697268651Skris 697368651Skris *) Don't set the two most significant bits to one when generating a 697468651Skris random number < q in the DSA library. 697568651Skris [Ulf M�ller] 697668651Skris 697768651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 697868651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 697968651Skris the underlying transport is blocking) if a handshake took place. 698068651Skris (The default behaviour is needed by applications such as s_client 698168651Skris and s_server that use select() to determine when to use SSL_read; 698268651Skris but for applications that know in advance when to expect data, it 698368651Skris just makes things more complicated.) 698468651Skris [Bodo Moeller] 698568651Skris 698668651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 698768651Skris from EGD. 698868651Skris [Ben Laurie] 698968651Skris 699068651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 699168651Skris work better on such systems. 699268651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 699368651Skris 699468651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 699568651Skris Update PKCS12_parse() so it copies the friendlyName and the 699668651Skris keyid to the certificates aux info. 699768651Skris [Steve Henson] 699868651Skris 699968651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 700068651Skris if there was more than one signature. 700168651Skris [Sven Uszpelkat <su@celocom.de>] 700268651Skris 700368651Skris *) Major change in util/mkdef.pl to include extra information 700468651Skris about each symbol, as well as presentig variables as well 700568651Skris as functions. This change means that there's n more need 700668651Skris to rebuild the .num files when some algorithms are excluded. 700768651Skris [Richard Levitte] 700868651Skris 700968651Skris *) Allow the verify time to be set by an application, 701068651Skris rather than always using the current time. 701168651Skris [Steve Henson] 701268651Skris 701368651Skris *) Phase 2 verify code reorganisation. The certificate 701468651Skris verify code now looks up an issuer certificate by a 701568651Skris number of criteria: subject name, authority key id 701668651Skris and key usage. It also verifies self signed certificates 701768651Skris by the same criteria. The main comparison function is 701868651Skris X509_check_issued() which performs these checks. 701968651Skris 702068651Skris Lot of changes were necessary in order to support this 702168651Skris without completely rewriting the lookup code. 702268651Skris 702368651Skris Authority and subject key identifier are now cached. 702468651Skris 702568651Skris The LHASH 'certs' is X509_STORE has now been replaced 702668651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 702768651Skris LHASH can't store or retrieve multiple objects with 702868651Skris the same hash value. 702968651Skris 703068651Skris As a result various functions (which were all internal 703168651Skris use only) have changed to handle the new X509_STORE 703268651Skris structure. This will break anything that messed round 703368651Skris with X509_STORE internally. 703468651Skris 703568651Skris The functions X509_STORE_add_cert() now checks for an 703668651Skris exact match, rather than just subject name. 703768651Skris 703868651Skris The X509_STORE API doesn't directly support the retrieval 703968651Skris of multiple certificates matching a given criteria, however 704068651Skris this can be worked round by performing a lookup first 704168651Skris (which will fill the cache with candidate certificates) 704268651Skris and then examining the cache for matches. This is probably 704368651Skris the best we can do without throwing out X509_LOOKUP 704468651Skris entirely (maybe later...). 704568651Skris 704668651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 704768651Skris 704868651Skris All certificate lookup operations now go via a get_issuer() 704968651Skris callback. Although this currently uses an X509_STORE it 705068651Skris can be replaced by custom lookups. This is a simple way 705168651Skris to bypass the X509_STORE hackery necessary to make this 705268651Skris work and makes it possible to use more efficient techniques 705368651Skris in future. A very simple version which uses a simple 705468651Skris STACK for its trusted certificate store is also provided 705568651Skris using X509_STORE_CTX_trusted_stack(). 705668651Skris 705768651Skris The verify_cb() and verify() callbacks now have equivalents 705868651Skris in the X509_STORE_CTX structure. 705968651Skris 706068651Skris X509_STORE_CTX also has a 'flags' field which can be used 706168651Skris to customise the verify behaviour. 706268651Skris [Steve Henson] 706368651Skris 706468651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 706568651Skris excludes S/MIME capabilities. 706668651Skris [Steve Henson] 706768651Skris 706868651Skris *) When a certificate request is read in keep a copy of the 706968651Skris original encoding of the signed data and use it when outputing 707068651Skris again. Signatures then use the original encoding rather than 707168651Skris a decoded, encoded version which may cause problems if the 707268651Skris request is improperly encoded. 707368651Skris [Steve Henson] 707468651Skris 707568651Skris *) For consistency with other BIO_puts implementations, call 707668651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 707768651Skris BIO_write(b, ...). 707868651Skris 707968651Skris In BIO_puts, increment b->num_write as in BIO_write. 708068651Skris [Peter.Sylvester@EdelWeb.fr] 708168651Skris 708268651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 708368651Skris BN_zero, we may not return a BIGNUM with an array consisting of 708468651Skris words set to zero.) 708568651Skris [Bodo Moeller] 708668651Skris 708768651Skris *) Avoid calling abort() from within the library when problems are 708868651Skris detected, except if preprocessor symbols have been defined 708968651Skris (such as REF_CHECK, BN_DEBUG etc.). 709068651Skris [Bodo Moeller] 709168651Skris 709268651Skris *) New openssl application 'rsautl'. This utility can be 709368651Skris used for low level RSA operations. DER public key 709468651Skris BIO/fp routines also added. 709568651Skris [Steve Henson] 709668651Skris 709768651Skris *) New Configure entry and patches for compiling on QNX 4. 709868651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 709968651Skris 710068651Skris *) A demo state-machine implementation was sponsored by 710168651Skris Nuron (http://www.nuron.com/) and is now available in 710268651Skris demos/state_machine. 710368651Skris [Ben Laurie] 710468651Skris 710568651Skris *) New options added to the 'dgst' utility for signature 710668651Skris generation and verification. 710768651Skris [Steve Henson] 710868651Skris 710968651Skris *) Unrecognized PKCS#7 content types are now handled via a 711068651Skris catch all ASN1_TYPE structure. This allows unsupported 711168651Skris types to be stored as a "blob" and an application can 711268651Skris encode and decode it manually. 711368651Skris [Steve Henson] 711468651Skris 711568651Skris *) Fix various signed/unsigned issues to make a_strex.c 711668651Skris compile under VC++. 711768651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 711868651Skris 711968651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 712068651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 712168651Skris if passed a NULL BN and its argument was negative. 712268651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 712368651Skris 712468651Skris *) Modification to PKCS#7 encoding routines to output definite 712568651Skris length encoding. Since currently the whole structures are in 712668651Skris memory there's not real point in using indefinite length 712768651Skris constructed encoding. However if OpenSSL is compiled with 712868651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 712968651Skris [Steve Henson] 713068651Skris 713168651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 713268651Skris [Richard Levitte] 713368651Skris 713468651Skris *) Added more prefixes to parse for in the the strings written 713568651Skris through a logging bio, to cover all the levels that are available 713668651Skris through syslog. The prefixes are now: 713768651Skris 713868651Skris PANIC, EMERG, EMR => LOG_EMERG 713968651Skris ALERT, ALR => LOG_ALERT 714068651Skris CRIT, CRI => LOG_CRIT 714168651Skris ERROR, ERR => LOG_ERR 714268651Skris WARNING, WARN, WAR => LOG_WARNING 714368651Skris NOTICE, NOTE, NOT => LOG_NOTICE 714468651Skris INFO, INF => LOG_INFO 714568651Skris DEBUG, DBG => LOG_DEBUG 714668651Skris 714768651Skris and as before, if none of those prefixes are present at the 714868651Skris beginning of the string, LOG_ERR is chosen. 714968651Skris 715068651Skris On Win32, the LOG_* levels are mapped according to this: 715168651Skris 715268651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 715368651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 715468651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 715568651Skris 715668651Skris [Richard Levitte] 715768651Skris 715868651Skris *) Made it possible to reconfigure with just the configuration 715968651Skris argument "reconf" or "reconfigure". The command line arguments 716068651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 716168651Skris and are retrieved from there when reconfiguring. 716268651Skris [Richard Levitte] 716368651Skris 716468651Skris *) MD4 implemented. 716568651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 716668651Skris 716768651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 716868651Skris [Richard Levitte] 716968651Skris 717068651Skris *) The obj_dat.pl script was messing up the sorting of object 717168651Skris names. The reason was that it compared the quoted version 717268651Skris of strings as a result "OCSP" > "OCSP Signing" because 717368651Skris " > SPACE. Changed script to store unquoted versions of 717468651Skris names and add quotes on output. It was also omitting some 717568651Skris names from the lookup table if they were given a default 717668651Skris value (that is if SN is missing it is given the same 717768651Skris value as LN and vice versa), these are now added on the 717868651Skris grounds that if an object has a name we should be able to 717968651Skris look it up. Finally added warning output when duplicate 718068651Skris short or long names are found. 718168651Skris [Steve Henson] 718268651Skris 718368651Skris *) Changes needed for Tandem NSK. 718468651Skris [Scott Uroff <scott@xypro.com>] 718568651Skris 718668651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 718768651Skris RSA_padding_check_SSLv23(), special padding was never detected 718868651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 718968651Skris version rollback attacks was not effective. 719068651Skris 719168651Skris In s23_clnt.c, don't use special rollback-attack detection padding 719268651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 719368651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 719468651Skris SSL 2.0 is the only protocol enabled in the server. 719568651Skris [Bodo Moeller] 719668651Skris 719768651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 719868651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 719968651Skris BIO_dump_indent() are added. 720068651Skris [Richard Levitte] 720168651Skris 720268651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 720368651Skris these print out strings and name structures based on various 720468651Skris flags including RFC2253 support and proper handling of 720568651Skris multibyte characters. Added options to the 'x509' utility 720668651Skris to allow the various flags to be set. 720768651Skris [Steve Henson] 720868651Skris 720968651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 721068651Skris Also change the functions X509_cmp_current_time() and 721168651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 721268651Skris this will enable certificates using GeneralizedTime in validity 721368651Skris dates to be checked. 721468651Skris [Steve Henson] 721568651Skris 721668651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 721768651Skris negative public key encodings) on by default, 721868651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 721968651Skris [Steve Henson] 722068651Skris 722168651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 722268651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 722368651Skris the encoding can be trivially obtained from the structure. 722468651Skris [Steve Henson] 722568651Skris 722668651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 722768651Skris not read locks (CRYPTO_r_[un]lock). 722868651Skris [Bodo Moeller] 722968651Skris 723068651Skris *) A first attempt at creating official support for shared 723168651Skris libraries through configuration. I've kept it so the 723268651Skris default is static libraries only, and the OpenSSL programs 723368651Skris are always statically linked for now, but there are 723468651Skris preparations for dynamic linking in place. 723589837Skris This has been tested on Linux and Tru64. 723668651Skris [Richard Levitte] 723768651Skris 723868651Skris *) Randomness polling function for Win9x, as described in: 723968651Skris Peter Gutmann, Software Generation of Practically Strong 724068651Skris Random Numbers. 724168651Skris [Ulf M�ller] 724268651Skris 724368651Skris *) Fix so PRNG is seeded in req if using an already existing 724468651Skris DSA key. 724568651Skris [Steve Henson] 724668651Skris 724768651Skris *) New options to smime application. -inform and -outform 724868651Skris allow alternative formats for the S/MIME message including 724968651Skris PEM and DER. The -content option allows the content to be 725068651Skris specified separately. This should allow things like Netscape 725168651Skris form signing output easier to verify. 725268651Skris [Steve Henson] 725368651Skris 725468651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 725568651Skris [Steve Henson] 725668651Skris 725768651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 725868651Skris STRING types. These convert content octets to and from the 725968651Skris underlying type. The actual tag and length octets are 726068651Skris already assumed to have been read in and checked. These 726168651Skris are needed because all other string types have virtually 726268651Skris identical handling apart from the tag. By having versions 726368651Skris of the ASN1 functions that just operate on content octets 726468651Skris IMPLICIT tagging can be handled properly. It also allows 726568651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 726668651Skris and ASN1_INTEGER are identical apart from the tag. 726768651Skris [Steve Henson] 726868651Skris 726968651Skris *) Change the handling of OID objects as follows: 727068651Skris 727168651Skris - New object identifiers are inserted in objects.txt, following 727268651Skris the syntax given in objects.README. 727368651Skris - objects.pl is used to process obj_mac.num and create a new 727468651Skris obj_mac.h. 727568651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 727668651Skris obj_mac.h. 727768651Skris 727868651Skris This is currently kind of a hack, and the perl code in objects.pl 727968651Skris isn't very elegant, but it works as I intended. The simplest way 728068651Skris to check that it worked correctly is to look in obj_dat.h and 728168651Skris check the array nid_objs and make sure the objects haven't moved 728268651Skris around (this is important!). Additions are OK, as well as 728368651Skris consistent name changes. 728468651Skris [Richard Levitte] 728568651Skris 728668651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 728768651Skris [Bodo Moeller] 728868651Skris 728968651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 729068651Skris The given file adds to whatever has already been seeded into the 729168651Skris random pool through the RANDFILE configuration file option or 729268651Skris environment variable, or the default random state file. 729368651Skris [Richard Levitte] 729468651Skris 729568651Skris *) mkstack.pl now sorts each macro group into lexical order. 729668651Skris Previously the output order depended on the order the files 729768651Skris appeared in the directory, resulting in needless rewriting 729868651Skris of safestack.h . 729968651Skris [Steve Henson] 730068651Skris 730168651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 730268651Skris work arounds for the VC++ problem that it treats func() as 730368651Skris func(void). Also stripped out the parts of mkdef.pl that 730468651Skris added extra typesafe functions: these no longer exist. 730568651Skris [Steve Henson] 730668651Skris 730768651Skris *) Reorganisation of the stack code. The macros are now all 730868651Skris collected in safestack.h . Each macro is defined in terms of 730968651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 731068651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 731168651Skris this has the advantage of retaining type safety without the 731268651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 731368651Skris then the non typesafe macros are used instead. Also modified the 731468651Skris mkstack.pl script to handle the new form. Needs testing to see 731568651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 731668651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 731768651Skris and PKCS12_STACK_OF. 731868651Skris [Steve Henson] 731968651Skris 732068651Skris *) When some versions of IIS use the 'NET' form of private key the 732168651Skris key derivation algorithm is different. Normally MD5(password) is 732268651Skris used as a 128 bit RC4 key. In the modified case 732368651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 732468651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 732568651Skris as the old Netscape_RSA functions except they have an additional 732668651Skris 'sgckey' parameter which uses the modified algorithm. Also added 732768651Skris an -sgckey command line option to the rsa utility. Thanks to 732868651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 732968651Skris algorithm to openssl-dev. 733068651Skris [Steve Henson] 733168651Skris 733268651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 733368651Skris invalid expansion on some systems (SCO 5.0.5 for example). 733468651Skris Corrected to 'c.kname'. 733568651Skris [Phillip Porch <root@theporch.com>] 733668651Skris 733768651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 733868651Skris a STACK of email addresses from a certificate or request, these look 733968651Skris in the subject name and the subject alternative name extensions and 734068651Skris omit any duplicate addresses. 734168651Skris [Steve Henson] 734268651Skris 734368651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 734468651Skris This makes DSA verification about 2 % faster. 734568651Skris [Bodo Moeller] 734668651Skris 734768651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 734868651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 734968651Skris plus overhead for 1024 bit moduli). 735068651Skris This makes exponentiations about 0.5 % faster for 1024 bit 735168651Skris exponents (as measured by "openssl speed rsa2048"). 735268651Skris [Bodo Moeller] 735368651Skris 735468651Skris *) Rename memory handling macros to avoid conflicts with other 735568651Skris software: 735668651Skris Malloc => OPENSSL_malloc 735768651Skris Malloc_locked => OPENSSL_malloc_locked 735868651Skris Realloc => OPENSSL_realloc 735968651Skris Free => OPENSSL_free 736068651Skris [Richard Levitte] 736168651Skris 736268651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 736368651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 736468651Skris [Bodo Moeller] 736568651Skris 736668651Skris *) CygWin32 support. 736768651Skris [John Jarvie <jjarvie@newsguy.com>] 736868651Skris 736968651Skris *) The type-safe stack code has been rejigged. It is now only compiled 737068651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 737168651Skris by default all type-specific stack functions are "#define"d back to 737268651Skris standard stack functions. This results in more streamlined output 737368651Skris but retains the type-safety checking possibilities of the original 737468651Skris approach. 737568651Skris [Geoff Thorpe] 737668651Skris 737768651Skris *) The STACK code has been cleaned up, and certain type declarations 737868651Skris that didn't make a lot of sense have been brought in line. This has 737968651Skris also involved a cleanup of sorts in safestack.h to more correctly 738068651Skris map type-safe stack functions onto their plain stack counterparts. 738168651Skris This work has also resulted in a variety of "const"ifications of 738268651Skris lots of the code, especially "_cmp" operations which should normally 738368651Skris be prototyped with "const" parameters anyway. 738468651Skris [Geoff Thorpe] 738568651Skris 738668651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 738768651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 738868651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 738968651Skris where all of 'md' is used each time the PRNG is used, but 'state' 739068651Skris is used only indexed by a cyclic counter. As entropy may not be 739168651Skris well distributed from the beginning, 'md' is important as a 739268651Skris chaining variable. However, the output function chains only half 739368651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 739468651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 739568651Skris in all of 'state' being rewritten, with the new values depending 739668651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 739768651Skris [Bodo Moeller] 739868651Skris 739968651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 740068651Skris the handshake is continued after ssl_verify_cert_chain(); 740168651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 740268651Skris can lead to 'unexplainable' connection aborts later. 740368651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 740468651Skris 740568651Skris *) Major EVP API cipher revision. 740668651Skris Add hooks for extra EVP features. This allows various cipher 740768651Skris parameters to be set in the EVP interface. Support added for variable 740868651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 740968651Skris setting of RC2 and RC5 parameters. 741068651Skris 741168651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 741268651Skris ciphers. 741368651Skris 741468651Skris Remove lots of duplicated code from the EVP library. For example *every* 741568651Skris cipher init() function handles the 'iv' in the same way according to the 741668651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 741768651Skris for CFB and OFB modes they zero ctx->num. 741868651Skris 741968651Skris New functionality allows removal of S/MIME code RC2 hack. 742068651Skris 742168651Skris Most of the routines have the same form and so can be declared in terms 742268651Skris of macros. 742368651Skris 742468651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 742568651Skris all individual ciphers. If the cipher wants to handle IVs or keys 742668651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 742768651Skris flags. 742868651Skris 742968651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 743068651Skris value: although software versions of the algorithms cannot fail 743168651Skris any installed hardware versions can. 743268651Skris [Steve Henson] 743368651Skris 743468651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 743568651Skris this option is set, tolerate broken clients that send the negotiated 743668651Skris protocol version number instead of the requested protocol version 743768651Skris number. 743868651Skris [Bodo Moeller] 743968651Skris 744068651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 744168651Skris i.e. non-zero for export ciphersuites, zero otherwise. 744268651Skris Previous versions had this flag inverted, inconsistent with 744368651Skris rsa_tmp_cb (..._TMP_RSA_CB). 744468651Skris [Bodo Moeller; problem reported by Amit Chopra] 744568651Skris 744668651Skris *) Add missing DSA library text string. Work around for some IIS 744768651Skris key files with invalid SEQUENCE encoding. 744868651Skris [Steve Henson] 744968651Skris 745068651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 745168651Skris and so on that are implemented in OpenSSL. 745268651Skris [Richard Levitte] 745368651Skris 745468651Skris *) Enhance c_rehash script. Old version would mishandle certificates 745568651Skris with the same subject name hash and wouldn't handle CRLs at all. 745668651Skris Added -fingerprint option to crl utility, to support new c_rehash 745768651Skris features. 745868651Skris [Steve Henson] 745968651Skris 746068651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 746168651Skris [Ulf M�ller] 746268651Skris 746368651Skris *) Fix for SSL server purpose checking. Server checking was 746468651Skris rejecting certificates which had extended key usage present 746568651Skris but no ssl client purpose. 746668651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 746768651Skris 746868651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 746968651Skris is a little unclear about how a blank password is handled. 747068651Skris Since the password in encoded as a BMPString with terminating 747168651Skris double NULL a zero length password would end up as just the 747268651Skris double NULL. However no password at all is different and is 747368651Skris handled differently in the PKCS#12 key generation code. NS 747468651Skris treats a blank password as zero length. MSIE treats it as no 747568651Skris password on export: but it will try both on import. We now do 747668651Skris the same: PKCS12_parse() tries zero length and no password if 747768651Skris the password is set to "" or NULL (NULL is now a valid password: 747868651Skris it wasn't before) as does the pkcs12 application. 747968651Skris [Steve Henson] 748068651Skris 748168651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 748268651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 748368651Skris be obtained from the error queue. 748468651Skris [Bodo Moeller] 748568651Skris 748668651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 748768651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 748868651Skris accordingly to avoid race conditions (this is necessary because 748968651Skris thread_hash is no longer constant once set). 749068651Skris [Bodo Moeller] 749168651Skris 749268651Skris *) Bugfix for linux-elf makefile.one. 749368651Skris [Ulf M�ller] 749468651Skris 749568651Skris *) RSA_get_default_method() will now cause a default 749668651Skris RSA_METHOD to be chosen if one doesn't exist already. 749768651Skris Previously this was only set during a call to RSA_new() 749868651Skris or RSA_new_method(NULL) meaning it was possible for 749968651Skris RSA_get_default_method() to return NULL. 750068651Skris [Geoff Thorpe] 750168651Skris 750268651Skris *) Added native name translation to the existing DSO code 750368651Skris that will convert (if the flag to do so is set) filenames 750468651Skris that are sufficiently small and have no path information 750568651Skris into a canonical native form. Eg. "blah" converted to 750668651Skris "libblah.so" or "blah.dll" etc. 750768651Skris [Geoff Thorpe] 750868651Skris 750968651Skris *) New function ERR_error_string_n(e, buf, len) which is like 751068651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 751168651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 751268651Skris may not be NULL. 751368651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 751468651Skris 751568651Skris *) CONF library reworked to become more general. A new CONF 751668651Skris configuration file reader "class" is implemented as well as a 751768651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 751868651Skris old CONF_* functions are still there, but are reimplemented to 751968651Skris work in terms of the new functions. Also, a set of functions 752068651Skris to handle the internal storage of the configuration data is 752168651Skris provided to make it easier to write new configuration file 752268651Skris reader "classes" (I can definitely see something reading a 752368651Skris configuration file in XML format, for example), called _CONF_*, 752468651Skris or "the configuration storage API"... 752568651Skris 752668651Skris The new configuration file reading functions are: 752768651Skris 752868651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 752968651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 753068651Skris 753168651Skris NCONF_default, NCONF_WIN32 753268651Skris 753368651Skris NCONF_dump_fp, NCONF_dump_bio 753468651Skris 753568651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 753668651Skris NCONF_new creates a new CONF object. This works in the same way 753768651Skris as other interfaces in OpenSSL, like the BIO interface. 753868651Skris NCONF_dump_* dump the internal storage of the configuration file, 753968651Skris which is useful for debugging. All other functions take the same 754068651Skris arguments as the old CONF_* functions wth the exception of the 754168651Skris first that must be a `CONF *' instead of a `LHASH *'. 754268651Skris 754368651Skris To make it easer to use the new classes with the old CONF_* functions, 754468651Skris the function CONF_set_default_method is provided. 754568651Skris [Richard Levitte] 754668651Skris 754768651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 754868651Skris mentioned in the documentation but had not been implemented. 754968651Skris (This option is not yet really useful because even the additional 755068651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 755168651Skris [Bodo Moeller] 755268651Skris 755368651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 755468651Skris OpenSSL-based applications) load shared libraries and bind to 755568651Skris them in a portable way. 755668651Skris [Geoff Thorpe, with contributions from Richard Levitte] 755768651Skris 755859191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 755959191Skris 756059191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 756159191Skris 756259191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 756359191Skris (the default implementation of RAND_status). 756459191Skris 756559191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 756659191Skris to '-clrext' (= clear extensions), as intended and documented. 756759191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 756859191Skris <attili@amaxo.com>] 756959191Skris 757059191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 757159191Skris was larger than the MD block size. 757259191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 757359191Skris 757459191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 757559191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 757659191Skris using the passed key: if the passed key was a private key the result 757759191Skris of X509_print(), for example, would be to print out all the private key 757859191Skris components. 757959191Skris [Steve Henson] 758059191Skris 758159191Skris *) des_quad_cksum() byte order bug fix. 758259191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 758359191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 758459191Skris 758559191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 758659191Skris discouraged. 758759191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 758859191Skris 758959191Skris *) For easily testing in shell scripts whether some command 759059191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 759159191Skris returns with exit code 0 iff no command of the given name is available. 759259191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 759359191Skris the output goes to stdout and nothing is printed to stderr. 759459191Skris Additional arguments are always ignored. 759559191Skris 759659191Skris Since for each cipher there is a command of the same name, 759759191Skris the 'no-cipher' compilation switches can be tested this way. 759859191Skris 759959191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 760059191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 760159191Skris [Bodo Moeller] 760259191Skris 760359191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 760459191Skris [Bodo Moeller] 760559191Skris 760659191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 760759191Skris is set; it will be thrown away anyway because each handshake creates 760859191Skris its own key. 760959191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 761059191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 761159191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 761259191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 761359191Skris [Bodo Moeller] 761459191Skris 761559191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 761659191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 761759191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 761859191Skris does not suppress any output. 761959191Skris [Richard Levitte] 762059191Skris 762159191Skris *) Add compatibility options to the purpose and trust code. The 762259191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 762359191Skris accepts a certificate or CA, this was the previous behaviour, 762459191Skris with all the associated security issues. 762559191Skris 762659191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 762759191Skris automatically trust self signed roots in certificate store. A 762859191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 762959191Skris a purpose has no associated trust setting and it should instead 763059191Skris use the value in the default purpose. 763159191Skris [Steve Henson] 763259191Skris 763359191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 763459191Skris and fix a memory leak. 763559191Skris [Steve Henson] 763659191Skris 763759191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 763859191Skris reason strings from the previous version of the .c file, as 763959191Skris the default to have only downcase letters (and digits) in 764059191Skris automatically generated reasons codes is not always appropriate. 764159191Skris [Bodo Moeller] 764259191Skris 764359191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 764459191Skris using strerror. Previously, ERR_reason_error_string() returned 764559191Skris library names as reason strings for SYSerr; but SYSerr is a special 764659191Skris case where small numbers are errno values, not library numbers. 764759191Skris [Bodo Moeller] 764859191Skris 764959191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 765059191Skris converts DSA parameters into DH parameters. (When creating parameters, 765159191Skris DSA_generate_parameters is used.) 765259191Skris [Bodo Moeller] 765359191Skris 765459191Skris *) Include 'length' (recommended exponent length) in C code generated 765559191Skris by 'openssl dhparam -C'. 765659191Skris [Bodo Moeller] 765759191Skris 765859191Skris *) The second argument to set_label in perlasm was already being used 765959191Skris so couldn't be used as a "file scope" flag. Moved to third argument 766059191Skris which was free. 766159191Skris [Steve Henson] 766259191Skris 766359191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 766459191Skris instead of RAND_bytes for encryption IVs and salts. 766559191Skris [Bodo Moeller] 766659191Skris 766759191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 766859191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 766959191Skris RAND_set_rand_method would be impossible. 767059191Skris [Bodo Moeller] 767159191Skris 767259191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 767359191Skris number generation fails. 767459191Skris [Bodo Moeller] 767559191Skris 767659191Skris *) New 'rand' application for creating pseudo-random output. 767759191Skris [Bodo Moeller] 767859191Skris 767959191Skris *) Added configuration support for Linux/IA64 768059191Skris [Rolf Haberrecker <rolf@suse.de>] 768159191Skris 768259191Skris *) Assembler module support for Mingw32. 768359191Skris [Ulf M�ller] 768459191Skris 768559191Skris *) Shared library support for HPUX (in shlib/). 768659191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 768759191Skris 768859191Skris *) Shared library support for Solaris gcc. 768959191Skris [Lutz Behnke <behnke@trustcenter.de>] 769059191Skris 769159191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 769259191Skris 769359191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 769459191Skris were added manually and by SMIME_crlf_copy(). 769559191Skris [Steve Henson] 769659191Skris 769759191Skris *) In bntest.c don't call BN_rand with zero bits argument. 769859191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 769959191Skris 770059191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 770159191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 770259191Skris [Ulf M�ller] 770359191Skris 770459191Skris *) Add an optional second argument to the set_label() in the perl 770559191Skris assembly language builder. If this argument exists and is set 770659191Skris to 1 it signals that the assembler should use a symbol whose 770759191Skris scope is the entire file, not just the current function. This 770859191Skris is needed with MASM which uses the format label:: for this scope. 770959191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 771059191Skris 771159191Skris *) Change the ASN1 types so they are typedefs by default. Before 771259191Skris almost all types were #define'd to ASN1_STRING which was causing 771359191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 771459191Skris for example. 771559191Skris [Steve Henson] 771659191Skris 771759191Skris *) Change names of new functions to the new get1/get0 naming 771859191Skris convention: After 'get1', the caller owns a reference count 771959191Skris and has to call ..._free; 'get0' returns a pointer to some 772059191Skris data structure without incrementing reference counters. 772159191Skris (Some of the existing 'get' functions increment a reference 772259191Skris counter, some don't.) 772359191Skris Similarly, 'set1' and 'add1' functions increase reference 772459191Skris counters or duplicate objects. 772559191Skris [Steve Henson] 772659191Skris 772759191Skris *) Allow for the possibility of temp RSA key generation failure: 772859191Skris the code used to assume it always worked and crashed on failure. 772959191Skris [Steve Henson] 773059191Skris 773159191Skris *) Fix potential buffer overrun problem in BIO_printf(). 773259191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 773359191Skris pointed out by David Sacerdote <das33@cornell.edu>] 773459191Skris 773559191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 773659191Skris RAND_egd() and RAND_status(). In the command line application, 773759191Skris the EGD socket can be specified like a seed file using RANDFILE 773859191Skris or -rand. 773959191Skris [Ulf M�ller] 774059191Skris 774159191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 774259191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 774359191Skris [Steve Henson] 774459191Skris 774559191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 774659191Skris list to exclude them. This means that no special compilation option 774759191Skris is needed to use anonymous DH: it just needs to be included in the 774859191Skris cipher list. 774959191Skris [Steve Henson] 775059191Skris 775159191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 775259191Skris EVP_MD_type. The old functionality is available in a new macro called 775359191Skris EVP_MD_md(). Change code that uses it and update docs. 775459191Skris [Steve Henson] 775559191Skris 775659191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 775759191Skris where the 'void *' argument is replaced by a function pointer argument. 775859191Skris Previously 'void *' was abused to point to functions, which works on 775959191Skris many platforms, but is not correct. As these functions are usually 776059191Skris called by macros defined in OpenSSL header files, most source code 776159191Skris should work without changes. 776259191Skris [Richard Levitte] 776359191Skris 776459191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 776559191Skris sections with information on -D... compiler switches used for 776659191Skris compiling the library so that applications can see them. To enable 776759191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 776859191Skris must be defined. E.g., 776959191Skris #define OPENSSL_ALGORITHM_DEFINES 777059191Skris #include <openssl/opensslconf.h> 777159191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 777259191Skris [Richard Levitte, Ulf and Bodo M�ller] 777359191Skris 777459191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 777559191Skris record layer. 777659191Skris [Bodo Moeller] 777759191Skris 777859191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 777959191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 778059191Skris the required ASN1 format: arbitrary types determined by an OID. 778159191Skris [Steve Henson] 778259191Skris 778359191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 778459191Skris argument to 'req'. This is not because the function is newer or 778559191Skris better than others it just uses the work 'NEW' in the certificate 778659191Skris request header lines. Some software needs this. 778759191Skris [Steve Henson] 778859191Skris 778959191Skris *) Reorganise password command line arguments: now passwords can be 779059191Skris obtained from various sources. Delete the PEM_cb function and make 779159191Skris it the default behaviour: i.e. if the callback is NULL and the 779259191Skris usrdata argument is not NULL interpret it as a null terminated pass 779359191Skris phrase. If usrdata and the callback are NULL then the pass phrase 779459191Skris is prompted for as usual. 779559191Skris [Steve Henson] 779659191Skris 779759191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 779859191Skris the support is automatically enabled. The resulting binaries will 779959191Skris autodetect the card and use it if present. 780059191Skris [Ben Laurie and Compaq Inc.] 780159191Skris 780259191Skris *) Work around for Netscape hang bug. This sends certificate request 780359191Skris and server done in one record. Since this is perfectly legal in the 780459191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 780559191Skris the bugs/SSLv3 entry for more info. 780659191Skris [Steve Henson] 780759191Skris 780859191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 780959191Skris [Andy Polyakov] 781059191Skris 781159191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 781259191Skris of seed file. 781359191Skris [Steve Henson] 781459191Skris 781559191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 781659191Skris [Bodo Moeller] 781759191Skris 781859191Skris *) Add command line password options to the remaining applications. 781959191Skris [Steve Henson] 782059191Skris 782159191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 782259191Skris bits. 782359191Skris [Ulf M�ller] 782459191Skris 782559191Skris *) More tests in bntest.c, and changed test_bn output. 782659191Skris [Ulf M�ller] 782759191Skris 782859191Skris *) ./config recognizes MacOS X now. 782959191Skris [Andy Polyakov] 783059191Skris 783159191Skris *) Bug fix for BN_div() when the first words of num and divsor are 783259191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 783359191Skris [Ulf M�ller] 783459191Skris 783559191Skris *) Add support for various broken PKCS#8 formats, and command line 783659191Skris options to produce them. 783759191Skris [Steve Henson] 783859191Skris 783959191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 784059191Skris get temporary BIGNUMs from a BN_CTX. 784159191Skris [Ulf M�ller] 784259191Skris 784359191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 784459191Skris for p == 0. 784559191Skris [Ulf M�ller] 784659191Skris 784759191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 784859191Skris include a #define from the old name to the new. The original intent 784959191Skris was that statically linked binaries could for example just call 785059191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 785159191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 785259191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 785359191Skris one would link with the other. They are now in separate source files. 785459191Skris [Steve Henson] 785559191Skris 785659191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 785759191Skris [Steve Henson] 785859191Skris 785959191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 786059191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 786159191Skris loop, our standard modexp algorithms are faster). 786259191Skris [Bodo Moeller] 786359191Skris 786459191Skris *) Support for the EBCDIC character set completed. 786559191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 786659191Skris 786759191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 786859191Skris use void * instead of char * in lhash. 786959191Skris [Ulf M�ller] 787059191Skris 787159191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 787259191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 787359191Skris this the server could overwrite ephemeral keys that the client 787459191Skris has already seen). 787559191Skris [Bodo Moeller] 787659191Skris 787759191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 787859191Skris using 50 iterations of the Rabin-Miller test. 787959191Skris 788059191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 788159191Skris iterations of the Rabin-Miller test as required by the appendix 788259191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 788359191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 788459191Skris generation becomes much faster. 788559191Skris 788659191Skris This implies a change for the callback functions in DSA_is_prime 788759191Skris and DSA_generate_parameters: The callback function is called once 788859191Skris for each positive witness in the Rabin-Miller test, not just 788959191Skris occasionally in the inner loop; and the parameters to the 789059191Skris callback function now provide an iteration count for the outer 789159191Skris loop rather than for the current invocation of the inner loop. 789259191Skris DSA_generate_parameters additionally can call the callback 789359191Skris function with an 'iteration count' of -1, meaning that a 789459191Skris candidate has passed the trial division test (when q is generated 789559191Skris from an application-provided seed, trial division is skipped). 789659191Skris [Bodo Moeller] 789759191Skris 789859191Skris *) New function BN_is_prime_fasttest that optionally does trial 789959191Skris division before starting the Rabin-Miller test and has 790059191Skris an additional BN_CTX * argument (whereas BN_is_prime always 790159191Skris has to allocate at least one BN_CTX). 790259191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 790359191Skris trial division stage. 790459191Skris [Bodo Moeller] 790559191Skris 790659191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 790759191Skris as ASN1_TIME. 790859191Skris [Steve Henson] 790959191Skris 791059191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 791159191Skris [Steve Henson] 791259191Skris 791359191Skris *) New function BN_pseudo_rand(). 791459191Skris [Ulf M�ller] 791559191Skris 791659191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 791759191Skris bignum version of BN_from_montgomery() with the working code from 791859191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 791959191Skris the comments. 792059191Skris [Ulf M�ller] 792159191Skris 792259191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 792359191Skris made it impossible to use the same SSL_SESSION data structure in 792459191Skris SSL2 clients in multiple threads. 792559191Skris [Bodo Moeller] 792659191Skris 792759191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 792859191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 792959191Skris to seed the PRNG (previously an explicit byte count was required). 793059191Skris [Ulf M�ller, Bodo M�ller] 793159191Skris 793259191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 793359191Skris used (char *) instead of (void *) and had casts all over the place. 793459191Skris [Steve Henson] 793559191Skris 793659191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 793759191Skris [Ulf M�ller] 793859191Skris 793959191Skris *) Retain source code compatibility for BN_prime_checks macro: 794059191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 794159191Skris BN_prime_checks_for_size to determine the appropriate number of 794259191Skris Rabin-Miller iterations. 794359191Skris [Ulf M�ller] 794459191Skris 794559191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 794659191Skris DH_CHECK_P_NOT_SAFE_PRIME. 794759191Skris (Check if this is true? OpenPGP calls them "strong".) 794859191Skris [Ulf M�ller] 794959191Skris 795059191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 795159191Skris "dhparam". The old programs are retained for now but will handle DH keys 795259191Skris (instead of parameters) in future. 795359191Skris [Steve Henson] 795459191Skris 795559191Skris *) Make the ciphers, s_server and s_client programs check the return values 795659191Skris when a new cipher list is set. 795759191Skris [Steve Henson] 795859191Skris 795959191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 796059191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 796159191Skris wrong. 796259191Skris 796359191Skris The syntax for the cipher sorting has been extended to support sorting by 796459191Skris cipher-strength (using the strength_bits hard coded in the tables). 796559191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 796659191Skris 796759191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 796859191Skris string with an "undefined" symbol (neither command nor alphanumeric 796959191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 797059191Skris an error is flagged. 797159191Skris 797259191Skris Due to the strength-sorting extension, the code of the 797359191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 797459191Skris the readability was also increased :-) 797559191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 797659191Skris 797759191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 797859191Skris for the first serial number and places 2 in the serial number file. This 797959191Skris avoids problems when the root CA is created with serial number zero and 798059191Skris the first user certificate has the same issuer name and serial number 798159191Skris as the root CA. 798259191Skris [Steve Henson] 798359191Skris 798459191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 798559191Skris the new code. Add documentation for this stuff. 798659191Skris [Steve Henson] 798759191Skris 798859191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 798959191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 799059191Skris structures and behave in an analagous way to the X509v3 functions: 799159191Skris they shouldn't be called directly but wrapper functions should be used 799259191Skris instead. 799359191Skris 799459191Skris So we also now have some wrapper functions that call the X509at functions 799559191Skris when passed certificate requests. (TO DO: similar things can be done with 799659191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 799759191Skris things. Some of these need some d2i or i2d and print functionality 799859191Skris because they handle more complex structures.) 799959191Skris [Steve Henson] 800059191Skris 800159191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 800259191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 800359191Skris NO_RSA in ssl/s2*.c. 800459191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 800559191Skris 800659191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 800759191Skris has a return value which indicates the quality of the random data 800859191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 800959191Skris error queue. New function RAND_pseudo_bytes() generates output that is 801059191Skris guaranteed to be unique but not unpredictable. RAND_add is like 801159191Skris RAND_seed, but takes an extra argument for an entropy estimate 801259191Skris (RAND_seed always assumes full entropy). 801359191Skris [Ulf M�ller] 801459191Skris 801559191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 801659191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 801759191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 801859191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 801959191Skris false-positive rate of at most 2^-80 for random input. 802059191Skris [Bodo Moeller] 802159191Skris 802259191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 802359191Skris [Bodo Moeller] 802459191Skris 802559191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 802659191Skris in the 0.9.5 release), this returns the chain 802759191Skris from an X509_CTX structure with a dup of the stack and all 802859191Skris the X509 reference counts upped: so the stack will exist 802959191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 803059191Skris to use this. 803159191Skris 803259191Skris Also make SSL_SESSION_print() print out the verify return 803359191Skris code. 803459191Skris [Steve Henson] 803559191Skris 803659191Skris *) Add manpage for the pkcs12 command. Also change the default 803759191Skris behaviour so MAC iteration counts are used unless the new 803859191Skris -nomaciter option is used. This improves file security and 803959191Skris only older versions of MSIE (4.0 for example) need it. 804059191Skris [Steve Henson] 804159191Skris 804259191Skris *) Honor the no-xxx Configure options when creating .DEF files. 804359191Skris [Ulf M�ller] 804459191Skris 804559191Skris *) Add PKCS#10 attributes to field table: challengePassword, 804659191Skris unstructuredName and unstructuredAddress. These are taken from 804759191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 804859191Skris international characters are used. 804959191Skris 805059191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 805159191Skris based on strings. Remove the 'loc' parameter when adding 805259191Skris attributes because these will be a SET OF encoding which is sorted 805359191Skris in ASN1 order. 805459191Skris [Steve Henson] 805559191Skris 805659191Skris *) Initial changes to the 'req' utility to allow request generation 805759191Skris automation. This will allow an application to just generate a template 805859191Skris file containing all the field values and have req construct the 805959191Skris request. 806059191Skris 806159191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 806259191Skris used all over the place including certificate requests and PKCS#7 806359191Skris structures. They are currently handled manually where necessary with 806459191Skris some primitive wrappers for PKCS#7. The new functions behave in a 806559191Skris manner analogous to the X509 extension functions: they allow 806659191Skris attributes to be looked up by NID and added. 806759191Skris 806859191Skris Later something similar to the X509V3 code would be desirable to 806959191Skris automatically handle the encoding, decoding and printing of the 807059191Skris more complex types. The string types like challengePassword can 807159191Skris be handled by the string table functions. 807259191Skris 807359191Skris Also modified the multi byte string table handling. Now there is 807459191Skris a 'global mask' which masks out certain types. The table itself 807559191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 807659191Skris is useful when for example there is only one permissible type 807759191Skris (as in countryName) and using the mask might result in no valid 807859191Skris types at all. 807959191Skris [Steve Henson] 808059191Skris 808159191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 808259191Skris SSL_get_peer_finished to allow applications to obtain the latest 808359191Skris Finished messages sent to the peer or expected from the peer, 808459191Skris respectively. (SSL_get_peer_finished is usually the Finished message 808559191Skris actually received from the peer, otherwise the protocol will be aborted.) 808659191Skris 808759191Skris As the Finished message are message digests of the complete handshake 808859191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 808959191Skris be used for external authentication procedures when the authentication 809059191Skris provided by SSL/TLS is not desired or is not enough. 809159191Skris [Bodo Moeller] 809259191Skris 809359191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 809459191Skris the host supports BWX extension and if Compaq C is present on the 809559191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 809659191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 809759191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 809859191Skris SHA1. 809959191Skris [Andy Polyakov] 810059191Skris 810159191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 810259191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 810359191Skris weak crypto and after checking the certificate is SGC a second one 810459191Skris with strong crypto. MS SGC stops the first handshake after receiving 810559191Skris the server certificate message and sends a second client hello. Since 810659191Skris a server will typically do all the time consuming operations before 810759191Skris expecting any further messages from the client (server key exchange 810859191Skris is the most expensive) there is little difference between the two. 810959191Skris 811059191Skris To get OpenSSL to support MS SGC we have to permit a second client 811159191Skris hello message after we have sent server done. In addition we have to 811259191Skris reset the MAC if we do get this second client hello. 811359191Skris [Steve Henson] 811459191Skris 811559191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 811659191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 811759191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 811859191Skris format DER encoded private key. Newer code should use PKCS#8 format which 811959191Skris has the key type encoded in the ASN1 structure. Added DER private key 812059191Skris support to pkcs8 application. 812159191Skris [Steve Henson] 812259191Skris 812359191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 812459191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 812559191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 812659191Skris is set, we interpret this as a request to violate the specification 812759191Skris (the worst that can happen is a handshake failure, and 'correct' 812859191Skris behaviour would result in a handshake failure anyway). 812959191Skris [Bodo Moeller] 813059191Skris 813159191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 813259191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 813359191Skris concurrently obtain them from an external cache). 813459191Skris The internal cache can handle only one SSL_SESSION with a given ID, 813559191Skris so if there's a conflict, we now throw out the old one to achieve 813659191Skris consistency. 813759191Skris [Bodo Moeller] 813859191Skris 813959191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 814059191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 814159191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 814259191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 814359191Skris example. 814459191Skris [Steve Henson] 814559191Skris 814659191Skris *) Simplify the trust setting structure and code. Now we just have 814759191Skris two sequences of OIDs for trusted and rejected settings. These will 814859191Skris typically have values the same as the extended key usage extension 814959191Skris and any application specific purposes. 815059191Skris 815159191Skris The trust checking code now has a default behaviour: it will just 815259191Skris check for an object with the same NID as the passed id. Functions can 815359191Skris be provided to override either the default behaviour or the behaviour 815459191Skris for a given id. SSL client, server and email already have functions 815559191Skris in place for compatibility: they check the NID and also return "trusted" 815659191Skris if the certificate is self signed. 815759191Skris [Steve Henson] 815859191Skris 815959191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 816059191Skris traditional format into an EVP_PKEY structure. 816159191Skris [Steve Henson] 816259191Skris 816359191Skris *) Add a password callback function PEM_cb() which either prompts for 816459191Skris a password if usr_data is NULL or otherwise assumes it is a null 816559191Skris terminated password. Allow passwords to be passed on command line 816659191Skris environment or config files in a few more utilities. 816759191Skris [Steve Henson] 816859191Skris 816959191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 817059191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 817159191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 817259191Skris Update documentation. 817359191Skris [Steve Henson] 817459191Skris 817559191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 817659191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 817759191Skris and produce an error if it couldn't. For compatibility we also have 817859191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 817959191Skris don't allocate anything because they don't need to. 818059191Skris [Steve Henson] 818159191Skris 818259191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 818359191Skris for details. 818459191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 818559191Skris 818659191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 818759191Skris possibly others as well. The purpose is to make an interface that 818859191Skris provide hooks so anyone can build a separate set of allocation and 818959191Skris deallocation routines to be used by OpenSSL, for example memory 819059191Skris pool implementations, or something else, which was previously hard 819159191Skris since Malloc(), Realloc() and Free() were defined as macros having 819259191Skris the values malloc, realloc and free, respectively (except for Win32 819359191Skris compilations). The same is provided for memory debugging code. 819459191Skris OpenSSL already comes with functionality to find memory leaks, but 819559191Skris this gives people a chance to debug other memory problems. 819659191Skris 819759191Skris With these changes, a new set of functions and macros have appeared: 819859191Skris 819968651Skris CRYPTO_set_mem_debug_functions() [F] 820068651Skris CRYPTO_get_mem_debug_functions() [F] 820168651Skris CRYPTO_dbg_set_options() [F] 820268651Skris CRYPTO_dbg_get_options() [F] 820368651Skris CRYPTO_malloc_debug_init() [M] 820459191Skris 820559191Skris The memory debug functions are NULL by default, unless the library 820659191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 820759191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 820859191Skris gives the standard debugging functions that come with OpenSSL) or 820959191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 821059191Skris provided by the library user) must be used. When the standard 821159191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 821259191Skris request additional information: 821359191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 821459191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 821559191Skris 821659191Skris Also, things like CRYPTO_set_mem_functions will always give the 821759191Skris expected result (the new set of functions is used for allocation 821859191Skris and deallocation) at all times, regardless of platform and compiler 821959191Skris options. 822059191Skris 822159191Skris To finish it up, some functions that were never use in any other 822259191Skris way than through macros have a new API and new semantic: 822359191Skris 822459191Skris CRYPTO_dbg_malloc() 822559191Skris CRYPTO_dbg_realloc() 822659191Skris CRYPTO_dbg_free() 822759191Skris 822859191Skris All macros of value have retained their old syntax. 822959191Skris [Richard Levitte and Bodo Moeller] 823059191Skris 823159191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 823259191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 823359191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 823459191Skris algorithm. 823559191Skris [Steve Henson] 823659191Skris 823759191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 823859191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 823959191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 824059191Skris 824159191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 824259191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 824359191Skris functionality to handle multipart/signed properly) and a utility 824459191Skris called 'smime' to call all this stuff. This is based on code I 824559191Skris originally wrote for Celo who have kindly allowed it to be 824659191Skris included in OpenSSL. 824759191Skris [Steve Henson] 824859191Skris 824959191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 825059191Skris des_set_key (aka des_key_sched). Global variable des_check_key 825159191Skris decides which of these is called by des_set_key; this way 825259191Skris des_check_key behaves as it always did, but applications and 825359191Skris the library itself, which was buggy for des_check_key == 1, 825459191Skris have a cleaner way to pick the version they need. 825559191Skris [Bodo Moeller] 825659191Skris 825759191Skris *) New function PKCS12_newpass() which changes the password of a 825859191Skris PKCS12 structure. 825959191Skris [Steve Henson] 826059191Skris 826159191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 826259191Skris dynamic mix. In both cases the ids can be used as an index into the 826359191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 826459191Skris functions so they accept a list of the field values and the 826559191Skris application doesn't need to directly manipulate the X509_TRUST 826659191Skris structure. 826759191Skris [Steve Henson] 826859191Skris 826959191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 827059191Skris need initialising. 827159191Skris [Steve Henson] 827259191Skris 827359191Skris *) Modify the way the V3 extension code looks up extensions. This now 827459191Skris works in a similar way to the object code: we have some "standard" 827559191Skris extensions in a static table which is searched with OBJ_bsearch() 827659191Skris and the application can add dynamic ones if needed. The file 827759191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 827859191Skris updated whenever a new extension is added to the core code and kept 827959191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 828059191Skris this. New extensions are not added too often so this file can readily 828159191Skris be maintained manually. 828259191Skris 828359191Skris There are two big advantages in doing things this way. The extensions 828459191Skris can be looked up immediately and no longer need to be "added" using 828559191Skris X509V3_add_standard_extensions(): this function now does nothing. 828659191Skris [Side note: I get *lots* of email saying the extension code doesn't 828759191Skris work because people forget to call this function] 828859191Skris Also no dynamic allocation is done unless new extensions are added: 828959191Skris so if we don't add custom extensions there is no need to call 829059191Skris X509V3_EXT_cleanup(). 829159191Skris [Steve Henson] 829259191Skris 829359191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 829459191Skris magic header, so unsalted files fail gracefully instead of just decrypting 829559191Skris to garbage. This is because not salting is a big security hole, so people 829659191Skris should be discouraged from doing it. 829759191Skris [Ben Laurie] 829859191Skris 829959191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 830059191Skris digest to be passed on the command line but it only used this 830159191Skris parameter when signing a certificate. Modified so all relevant 830259191Skris operations are affected by the digest parameter including the 830359191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 830459191Skris DSA key was used because it didn't fix the digest. 830559191Skris [Steve Henson] 830659191Skris 830759191Skris *) Initial certificate chain verify code. Currently tests the untrusted 830859191Skris certificates for consistency with the verify purpose (which is set 830959191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 831059191Skris 831159191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 831259191Skris this is because it will reject chains with invalid extensions whereas 831359191Skris every previous version of OpenSSL and SSLeay made no checks at all. 831459191Skris 831559191Skris Trust code: checks the root CA for the relevant trust settings. Trust 831659191Skris settings have an initial value consistent with the verify purpose: e.g. 831759191Skris if the verify purpose is for SSL client use it expects the CA to be 831859191Skris trusted for SSL client use. However the default value can be changed to 831959191Skris permit custom trust settings: one example of this would be to only trust 832059191Skris certificates from a specific "secure" set of CAs. 832159191Skris 832259191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 832359191Skris which should be used for version portability: especially since the 832459191Skris verify structure is likely to change more often now. 832559191Skris 832659191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 832759191Skris to set them. If not set then assume SSL clients will verify SSL servers 832859191Skris and vice versa. 832959191Skris 833059191Skris Two new options to the verify program: -untrusted allows a set of 833159191Skris untrusted certificates to be passed in and -purpose which sets the 833259191Skris intended purpose of the certificate. If a purpose is set then the 833359191Skris new chain verify code is used to check extension consistency. 833459191Skris [Steve Henson] 833559191Skris 833659191Skris *) Support for the authority information access extension. 833759191Skris [Steve Henson] 833859191Skris 833959191Skris *) Modify RSA and DSA PEM read routines to transparently handle 834059191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 834159191Skris public keys in a format compatible with certificate 834259191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 834359191Skris functions called *_PublicKey_* which used various odd formats so 834459191Skris these are retained for compatibility: however the DSA variants were 834559191Skris never in a public release so they have been deleted. Changed dsa/rsa 834659191Skris utilities to handle the new format: note no releases ever handled public 834759191Skris keys so we should be OK. 834859191Skris 834959191Skris The primary motivation for this change is to avoid the same fiasco 835059191Skris that dogs private keys: there are several incompatible private key 835159191Skris formats some of which are standard and some OpenSSL specific and 835259191Skris require various evil hacks to allow partial transparent handling and 835359191Skris even then it doesn't work with DER formats. Given the option anything 835459191Skris other than PKCS#8 should be dumped: but the other formats have to 835559191Skris stay in the name of compatibility. 835659191Skris 835759191Skris With public keys and the benefit of hindsight one standard format 835859191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 835959191Skris it clearly returns an error if you try to read the wrong kind of key. 836059191Skris 836159191Skris Added a -pubkey option to the 'x509' utility to output the public key. 836259191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 836359191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 836459191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 836559191Skris that do the same as the EVP_PKEY_assign_*() except they up the 836659191Skris reference count of the added key (they don't "swallow" the 836759191Skris supplied key). 836859191Skris [Steve Henson] 836959191Skris 837059191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 837159191Skris CRLs would fail if the file contained no certificates or no CRLs: 837259191Skris added a new function to read in both types and return the number 837359191Skris read: this means that if none are read it will be an error. The 837459191Skris DER versions of the certificate and CRL reader would always fail 837559191Skris because it isn't possible to mix certificates and CRLs in DER format 837659191Skris without choking one or the other routine. Changed this to just read 837759191Skris a certificate: this is the best we can do. Also modified the code 837859191Skris in apps/verify.c to take notice of return codes: it was previously 837959191Skris attempting to read in certificates from NULL pointers and ignoring 838059191Skris any errors: this is one reason why the cert and CRL reader seemed 838159191Skris to work. It doesn't check return codes from the default certificate 838259191Skris routines: these may well fail if the certificates aren't installed. 838359191Skris [Steve Henson] 838459191Skris 838559191Skris *) Code to support otherName option in GeneralName. 838659191Skris [Steve Henson] 838759191Skris 838859191Skris *) First update to verify code. Change the verify utility 838959191Skris so it warns if it is passed a self signed certificate: 839059191Skris for consistency with the normal behaviour. X509_verify 839159191Skris has been modified to it will now verify a self signed 839259191Skris certificate if *exactly* the same certificate appears 839359191Skris in the store: it was previously impossible to trust a 839459191Skris single self signed certificate. This means that: 839559191Skris openssl verify ss.pem 839659191Skris now gives a warning about a self signed certificate but 839759191Skris openssl verify -CAfile ss.pem ss.pem 839859191Skris is OK. 839959191Skris [Steve Henson] 840059191Skris 840159191Skris *) For servers, store verify_result in SSL_SESSION data structure 840259191Skris (and add it to external session representation). 840359191Skris This is needed when client certificate verifications fails, 840459191Skris but an application-provided verification callback (set by 840559191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 840659191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 840759191Skris but returns 1): When the session is reused, we have to set 840859191Skris ssl->verify_result to the appropriate error code to avoid 840959191Skris security holes. 841059191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 841159191Skris 841259191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 841359191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 841459191Skris didn't contain any existing data because it was being created. 841559191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 841659191Skris 841759191Skris *) Add a salt to the key derivation routines in enc.c. This 841859191Skris forms the first 8 bytes of the encrypted file. Also add a 841959191Skris -S option to allow a salt to be input on the command line. 842059191Skris [Steve Henson] 842159191Skris 842259191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 842359191Skris to compare two certificates. We do this by working out the SHA1 842459191Skris hash and comparing that. X509_cmp() will be needed by the trust 842559191Skris code. 842659191Skris [Steve Henson] 842759191Skris 842859191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 842959191Skris the reference count in the SSL_SESSION returned. 843059191Skris [Geoff Thorpe <geoff@eu.c2.net>] 843159191Skris 843259191Skris *) Fix for 'req': it was adding a null to request attributes. 843359191Skris Also change the X509_LOOKUP and X509_INFO code to handle 843459191Skris certificate auxiliary information. 843559191Skris [Steve Henson] 843659191Skris 843759191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 843859191Skris the 'enc' command. 843959191Skris [Steve Henson] 844059191Skris 844159191Skris *) Add the possibility to add extra information to the memory leak 844259191Skris detecting output, to form tracebacks, showing from where each 844359191Skris allocation was originated: CRYPTO_push_info("constant string") adds 844459191Skris the string plus current file name and line number to a per-thread 844559191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 844659191Skris is like calling CYRPTO_pop_info() until the stack is empty. 844759191Skris Also updated memory leak detection code to be multi-thread-safe. 844859191Skris [Richard Levitte] 844959191Skris 845059191Skris *) Add options -text and -noout to pkcs7 utility and delete the 845159191Skris encryption options which never did anything. Update docs. 845259191Skris [Steve Henson] 845359191Skris 845459191Skris *) Add options to some of the utilities to allow the pass phrase 845559191Skris to be included on either the command line (not recommended on 845659191Skris OSes like Unix) or read from the environment. Update the 845759191Skris manpages and fix a few bugs. 845859191Skris [Steve Henson] 845959191Skris 846059191Skris *) Add a few manpages for some of the openssl commands. 846159191Skris [Steve Henson] 846259191Skris 846359191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 846459191Skris leaking and not finding already revoked certificates. 846559191Skris [Steve Henson] 846659191Skris 846759191Skris *) Extensive changes to support certificate auxiliary information. 846859191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 846959191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 847059191Skris can still read in a certificate file in the usual way but it 847159191Skris will also read in any additional "auxiliary information". By 847259191Skris doing things this way a fair degree of compatibility can be 847359191Skris retained: existing certificates can have this information added 847459191Skris using the new 'x509' options. 847559191Skris 847659191Skris Current auxiliary information includes an "alias" and some trust 847759191Skris settings. The trust settings will ultimately be used in enhanced 847859191Skris certificate chain verification routines: currently a certificate 847959191Skris can only be trusted if it is self signed and then it is trusted 848059191Skris for all purposes. 848159191Skris [Steve Henson] 848259191Skris 848359191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 848459191Skris The problem was that one of the replacement routines had not been working 848559191Skris since SSLeay releases. For now the offending routine has been replaced 848659191Skris with non-optimised assembler. Even so, this now gives around 95% 848759191Skris performance improvement for 1024 bit RSA signs. 848859191Skris [Mark Cox] 848959191Skris 849059191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 849159191Skris handling. Most clients have the effective key size in bits equal to 849259191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 849359191Skris A few however don't do this and instead use the size of the decrypted key 849459191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 849559191Skris the effective key length. In this case the effective key length can still 849659191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 849759191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 849859191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 849959191Skris the key length and effective key length are equal. 850059191Skris [Steve Henson] 850159191Skris 850259191Skris *) Add a bunch of functions that should simplify the creation of 850359191Skris X509_NAME structures. Now you should be able to do: 850459191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 850559191Skris and have it automatically work out the correct field type and fill in 850659191Skris the structures. The more adventurous can try: 850759191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 850859191Skris and it will (hopefully) work out the correct multibyte encoding. 850959191Skris [Steve Henson] 851059191Skris 851159191Skris *) Change the 'req' utility to use the new field handling and multibyte 851259191Skris copy routines. Before the DN field creation was handled in an ad hoc 851359191Skris way in req, ca, and x509 which was rather broken and didn't support 851459191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 851559191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 851659191Skris using the dirstring_type option. See the new comment in the default 851759191Skris openssl.cnf for more info. 851859191Skris [Steve Henson] 851959191Skris 852059191Skris *) Make crypto/rand/md_rand.c more robust: 852159191Skris - Assure unique random numbers after fork(). 852259191Skris - Make sure that concurrent threads access the global counter and 852359191Skris md serializably so that we never lose entropy in them 852459191Skris or use exactly the same state in multiple threads. 852559191Skris Access to the large state is not always serializable because 852659191Skris the additional locking could be a performance killer, and 852759191Skris md should be large enough anyway. 852859191Skris [Bodo Moeller] 852959191Skris 853059191Skris *) New file apps/app_rand.c with commonly needed functionality 853159191Skris for handling the random seed file. 853259191Skris 853359191Skris Use the random seed file in some applications that previously did not: 853459191Skris ca, 853559191Skris dsaparam -genkey (which also ignored its '-rand' option), 853659191Skris s_client, 853759191Skris s_server, 853859191Skris x509 (when signing). 853959191Skris Except on systems with /dev/urandom, it is crucial to have a random 854059191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 854159191Skris for RSA signatures we could do without one. 854259191Skris 854359191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 854459191Skris of each file listed in the '-rand' option. The function as previously 854559191Skris found in genrsa is now in app_rand.c and is used by all programs 854659191Skris that support '-rand'. 854759191Skris [Bodo Moeller] 854859191Skris 854959191Skris *) In RAND_write_file, use mode 0600 for creating files; 855059191Skris don't just chmod when it may be too late. 855159191Skris [Bodo Moeller] 855259191Skris 855359191Skris *) Report an error from X509_STORE_load_locations 855459191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 855559191Skris [Bill Perry] 855659191Skris 855759191Skris *) New function ASN1_mbstring_copy() this copies a string in either 855859191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 855959191Skris into an ASN1_STRING type. A mask of permissible types is passed 856059191Skris and it chooses the "minimal" type to use or an error if not type 856159191Skris is suitable. 856259191Skris [Steve Henson] 856359191Skris 856459191Skris *) Add function equivalents to the various macros in asn1.h. The old 856559191Skris macros are retained with an M_ prefix. Code inside the library can 856659191Skris use the M_ macros. External code (including the openssl utility) 856759191Skris should *NOT* in order to be "shared library friendly". 856859191Skris [Steve Henson] 856959191Skris 857059191Skris *) Add various functions that can check a certificate's extensions 857159191Skris to see if it usable for various purposes such as SSL client, 857259191Skris server or S/MIME and CAs of these types. This is currently 857359191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 857459191Skris verification. Also added a -purpose flag to x509 utility to 857559191Skris print out all the purposes. 857659191Skris [Steve Henson] 857759191Skris 857859191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 857959191Skris functions. 858059191Skris [Steve Henson] 858159191Skris 858259191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 858359191Skris for, obtain and decode and extension and obtain its critical flag. 858459191Skris This allows all the necessary extension code to be handled in a 858559191Skris single function call. 858659191Skris [Steve Henson] 858759191Skris 858859191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 858959191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 859059191Skris [Andy Polyakov] 859159191Skris 859259191Skris *) New -noout option to asn1parse. This causes no output to be produced 859359191Skris its main use is when combined with -strparse and -out to extract data 859459191Skris from a file (which may not be in ASN.1 format). 859559191Skris [Steve Henson] 859659191Skris 859759191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 859859191Skris when producing the local key id. 859959191Skris [Richard Levitte <levitte@stacken.kth.se>] 860059191Skris 860159191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 860259191Skris stated explicitly. If it is not stated then it tries the first server 860359191Skris certificate file. The previous behaviour hard coded the filename 860459191Skris "server.pem". 860559191Skris [Steve Henson] 860659191Skris 860759191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 860859191Skris a public key to be input or output. For example: 860959191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 861059191Skris Also added necessary DSA public key functions to handle this. 861159191Skris [Steve Henson] 861259191Skris 861359191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 861459191Skris in the message. This was handled by allowing 861559191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 861659191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 861759191Skris 861859191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 861959191Skris to the end of the strings whereas this didn't. This would cause problems 862059191Skris if strings read with d2i_ASN1_bytes() were later modified. 862159191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 862259191Skris 862359191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 862459191Skris data and it contains EOF it will end up returning an error. This is 862559191Skris caused by input 46 bytes long. The cause is due to the way base64 862659191Skris BIOs find the start of base64 encoded data. They do this by trying a 862759191Skris trial decode on each line until they find one that works. When they 862859191Skris do a flag is set and it starts again knowing it can pass all the 862959191Skris data directly through the decoder. Unfortunately it doesn't reset 863059191Skris the context it uses. This means that if EOF is reached an attempt 863159191Skris is made to pass two EOFs through the context and this causes the 863259191Skris resulting error. This can also cause other problems as well. As is 863359191Skris usual with these problems it takes *ages* to find and the fix is 863459191Skris trivial: move one line. 863559191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 863659191Skris 863759191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 863859191Skris old code wouldn't work because it needed to select() on sockets and the 863959191Skris tty (for keypresses and to see if data could be written). Win32 only 864059191Skris supports select() on sockets so we select() with a 1s timeout on the 864159191Skris sockets and then see if any characters are waiting to be read, if none 864259191Skris are present then we retry, we also assume we can always write data to 864359191Skris the tty. This isn't nice because the code then blocks until we've 864459191Skris received a complete line of data and it is effectively polling the 864559191Skris keyboard at 1s intervals: however it's quite a bit better than not 864659191Skris working at all :-) A dedicated Windows application might handle this 864759191Skris with an event loop for example. 864859191Skris [Steve Henson] 864959191Skris 865059191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 865159191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 865259191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 865359191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 865459191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 865559191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 865659191Skris This necessitated the support of an extra signature type NID_md5_sha1 865759191Skris for SSL signatures and modifications to the SSL library to use it instead 865859191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 865959191Skris [Steve Henson] 866059191Skris 866159191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 866259191Skris will lookup a CRL issuers certificate and verify the signature in a 866359191Skris similar way to the verify program. Tidy up the crl program so it 866459191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 866559191Skris less strict. It will now permit CRL extensions even if it is not 866659191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 866759191Skris [Steve Henson] 866859191Skris 866959191Skris *) Initialize all non-automatic variables each time one of the openssl 867059191Skris sub-programs is started (this is necessary as they may be started 867159191Skris multiple times from the "OpenSSL>" prompt). 867259191Skris [Lennart Bang, Bodo Moeller] 867359191Skris 867459191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 867559191Skris removing all other RSA functionality (this is what NO_RSA does). This 867659191Skris is so (for example) those in the US can disable those operations covered 867759191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 867859191Skris key generation. 867959191Skris [Steve Henson] 868059191Skris 868159191Skris *) Non-copying interface to BIO pairs. 868259191Skris (still largely untested) 868359191Skris [Bodo Moeller] 868459191Skris 868559191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 868659191Skris ASCII string. This was handled independently in various places before. 868759191Skris [Steve Henson] 868859191Skris 868959191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 869059191Skris UTF8 strings a character at a time. 869159191Skris [Steve Henson] 869259191Skris 869359191Skris *) Use client_version from client hello to select the protocol 869459191Skris (s23_srvr.c) and for RSA client key exchange verification 869559191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 869659191Skris [Bodo Moeller] 869759191Skris 869859191Skris *) Add various utility functions to handle SPKACs, these were previously 869959191Skris handled by poking round in the structure internals. Added new function 870059191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 870159191Skris print, verify and generate SPKACs. Based on an original idea from 870259191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 870359191Skris [Steve Henson] 870459191Skris 870559191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 870659191Skris [Andy Polyakov] 870759191Skris 870859191Skris *) Allow the config file extension section to be overwritten on the 870959191Skris command line. Based on an original idea from Massimiliano Pala 871059191Skris <madwolf@comune.modena.it>. The new option is called -extensions 871159191Skris and can be applied to ca, req and x509. Also -reqexts to override 871259191Skris the request extensions in req and -crlexts to override the crl extensions 871359191Skris in ca. 871459191Skris [Steve Henson] 871559191Skris 871659191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 871759191Skris the same field multiple times by preceding it by "XXXX." for example: 871859191Skris 1.OU="Unit name 1" 871959191Skris 2.OU="Unit name 2" 872059191Skris this is the same syntax as used in the req config file. 872159191Skris [Steve Henson] 872259191Skris 872359191Skris *) Allow certificate extensions to be added to certificate requests. These 872459191Skris are specified in a 'req_extensions' option of the req section of the 872559191Skris config file. They can be printed out with the -text option to req but 872659191Skris are otherwise ignored at present. 872759191Skris [Steve Henson] 872859191Skris 872959191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 873059191Skris data read consists of only the final block it would not decrypted because 873159191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 873259191Skris A misplaced 'break' also meant the decrypted final block might not be 873359191Skris copied until the next read. 873459191Skris [Steve Henson] 873559191Skris 873659191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 873759191Skris a few extra parameters to the DH structure: these will be useful if 873859191Skris for example we want the value of 'q' or implement X9.42 DH. 873959191Skris [Steve Henson] 874059191Skris 874159191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 874259191Skris provides hooks that allow the default DSA functions or functions on a 874359191Skris "per key" basis to be replaced. This allows hardware acceleration and 874459191Skris hardware key storage to be handled without major modification to the 874559191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 874659191Skris associated functions. 874759191Skris [Steve Henson] 874859191Skris 874959191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 875059191Skris as "read only": it can't be written to and the buffer it points to will 875159191Skris not be freed. Reading from a read only BIO is much more efficient than 875259191Skris a normal memory BIO. This was added because there are several times when 875359191Skris an area of memory needs to be read from a BIO. The previous method was 875459191Skris to create a memory BIO and write the data to it, this results in two 875559191Skris copies of the data and an O(n^2) reading algorithm. There is a new 875659191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 875759191Skris an area of memory. Also modified the PKCS#7 routines to use read only 875859191Skris memory BIOs. 875959191Skris [Steve Henson] 876059191Skris 876159191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 876259191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 876359191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 876459191Skris but a retry condition occured while trying to read the rest. 876559191Skris [Bodo Moeller] 876659191Skris 876759191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 876859191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 876959191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 877059191Skris the encrypted data type: this is a more sensible place to put it and it 877159191Skris allows the PKCS#12 code to be tidied up that duplicated this 877259191Skris functionality. 877359191Skris [Steve Henson] 877459191Skris 877559191Skris *) Changed obj_dat.pl script so it takes its input and output files on 877659191Skris the command line. This should avoid shell escape redirection problems 877759191Skris under Win32. 877859191Skris [Steve Henson] 877959191Skris 878059191Skris *) Initial support for certificate extension requests, these are included 878159191Skris in things like Xenroll certificate requests. Included functions to allow 878259191Skris extensions to be obtained and added. 878359191Skris [Steve Henson] 878459191Skris 878559191Skris *) -crlf option to s_client and s_server for sending newlines as 878659191Skris CRLF (as required by many protocols). 878759191Skris [Bodo Moeller] 878859191Skris 878955714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 879055714Skris 879155714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 879255714Skris [Ralf S. Engelschall] 879355714Skris 879455714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 879555714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 879655714Skris 879755714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 879855714Skris program. 879955714Skris [Steve Henson] 880055714Skris 880155714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 880255714Skris DH parameters/keys (q is lost during that conversion, but the resulting 880355714Skris DH parameters contain its length). 880455714Skris 880555714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 880655714Skris much faster than DH_generate_parameters (which creates parameters 880755714Skris where p = 2*q + 1), and also the smaller q makes DH computations 880855714Skris much more efficient (160-bit exponentiation instead of 1024-bit 880955714Skris exponentiation); so this provides a convenient way to support DHE 881055714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 881155714Skris utter importance to use 881255714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 881355714Skris or 881455714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 881555714Skris when such DH parameters are used, because otherwise small subgroup 881655714Skris attacks may become possible! 881755714Skris [Bodo Moeller] 881855714Skris 881955714Skris *) Avoid memory leak in i2d_DHparams. 882055714Skris [Bodo Moeller] 882155714Skris 882255714Skris *) Allow the -k option to be used more than once in the enc program: 882355714Skris this allows the same encrypted message to be read by multiple recipients. 882455714Skris [Steve Henson] 882555714Skris 882655714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 882755714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 882855714Skris it will always use the numerical form of the OID, even if it has a short 882955714Skris or long name. 883055714Skris [Steve Henson] 883155714Skris 883255714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 883355714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 883455714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 883555714Skris no private key components need be present and it might store extra data 883659191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 883759191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 883859191Skris private key operations. 883955714Skris [Steve Henson] 884055714Skris 884155714Skris *) Added support for SPARC Linux. 884255714Skris [Andy Polyakov] 884355714Skris 884455714Skris *) pem_password_cb function type incompatibly changed from 884555714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 884655714Skris to 884755714Skris ....(char *buf, int size, int rwflag, void *userdata); 884855714Skris so that applications can pass data to their callbacks: 884955714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 885055714Skris additional void * argument, which is just handed through whenever 885155714Skris the password callback is called. 885259191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 885355714Skris 885455714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 885555714Skris 885655714Skris Compatibility note: As many C implementations push function arguments 885755714Skris onto the stack in reverse order, the new library version is likely to 885855714Skris interoperate with programs that have been compiled with the old 885955714Skris pem_password_cb definition (PEM_whatever takes some data that 886055714Skris happens to be on the stack as its last argument, and the callback 886155714Skris just ignores this garbage); but there is no guarantee whatsoever that 886255714Skris this will work. 886355714Skris 886455714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 886555714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 886655714Skris problems not only on Windows, but also on some Unix platforms. 886755714Skris To avoid problematic command lines, these definitions are now in an 886855714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 886955714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 887055714Skris [Bodo Moeller] 887155714Skris 887255714Skris *) MIPS III/IV assembler module is reimplemented. 887355714Skris [Andy Polyakov] 887455714Skris 887555714Skris *) More DES library cleanups: remove references to srand/rand and 887655714Skris delete an unused file. 887755714Skris [Ulf M�ller] 887855714Skris 887955714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 888055714Skris since not many people have MASM (ml) and it can be hard to obtain. 888155714Skris This is currently experimental but it seems to work OK and pass all 888255714Skris the tests. Check out INSTALL.W32 for info. 888355714Skris [Steve Henson] 888455714Skris 888555714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 888655714Skris without temporary keys kept an extra copy of the server key, 888755714Skris and connections with temporary keys did not free everything in case 888855714Skris of an error. 888955714Skris [Bodo Moeller] 889055714Skris 889155714Skris *) New function RSA_check_key and new openssl rsa option -check 889255714Skris for verifying the consistency of RSA keys. 889355714Skris [Ulf Moeller, Bodo Moeller] 889455714Skris 889555714Skris *) Various changes to make Win32 compile work: 889655714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 889755714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 889855714Skris comparison" warnings. 889955714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 890055714Skris [Steve Henson] 890155714Skris 890255714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 890355714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 890455714Skris derived keys are printed to stderr. 890555714Skris [Steve Henson] 890655714Skris 890755714Skris *) Copy the flags in ASN1_STRING_dup(). 890855714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 890955714Skris 891055714Skris *) The x509 application mishandled signing requests containing DSA 891155714Skris keys when the signing key was also DSA and the parameters didn't match. 891255714Skris 891355714Skris It was supposed to omit the parameters when they matched the signing key: 891455714Skris the verifying software was then supposed to automatically use the CA's 891555714Skris parameters if they were absent from the end user certificate. 891655714Skris 891755714Skris Omitting parameters is no longer recommended. The test was also 891855714Skris the wrong way round! This was probably due to unusual behaviour in 891955714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 892055714Skris This meant that parameters were omitted when they *didn't* match and 892155714Skris the certificate was useless. Certificates signed with 'ca' didn't have 892255714Skris this bug. 892355714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 892455714Skris 892555714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 892655714Skris The interface is as follows: 892755714Skris Applications can use 892855714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 892955714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 893055714Skris "off" is now the default. 893155714Skris The library internally uses 893255714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 893355714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 893455714Skris to disable memory-checking temporarily. 893555714Skris 893655714Skris Some inconsistent states that previously were possible (and were 893755714Skris even the default) are now avoided. 893855714Skris 893955714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 894055714Skris with each memory chunk allocated; this is occasionally more helpful 894155714Skris than just having a counter. 894255714Skris 894355714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 894455714Skris 894555714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 894655714Skris extensions. 894755714Skris [Bodo Moeller] 894855714Skris 894955714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 895055714Skris which largely parallels "options", but is for changing API behaviour, 895155714Skris whereas "options" are about protocol behaviour. 895255714Skris Initial "mode" flags are: 895355714Skris 895455714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 895555714Skris a single record has been written. 895655714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 895755714Skris retries use the same buffer location. 895855714Skris (But all of the contents must be 895955714Skris copied!) 896055714Skris [Bodo Moeller] 896155714Skris 896279998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 896355714Skris worked. 896455714Skris 896555714Skris *) Fix problems with no-hmac etc. 896655714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 896755714Skris 896855714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 896955714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 897055714Skris to mess around with the internals of an RSA structure. 897155714Skris [Steve Henson] 897255714Skris 897355714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 897455714Skris Also really enable memory leak checks in openssl.c and in some 897555714Skris test programs. 897655714Skris [Chad C. Mulligan, Bodo Moeller] 897755714Skris 897855714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 897955714Skris up the length of negative integers. This has now been simplified to just 898055714Skris store the length when it is first determined and use it later, rather 898155714Skris than trying to keep track of where data is copied and updating it to 898255714Skris point to the end. 898355714Skris [Steve Henson, reported by Brien Wheeler 898455714Skris <bwheeler@authentica-security.com>] 898555714Skris 898655714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 898755714Skris of a PKCS#7 signature but with the signing certificate passed to the 898855714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 898955714Skris certificate is present in the PKCS#7 structure. This isn't always the 899055714Skris case: certificates can be omitted from a PKCS#7 structure and be 899155714Skris distributed by "out of band" means (such as a certificate database). 899255714Skris [Steve Henson] 899355714Skris 899455714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 899555714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 899655714Skris necessary function names. 899755714Skris [Steve Henson] 899855714Skris 899955714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 900055714Skris options set by Configure in the top level Makefile, and Configure 900155714Skris was not even able to write more than one option correctly. 900255714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 900355714Skris [Bodo Moeller] 900455714Skris 900555714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 900655714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 900755714Skris for example allow memory BIOs to contain config info. 900855714Skris [Steve Henson] 900955714Skris 901055714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 901155714Skris Whoever hopes to achieve shared-library compatibility across versions 901255714Skris must use this, not the compile-time macro. 901355714Skris (Exercise 0.9.4: Which is the minimum library version required by 901455714Skris such programs?) 901555714Skris Note: All this applies only to multi-threaded programs, others don't 901655714Skris need locks. 901755714Skris [Bodo Moeller] 901855714Skris 901955714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 902055714Skris through a BIO pair triggered the default case, i.e. 902155714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 902255714Skris [Bodo Moeller] 902355714Skris 902455714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 902555714Skris can use the SSL library even if none of the specific BIOs is 902655714Skris appropriate. 902755714Skris [Bodo Moeller] 902855714Skris 902955714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 903055714Skris for the encoded length. 903155714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 903255714Skris 903355714Skris *) Add initial documentation of the X509V3 functions. 903455714Skris [Steve Henson] 903555714Skris 903655714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 903755714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 903855714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 903955714Skris secure PKCS#8 private key format with a high iteration count. 904055714Skris [Steve Henson] 904155714Skris 904255714Skris *) Fix determination of Perl interpreter: A perl or perl5 904355714Skris _directory_ in $PATH was also accepted as the interpreter. 904455714Skris [Ralf S. Engelschall] 904555714Skris 904655714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 904755714Skris wrong with it but it was very old and did things like calling 904855714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 904955714Skris unusual formatting. 905055714Skris [Steve Henson] 905155714Skris 905255714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 905355714Skris to use the new extension code. 905455714Skris [Steve Henson] 905555714Skris 905655714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 905755714Skris with macros. This should make it easier to change their form, add extra 905855714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 905955714Skris constant. 906055714Skris [Steve Henson] 906155714Skris 906255714Skris *) Add to configuration table a new entry that can specify an alternative 906355714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 906455714Skris according to Mark Crispin <MRC@Panda.COM>. 906555714Skris [Bodo Moeller] 906655714Skris 906755714Skris#if 0 906855714Skris *) DES CBC did not update the IV. Weird. 906955714Skris [Ben Laurie] 907055714Skris#else 907155714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 907255714Skris Changing the behaviour of the former might break existing programs -- 907355714Skris where IV updating is needed, des_ncbc_encrypt can be used. 907455714Skris#endif 907555714Skris 907655714Skris *) When bntest is run from "make test" it drives bc to check its 907755714Skris calculations, as well as internally checking them. If an internal check 907855714Skris fails, it needs to cause bc to give a non-zero result or make test carries 907955714Skris on without noticing the failure. Fixed. 908055714Skris [Ben Laurie] 908155714Skris 908255714Skris *) DES library cleanups. 908355714Skris [Ulf M�ller] 908455714Skris 908555714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 908655714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 908755714Skris ciphers. NOTE: although the key derivation function has been verified 908855714Skris against some published test vectors it has not been extensively tested 908955714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 909055714Skris of v2.0. 909155714Skris [Steve Henson] 909255714Skris 909355714Skris *) Instead of "mkdir -p", which is not fully portable, use new 909455714Skris Perl script "util/mkdir-p.pl". 909555714Skris [Bodo Moeller] 909655714Skris 909755714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 909855714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 909955714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 910055714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 910155714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 910255714Skris underlying key generation function so it must do its own ASN1 parsing. 910355714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 910455714Skris 'parameter' argument instead of literal salt and iteration count values 910555714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 910655714Skris [Steve Henson] 910755714Skris 910855714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 910955714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 911055714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 911155714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 911255714Skris value was just used as a "magic string" and not used directly its 911355714Skris value doesn't matter. 911455714Skris [Steve Henson] 911555714Skris 911655714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 911755714Skris support mutable. 911855714Skris [Ben Laurie] 911955714Skris 912055714Skris *) "linux-sparc64" configuration (ultrapenguin). 912155714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 912255714Skris "linux-sparc" configuration. 912355714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 912455714Skris 912555714Skris *) config now generates no-xxx options for missing ciphers. 912655714Skris [Ulf M�ller] 912755714Skris 912855714Skris *) Support the EBCDIC character set (work in progress). 912955714Skris File ebcdic.c not yet included because it has a different license. 913055714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 913155714Skris 913255714Skris *) Support BS2000/OSD-POSIX. 913355714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 913455714Skris 913555714Skris *) Make callbacks for key generation use void * instead of char *. 913655714Skris [Ben Laurie] 913755714Skris 913855714Skris *) Make S/MIME samples compile (not yet tested). 913955714Skris [Ben Laurie] 914055714Skris 914155714Skris *) Additional typesafe stacks. 914255714Skris [Ben Laurie] 914355714Skris 914455714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 914555714Skris [Bodo Moeller] 914655714Skris 914755714Skris 914855714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 914955714Skris 915055714Skris *) New configuration variant "sco5-gcc". 915155714Skris 915255714Skris *) Updated some demos. 915355714Skris [Sean O Riordain, Wade Scholine] 915455714Skris 915555714Skris *) Add missing BIO_free at exit of pkcs12 application. 915655714Skris [Wu Zhigang] 915755714Skris 915855714Skris *) Fix memory leak in conf.c. 915955714Skris [Steve Henson] 916055714Skris 916155714Skris *) Updates for Win32 to assembler version of MD5. 916255714Skris [Steve Henson] 916355714Skris 916455714Skris *) Set #! path to perl in apps/der_chop to where we found it 916555714Skris instead of using a fixed path. 916655714Skris [Bodo Moeller] 916755714Skris 916855714Skris *) SHA library changes for irix64-mips4-cc. 916955714Skris [Andy Polyakov] 917055714Skris 917155714Skris *) Improvements for VMS support. 917255714Skris [Richard Levitte] 917355714Skris 917455714Skris 917555714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 917655714Skris 917755714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 917855714Skris This also avoids the problems with SC4.2 and unpatched SC5. 917955714Skris [Andy Polyakov <appro@fy.chalmers.se>] 918055714Skris 918155714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 918255714Skris These are required because of the typesafe stack would otherwise break 918355714Skris existing code. If old code used a structure member which used to be STACK 918455714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 918555714Skris sk_num or sk_value it would produce an error because the num, data members 918655714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 918755714Skris replaces the old method of assigning a value to sk_value 918855714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 918955714Skris that does this will no longer work (and should use sk_set instead) but 919055714Skris this could be regarded as a "questionable" behaviour anyway. 919155714Skris [Steve Henson] 919255714Skris 919355714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 919455714Skris correctly handle encrypted S/MIME data. 919555714Skris [Steve Henson] 919655714Skris 919755714Skris *) Change type of various DES function arguments from des_cblock 919855714Skris (which means, in function argument declarations, pointer to char) 919955714Skris to des_cblock * (meaning pointer to array with 8 char elements), 920055714Skris which allows the compiler to do more typechecking; it was like 920155714Skris that back in SSLeay, but with lots of ugly casts. 920255714Skris 920355714Skris Introduce new type const_des_cblock. 920455714Skris [Bodo Moeller] 920555714Skris 920655714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 920755714Skris problems: find RecipientInfo structure that matches recipient certificate 920855714Skris and initialise the ASN1 structures properly based on passed cipher. 920955714Skris [Steve Henson] 921055714Skris 921155714Skris *) Belatedly make the BN tests actually check the results. 921255714Skris [Ben Laurie] 921355714Skris 921455714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 921555714Skris to and from BNs: it was completely broken. New compilation option 921655714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 921755714Skris key elements as negative integers. 921855714Skris [Steve Henson] 921955714Skris 922055714Skris *) Reorganize and speed up MD5. 922155714Skris [Andy Polyakov <appro@fy.chalmers.se>] 922255714Skris 922355714Skris *) VMS support. 922455714Skris [Richard Levitte <richard@levitte.org>] 922555714Skris 922655714Skris *) New option -out to asn1parse to allow the parsed structure to be 922755714Skris output to a file. This is most useful when combined with the -strparse 922855714Skris option to examine the output of things like OCTET STRINGS. 922955714Skris [Steve Henson] 923055714Skris 923155714Skris *) Make SSL library a little more fool-proof by not requiring any longer 923255714Skris that SSL_set_{accept,connect}_state be called before 923355714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 923455714Skris in many applications because usually everything *appeared* to work as 923555714Skris intended anyway -- now it really works as intended). 923655714Skris [Bodo Moeller] 923755714Skris 923855714Skris *) Move openssl.cnf out of lib/. 923955714Skris [Ulf M�ller] 924055714Skris 924155714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 924255714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 924355714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 924455714Skris [Ralf S. Engelschall] 924555714Skris 924655714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 924755714Skris handle PKCS#7 enveloped data properly. 924855714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 924955714Skris 925055714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 925155714Skris copying pointers. The cert_st handling is changed by this in 925255714Skris various ways (and thus what used to be known as ctx->default_cert 925355714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 925455714Skris any longer when s->cert does not give us what we need). 925555714Skris ssl_cert_instantiate becomes obsolete by this change. 925655714Skris As soon as we've got the new code right (possibly it already is?), 925755714Skris we have solved a couple of bugs of the earlier code where s->cert 925855714Skris was used as if it could not have been shared with other SSL structures. 925955714Skris 926055714Skris Note that using the SSL API in certain dirty ways now will result 926155714Skris in different behaviour than observed with earlier library versions: 926255714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 926355714Skris does not influence s as it used to. 926455714Skris 926555714Skris In order to clean up things more thoroughly, inside SSL_SESSION 926655714Skris we don't use CERT any longer, but a new structure SESS_CERT 926755714Skris that holds per-session data (if available); currently, this is 926855714Skris the peer's certificate chain and, for clients, the server's certificate 926955714Skris and temporary key. CERT holds only those values that can have 927055714Skris meaningful defaults in an SSL_CTX. 927155714Skris [Bodo Moeller] 927255714Skris 927355714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 927455714Skris from the internal representation. Various PKCS#7 fixes: remove some 927555714Skris evil casts and set the enc_dig_alg field properly based on the signing 927655714Skris key type. 927755714Skris [Steve Henson] 927855714Skris 927955714Skris *) Allow PKCS#12 password to be set from the command line or the 928055714Skris environment. Let 'ca' get its config file name from the environment 928155714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 928255714Skris and 'x509'). 928355714Skris [Steve Henson] 928455714Skris 928555714Skris *) Allow certificate policies extension to use an IA5STRING for the 928655714Skris organization field. This is contrary to the PKIX definition but 928755714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 928855714Skris extension option. 928955714Skris [Steve Henson] 929055714Skris 929155714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 929255714Skris without disallowing inline assembler and the like for non-pedantic builds. 929355714Skris [Ben Laurie] 929455714Skris 929555714Skris *) Support Borland C++ builder. 929655714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 929755714Skris 929855714Skris *) Support Mingw32. 929955714Skris [Ulf M�ller] 930055714Skris 930155714Skris *) SHA-1 cleanups and performance enhancements. 930255714Skris [Andy Polyakov <appro@fy.chalmers.se>] 930355714Skris 930455714Skris *) Sparc v8plus assembler for the bignum library. 930555714Skris [Andy Polyakov <appro@fy.chalmers.se>] 930655714Skris 930755714Skris *) Accept any -xxx and +xxx compiler options in Configure. 930855714Skris [Ulf M�ller] 930955714Skris 931055714Skris *) Update HPUX configuration. 931155714Skris [Anonymous] 931255714Skris 931355714Skris *) Add missing sk_<type>_unshift() function to safestack.h 931455714Skris [Ralf S. Engelschall] 931555714Skris 931655714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 931755714Skris "extra_cert"s in addition to the certificate. (This makes sense 931855714Skris only for "PEM" format files, as chains as a whole are not 931955714Skris DER-encoded.) 932055714Skris [Bodo Moeller] 932155714Skris 932255714Skris *) Support verify_depth from the SSL API. 932355714Skris x509_vfy.c had what can be considered an off-by-one-error: 932455714Skris Its depth (which was not part of the external interface) 932555714Skris was actually counting the number of certificates in a chain; 932655714Skris now it really counts the depth. 932755714Skris [Bodo Moeller] 932855714Skris 932955714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 933055714Skris instead of X509err, which often resulted in confusing error 933155714Skris messages since the error codes are not globally unique 933255714Skris (e.g. an alleged error in ssl3_accept when a certificate 933355714Skris didn't match the private key). 933455714Skris 933555714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 933655714Skris value (so that you don't need SSL_set_session_id_context for each 933755714Skris connection using the SSL_CTX). 933855714Skris [Bodo Moeller] 933955714Skris 934055714Skris *) OAEP decoding bug fix. 934155714Skris [Ulf M�ller] 934255714Skris 934355714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 934455714Skris David Harris. 934555714Skris [Bodo Moeller] 934655714Skris 934755714Skris *) New Configure options "threads" and "no-threads". For systems 934855714Skris where the proper compiler options are known (currently Solaris 934955714Skris and Linux), "threads" is the default. 935055714Skris [Bodo Moeller] 935155714Skris 935255714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 935355714Skris [Bodo Moeller] 935455714Skris 935555714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 935655714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 935755714Skris such as /usr/local/bin. 935855714Skris [Bodo Moeller] 935955714Skris 936055714Skris *) "make linux-shared" to build shared libraries. 936155714Skris [Niels Poppe <niels@netbox.org>] 936255714Skris 936355714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 936455714Skris [Ulf M�ller] 936555714Skris 936655714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 936755714Skris extension adding in x509 utility. 936855714Skris [Steve Henson] 936955714Skris 937055714Skris *) Remove NOPROTO sections and error code comments. 937155714Skris [Ulf M�ller] 937255714Skris 937355714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 937455714Skris prototypes. 937555714Skris [Steve Henson] 937655714Skris 937755714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 937855714Skris [Ulf M�ller] 937955714Skris 938055714Skris *) Complete rewrite of the error code script(s). It is all now handled 938155714Skris by one script at the top level which handles error code gathering, 938255714Skris header rewriting and C source file generation. It should be much better 938355714Skris than the old method: it now uses a modified version of Ulf's parser to 938455714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 938555714Skris aren't needed for error creation any more) and do a better job of 938655714Skris translating function codes into names. The old 'ASN1 error code imbedded 938755714Skris in a comment' is no longer necessary and it doesn't use .err files which 938855714Skris have now been deleted. Also the error code call doesn't have to appear all 938955714Skris on one line (which resulted in some large lines...). 939055714Skris [Steve Henson] 939155714Skris 939255714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 939355714Skris [Bodo Moeller] 939455714Skris 939555714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 939655714Skris 0 (which usually indicates a closed connection), but continue reading. 939755714Skris [Bodo Moeller] 939855714Skris 939955714Skris *) Fix some race conditions. 940055714Skris [Bodo Moeller] 940155714Skris 940255714Skris *) Add support for CRL distribution points extension. Add Certificate 940355714Skris Policies and CRL distribution points documentation. 940455714Skris [Steve Henson] 940555714Skris 940655714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 940755714Skris [Ulf M�ller] 940855714Skris 940955714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 941055714Skris 8 of keying material. Merlin has also confirmed interop with this fix 941155714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 941255714Skris [Merlin Hughes <merlin@baltimore.ie>] 941355714Skris 941455714Skris *) Fix lots of warnings. 941555714Skris [Richard Levitte <levitte@stacken.kth.se>] 941655714Skris 941755714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 941855714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 941955714Skris [Richard Levitte <levitte@stacken.kth.se>] 942055714Skris 942155714Skris *) Fix problems with sizeof(long) == 8. 942255714Skris [Andy Polyakov <appro@fy.chalmers.se>] 942355714Skris 942455714Skris *) Change functions to ANSI C. 942555714Skris [Ulf M�ller] 942655714Skris 942755714Skris *) Fix typos in error codes. 942855714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 942955714Skris 943055714Skris *) Remove defunct assembler files from Configure. 943155714Skris [Ulf M�ller] 943255714Skris 943355714Skris *) SPARC v8 assembler BIGNUM implementation. 943455714Skris [Andy Polyakov <appro@fy.chalmers.se>] 943555714Skris 943655714Skris *) Support for Certificate Policies extension: both print and set. 943755714Skris Various additions to support the r2i method this uses. 943855714Skris [Steve Henson] 943955714Skris 944055714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 944155714Skris return a const string when you are expecting an allocated buffer. 944255714Skris [Ben Laurie] 944355714Skris 944455714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 944555714Skris types DirectoryString and DisplayText. 944655714Skris [Steve Henson] 944755714Skris 944855714Skris *) Add code to allow r2i extensions to access the configuration database, 944955714Skris add an LHASH database driver and add several ctx helper functions. 945055714Skris [Steve Henson] 945155714Skris 945255714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 945355714Skris fail when they extended the size of a BIGNUM. 945455714Skris [Steve Henson] 945555714Skris 945655714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 945755714Skris support typesafe stack. 945855714Skris [Steve Henson] 945955714Skris 946055714Skris *) Fix typo in SSL_[gs]et_options(). 946155714Skris [Nils Frostberg <nils@medcom.se>] 946255714Skris 946355714Skris *) Delete various functions and files that belonged to the (now obsolete) 946455714Skris old X509V3 handling code. 946555714Skris [Steve Henson] 946655714Skris 946755714Skris *) New Configure option "rsaref". 946855714Skris [Ulf M�ller] 946955714Skris 947055714Skris *) Don't auto-generate pem.h. 947155714Skris [Bodo Moeller] 947255714Skris 947355714Skris *) Introduce type-safe ASN.1 SETs. 947455714Skris [Ben Laurie] 947555714Skris 947655714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 947755714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 947855714Skris 947955714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 948055714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 948155714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 948255714Skris few STACKed things have been converted already. Feel free to convert more. 948355714Skris In the fullness of time, I'll do away with the STACK type altogether. 948455714Skris [Ben Laurie] 948555714Skris 948655714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 948755714Skris specified in <certfile> by updating the entry in the index.txt file. 948855714Skris This way one no longer has to edit the index.txt file manually for 948955714Skris revoking a certificate. The -revoke option does the gory details now. 949055714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 949155714Skris 949255714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 949355714Skris `-text' option at all and this way the `-noout -text' combination was 949455714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 949555714Skris [Ralf S. Engelschall] 949655714Skris 949755714Skris *) Make sure a corresponding plain text error message exists for the 949855714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 949955714Skris verify callback function determined that a certificate was revoked. 950055714Skris [Ralf S. Engelschall] 950155714Skris 950255714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 950355714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 950455714Skris all available cipers including rc5, which was forgotten until now. 950555714Skris In order to let the testing shell script know which algorithms 950655714Skris are available, a new (up to now undocumented) command 950755714Skris "openssl list-cipher-commands" is used. 950855714Skris [Bodo Moeller] 950955714Skris 951055714Skris *) Bugfix: s_client occasionally would sleep in select() when 951155714Skris it should have checked SSL_pending() first. 951255714Skris [Bodo Moeller] 951355714Skris 951455714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 951555714Skris the raw DSA values prior to ASN.1 encoding. 951655714Skris [Ulf M�ller] 951755714Skris 951855714Skris *) Tweaks to Configure 951955714Skris [Niels Poppe <niels@netbox.org>] 952055714Skris 952155714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 952255714Skris yet... 952355714Skris [Steve Henson] 952455714Skris 952555714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 952655714Skris [Ulf M�ller] 952755714Skris 952855714Skris *) New config option to avoid instructions that are illegal on the 80386. 952955714Skris The default code is faster, but requires at least a 486. 953055714Skris [Ulf M�ller] 953155714Skris 953255714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 953355714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 953455714Skris same as SSL2_VERSION anyway. 953555714Skris [Bodo Moeller] 953655714Skris 953755714Skris *) New "-showcerts" option for s_client. 953855714Skris [Bodo Moeller] 953955714Skris 954055714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 954155714Skris application. Various cleanups and fixes. 954255714Skris [Steve Henson] 954355714Skris 954455714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 954555714Skris modify error routines to work internally. Add error codes and PBE init 954655714Skris to library startup routines. 954755714Skris [Steve Henson] 954855714Skris 954955714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 955055714Skris packing functions to asn1 and evp. Changed function names and error 955155714Skris codes along the way. 955255714Skris [Steve Henson] 955355714Skris 955455714Skris *) PKCS12 integration: and so it begins... First of several patches to 955555714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 955655714Skris objects to objects.h 955755714Skris [Steve Henson] 955855714Skris 955955714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 956055714Skris and display support for Thawte strong extranet extension. 956155714Skris [Steve Henson] 956255714Skris 956355714Skris *) Add LinuxPPC support. 956455714Skris [Jeff Dubrule <igor@pobox.org>] 956555714Skris 956655714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 956755714Skris bn_div_words in alpha.s. 956855714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 956955714Skris 957055714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 957155714Skris OAEP isn't supported when OpenSSL is built with RSAref. 957255714Skris [Ulf Moeller <ulf@fitug.de>] 957355714Skris 957455714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 957555714Skris so they no longer are missing under -DNOPROTO. 957655714Skris [Soren S. Jorvang <soren@t.dk>] 957755714Skris 957855714Skris 957955714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 958055714Skris 958155714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 958255714Skris doesn't work when the session is reused. Coming soon! 958355714Skris [Ben Laurie] 958455714Skris 958555714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 958655714Skris context thus bypassing client cert protection! All software that uses 958755714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 958855714Skris allow session reuse! A fuller solution is in the works. 958955714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 959055714Skris 959155714Skris *) Some more source tree cleanups (removed obsolete files 959255714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 959355714Skris permission on "config" script to be executable) and a fix for the INSTALL 959455714Skris document. 959555714Skris [Ulf Moeller <ulf@fitug.de>] 959655714Skris 959755714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 959855714Skris Malloc, Free. 959955714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 960055714Skris 960155714Skris *) Make rsa_oaep_test return non-zero on error. 960255714Skris [Ulf Moeller <ulf@fitug.de>] 960355714Skris 960455714Skris *) Add support for native Solaris shared libraries. Configure 960555714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 960655714Skris if someone would make that last step automatic. 960755714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 960855714Skris 960955714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 961055714Skris [Ben Laurie] 961155714Skris 961255714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 961355714Skris except NULL ciphers". This means the default cipher list will no longer 961455714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 961555714Skris the string "DEFAULT:eNULL". 961655714Skris [Steve Henson] 961755714Skris 961855714Skris *) Fix to RSA private encryption routines: if p < q then it would 961955714Skris occasionally produce an invalid result. This will only happen with 962055714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 962155714Skris [Steve Henson] 962255714Skris 962355714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 962455714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 962555714Skris because this way one can also use an interpreter named `perl5' (which is 962655714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 962755714Skris installed as `perl'). 962855714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 962955714Skris 963055714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 963155714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 963255714Skris 963355714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 963455714Skris advapi32.lib to Win32 build and change the pem test comparision 963555714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 963655714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 963755714Skris and crypto/des/ede_cbcm_enc.c. 963855714Skris [Steve Henson] 963955714Skris 964055714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 964155714Skris [Ben Laurie] 964255714Skris 964355714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 964455714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 964555714Skris is horrible: I feel ill.... 964655714Skris [Steve Henson] 964755714Skris 964855714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 964955714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 965055714Skris sections: 10 functions were absent from non ANSI section and not exported 965155714Skris from Windows DLLs. Fixed up libeay.num for new functions. 965255714Skris [Steve Henson] 965355714Skris 965455714Skris *) Make `openssl version' output lines consistent. 965555714Skris [Ralf S. Engelschall] 965655714Skris 965755714Skris *) Fix Win32 symbol export lists for BIO functions: Added 965855714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 965955714Skris to ms/libeay{16,32}.def. 966055714Skris [Ralf S. Engelschall] 966155714Skris 966255714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 966355714Skris fine under Unix and passes some trivial tests I've now added. But the 966455714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 966555714Skris added to make sure no one expects that this stuff really works in the 966655714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 966755714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 966855714Skris openssl_bio.xs. 966955714Skris [Ralf S. Engelschall] 967055714Skris 967155714Skris *) Fix the generation of two part addresses in perl. 967255714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 967355714Skris 967455714Skris *) Add config entry for Linux on MIPS. 967555714Skris [John Tobey <jtobey@channel1.com>] 967655714Skris 967755714Skris *) Make links whenever Configure is run, unless we are on Windoze. 967855714Skris [Ben Laurie] 967955714Skris 968055714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 968155714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 968255714Skris in CRLs. 968355714Skris [Steve Henson] 968455714Skris 968555714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 968655714Skris other platforms details on the command line without having to patch the 968755714Skris Configure script everytime: One now can use ``perl Configure 968855714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 968955714Skris to them (seperated by colons). This is treated as there would be a static 969055714Skris pre-configured entry in Configure's %table under key <id> with value 969155714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 969255714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 969355714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 969455714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 969555714Skris [Ralf S. Engelschall] 969655714Skris 969755714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 969855714Skris [Ben Laurie] 969955714Skris 970055714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 970155714Skris on the `perl Configure ...' command line. This way one can compile 970255714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 970355714Skris for linking it into DSOs. 970455714Skris [Ralf S. Engelschall] 970555714Skris 970655714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 970755714Skris Fixed. 970855714Skris [Ben Laurie] 970955714Skris 971055714Skris *) Cleaned up the LICENSE document: The official contact for any license 971155714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 971255714Skris And add a paragraph about the dual-license situation to make sure people 971355714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 971455714Skris to the OpenSSL toolkit. 971555714Skris [Ralf S. Engelschall] 971655714Skris 971755714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 971855714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 971955714Skris Additonally cleaned up the `make links' target: Remove unnecessary 972055714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 972155714Skris to speed processing and no longer clutter the display with confusing 972255714Skris stuff. Instead only the actually done links are displayed. 972355714Skris [Ralf S. Engelschall] 972455714Skris 972555714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 972655714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 972755714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 972855714Skris encryption. 972955714Skris [Ben Laurie] 973055714Skris 973155714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 973255714Skris signed attributes when verifying signatures (this would break them), 973355714Skris the detached data encoding was wrong and public keys obtained using 973455714Skris X509_get_pubkey() weren't freed. 973555714Skris [Steve Henson] 973655714Skris 973755714Skris *) Add text documentation for the BUFFER functions. Also added a work around 973855714Skris to a Win95 console bug. This was triggered by the password read stuff: the 973955714Skris last character typed gets carried over to the next fread(). If you were 974055714Skris generating a new cert request using 'req' for example then the last 974155714Skris character of the passphrase would be CR which would then enter the first 974255714Skris field as blank. 974355714Skris [Steve Henson] 974455714Skris 974555714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 974655714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 974755714Skris button and can be used by applications based on OpenSSL to show the 974855714Skris relationship to the OpenSSL project. 974955714Skris [Ralf S. Engelschall] 975055714Skris 975155714Skris *) Remove confusing variables in function signatures in files 975255714Skris ssl/ssl_lib.c and ssl/ssl.h. 975355714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 975455714Skris 975555714Skris *) Don't install bss_file.c under PREFIX/include/ 975655714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 975755714Skris 975855714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 975955714Skris functions that return function pointers and has support for NT specific 976055714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 976155714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 976255714Skris unsigned to signed types: this was killing the Win32 compile. 976355714Skris [Steve Henson] 976455714Skris 976555714Skris *) Add new certificate file to stack functions, 976655714Skris SSL_add_dir_cert_subjects_to_stack() and 976755714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 976855714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 976955714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 977055714Skris This means that Apache-SSL and similar packages don't have to mess around 977155714Skris to add as many CAs as they want to the preferred list. 977255714Skris [Ben Laurie] 977355714Skris 977455714Skris *) Experiment with doxygen documentation. Currently only partially applied to 977555714Skris ssl/ssl_lib.c. 977655714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 977755714Skris openssl.doxy as the configuration file. 977855714Skris [Ben Laurie] 977955714Skris 978055714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 978155714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 978255714Skris 978355714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 978455714Skris compiled in by default: it has problems with large keys. 978555714Skris [Steve Henson] 978655714Skris 978755714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 978855714Skris DH private keys and/or callback functions which directly correspond to 978955714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 979055714Skris is needed for applications which have to configure certificates on a 979155714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 979255714Skris (e.g. s_server). 979355714Skris For the RSA certificate situation is makes no difference, but 979455714Skris for the DSA certificate situation this fixes the "no shared cipher" 979555714Skris problem where the OpenSSL cipher selection procedure failed because the 979655714Skris temporary keys were not overtaken from the context and the API provided 979755714Skris no way to reconfigure them. 979855714Skris The new functions now let applications reconfigure the stuff and they 979955714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 980055714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 980155714Skris non-public-API function ssl_cert_instantiate() is used as a helper 980255714Skris function and also to reduce code redundancy inside ssl_rsa.c. 980355714Skris [Ralf S. Engelschall] 980455714Skris 980555714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 980655714Skris area because they are useful for the DSA situation and should be 980755714Skris recognized by the users. 980855714Skris [Ralf S. Engelschall] 980955714Skris 981055714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 981155714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 981255714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 981355714Skris already masked variable. 981455714Skris [Richard Levitte <levitte@stacken.kth.se>] 981555714Skris 981655714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 981755714Skris [Richard Levitte <levitte@stacken.kth.se>] 981855714Skris 981955714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 982055714Skris from `int' to `unsigned int' because it's a length and initialized by 982155714Skris EVP_DigestFinal() which expects an `unsigned int *'. 982255714Skris [Richard Levitte <levitte@stacken.kth.se>] 982355714Skris 982455714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 982555714Skris script. Instead use the usual Shell->Perl transition trick. 982655714Skris [Ralf S. Engelschall] 982755714Skris 982855714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 982955714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 983055714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 983155714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 983255714Skris currently the public key is printed (a decision which was already done by 983355714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 983455714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 983555714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 983655714Skris now, too. 983755714Skris [Ralf S. Engelschall] 983855714Skris 983955714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 984055714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 984155714Skris [Arne Ansper <arne@ats.cyber.ee>] 984255714Skris 984355714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 984455714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 984555714Skris config file. 984655714Skris [Steve Henson] 984755714Skris 984855714Skris *) Add cool BIO that does syslog (or event log on NT). 984955714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 985055714Skris 985155714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 985255714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 985355714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 985455714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 985555714Skris [Ben Laurie] 985655714Skris 985755714Skris *) Add preliminary config info for new extension code. 985855714Skris [Steve Henson] 985955714Skris 986055714Skris *) Make RSA_NO_PADDING really use no padding. 986155714Skris [Ulf Moeller <ulf@fitug.de>] 986255714Skris 986355714Skris *) Generate errors when private/public key check is done. 986455714Skris [Ben Laurie] 986555714Skris 986655714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 986755714Skris for some CRL extensions and new objects added. 986855714Skris [Steve Henson] 986955714Skris 987055714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 987155714Skris key usage extension and fuller support for authority key id. 987255714Skris [Steve Henson] 987355714Skris 987455714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 987555714Skris padding method for RSA, which is recommended for new applications in PKCS 987655714Skris #1 v2.0 (RFC 2437, October 1998). 987755714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 987855714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 987955714Skris against Bleichbacher's attack on RSA. 988055714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 988155714Skris Ben Laurie] 988255714Skris 988355714Skris *) Updates to the new SSL compression code 988455714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 988555714Skris 988655714Skris *) Fix so that the version number in the master secret, when passed 988755714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 988855714Skris (because the server will not accept higher), that the version number 988955714Skris is 0x03,0x01, not 0x03,0x00 989055714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 989155714Skris 989255714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 989355714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 989455714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 989555714Skris [Steve Henson] 989655714Skris 989755714Skris *) Support for RAW extensions where an arbitrary extension can be 989855714Skris created by including its DER encoding. See apps/openssl.cnf for 989955714Skris an example. 990055714Skris [Steve Henson] 990155714Skris 990255714Skris *) Make sure latest Perl versions don't interpret some generated C array 990355714Skris code as Perl array code in the crypto/err/err_genc.pl script. 990455714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 990555714Skris 990655714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 990755714Skris not many people have the assembler. Various Win32 compilation fixes and 990855714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 990955714Skris build instructions. 991055714Skris [Steve Henson] 991155714Skris 991255714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 991355714Skris file under Win32 and also build pem.h from pem.org. New script 991455714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 991555714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 991655714Skris [Steve Henson] 991755714Skris 991855714Skris *) Major rework of DES function declarations, in the pursuit of correctness 991955714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 992055714Skris too. You may find this causes warnings in your code. Zapping your evil 992155714Skris casts will probably fix them. Mostly. 992255714Skris [Ben Laurie] 992355714Skris 992455714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 992555714Skris obj_dat.pl. It considered a zero in an object definition to mean 992655714Skris "end of object": none of the objects in objects.h have any zeros 992755714Skris so it wasn't spotted. 992855714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 992955714Skris 993055714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 993155714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 993255714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 993355714Skris vectors if you have them. 993455714Skris [Ben Laurie] 993555714Skris 993655714Skris *) Correct calculation of key length for export ciphers (too much space was 993755714Skris allocated for null ciphers). This has not been tested! 993855714Skris [Ben Laurie] 993955714Skris 994055714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 994155714Skris message is now correct (it understands "crypto" and "ssl" on its 994255714Skris command line). There is also now an "update" option. This will update 994355714Skris the util/ssleay.num and util/libeay.num files with any new functions. 994455714Skris If you do a: 994555714Skris perl util/mkdef.pl crypto ssl update 994655714Skris it will update them. 994755714Skris [Steve Henson] 994855714Skris 994955714Skris *) Overhauled the Perl interface (perl/*): 995055714Skris - ported BN stuff to OpenSSL's different BN library 995155714Skris - made the perl/ source tree CVS-aware 995255714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 995355714Skris their history because I've copied them in the repository) 995455714Skris - removed obsolete files (the test scripts will be replaced 995555714Skris by better Test::Harness variants in the future) 995655714Skris [Ralf S. Engelschall] 995755714Skris 995855714Skris *) First cut for a very conservative source tree cleanup: 995955714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 996055714Skris where we collect the old documents and readme texts. 996155714Skris 2. remove the first part of files where I'm already sure that we no 996255714Skris longer need them because of three reasons: either they are just temporary 996355714Skris files which were left by Eric or they are preserved original files where 996455714Skris I've verified that the diff is also available in the CVS via "cvs diff 996555714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 996655714Skris the crypto/md/ stuff). 996755714Skris [Ralf S. Engelschall] 996855714Skris 996955714Skris *) More extension code. Incomplete support for subject and issuer alt 997055714Skris name, issuer and authority key id. Change the i2v function parameters 997155714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 997255714Skris what that's for :-) Fix to ASN1 macro which messed up 997355714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 997455714Skris [Steve Henson] 997555714Skris 997655714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 997755714Skris INTEGER code. 997855714Skris [Steve Henson] 997955714Skris 998055714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 998155714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 998255714Skris 998355714Skris *) Make sure `make rehash' target really finds the `openssl' program. 998455714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 998555714Skris 998655714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 998755714Skris like to hear about it if this slows down other processors. 998855714Skris [Ben Laurie] 998955714Skris 999055714Skris *) Add CygWin32 platform information to Configure script. 999155714Skris [Alan Batie <batie@aahz.jf.intel.com>] 999255714Skris 999355714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 999455714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 999555714Skris 999655714Skris *) New program nseq to manipulate netscape certificate sequences 999755714Skris [Steve Henson] 999855714Skris 999955714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 1000055714Skris few typos. 1000155714Skris [Steve Henson] 1000255714Skris 1000355714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 1000455714Skris but the BN code had some problems that would cause failures when 1000555714Skris doing certificate verification and some other functions. 1000655714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 1000755714Skris 1000855714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 1000955714Skris [Steve Henson] 1001055714Skris 1001155714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 1001255714Skris [Steve Henson] 1001355714Skris 1001455714Skris *) Add several PKIX and private extended key usage OIDs. 1001555714Skris [Steve Henson] 1001655714Skris 1001755714Skris *) Modify the 'ca' program to handle the new extension code. Modify 1001855714Skris openssl.cnf for new extension format, add comments. 1001955714Skris [Steve Henson] 1002055714Skris 1002155714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 1002255714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 1002355714Skris CA extensions. 1002455714Skris [Steve Henson] 1002555714Skris 1002655714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 1002755714Skris error code, add initial support to X509_print() and x509 application. 1002855714Skris [Steve Henson] 1002955714Skris 1003055714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 1003155714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 1003255714Skris stuff is currently isolated and isn't even compiled yet. 1003355714Skris [Steve Henson] 1003455714Skris 1003555714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 1003655714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 1003755714Skris Removed the versions check from X509 routines when loading extensions: 1003855714Skris this allows certain broken certificates that don't set the version 1003955714Skris properly to be processed. 1004055714Skris [Steve Henson] 1004155714Skris 1004255714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 1004355714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 1004455714Skris can still be regenerated with "make depend". 1004555714Skris [Ben Laurie] 1004655714Skris 1004755714Skris *) Spelling mistake in C version of CAST-128. 1004855714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 1004955714Skris 1005055714Skris *) Changes to the error generation code. The perl script err-code.pl 1005155714Skris now reads in the old error codes and retains the old numbers, only 1005255714Skris adding new ones if necessary. It also only changes the .err files if new 1005355714Skris codes are added. The makefiles have been modified to only insert errors 1005455714Skris when needed (to avoid needlessly modifying header files). This is done 1005555714Skris by only inserting errors if the .err file is newer than the auto generated 1005655714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 1005755714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 1005855714Skris or delete all the .err files. 1005955714Skris [Steve Henson] 1006055714Skris 1006155714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 1006255714Skris been fixed, but is untested. The assembler versions are also fixed, but 1006355714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 1006455714Skris to regenerate it if needed. 1006555714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 1006655714Skris Hagino <itojun@kame.net>] 1006755714Skris 1006855714Skris *) File was opened incorrectly in randfile.c. 1006955714Skris [Ulf M�ller <ulf@fitug.de>] 1007055714Skris 1007155714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 1007255714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 1007355714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 1007455714Skris al: it's just almost always a UTCTime. Note this patch adds new error 1007555714Skris codes so do a "make errors" if there are problems. 1007655714Skris [Steve Henson] 1007755714Skris 1007855714Skris *) Correct Linux 1 recognition in config. 1007955714Skris [Ulf M�ller <ulf@fitug.de>] 1008055714Skris 1008155714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 1008255714Skris [Anonymous <nobody@replay.com>] 1008355714Skris 1008455714Skris *) Generate an error if given an empty string as a cert directory. Also 1008555714Skris generate an error if handed NULL (previously returned 0 to indicate an 1008655714Skris error, but didn't set one). 1008755714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 1008855714Skris 1008955714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 1009055714Skris [Ben Laurie] 1009155714Skris 1009255714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 1009355714Skris parameters. This was causing a warning which killed off the Win32 compile. 1009455714Skris [Steve Henson] 1009555714Skris 1009655714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 1009755714Skris [Neil Costigan <neil.costigan@celocom.com>] 1009855714Skris 1009955714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 1010055714Skris based on a text string, looking up short and long names and finally 1010155714Skris "dot" format. The "dot" format stuff didn't work. Added new function 1010255714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 1010355714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 1010455714Skris OID is not part of the table. 1010555714Skris [Steve Henson] 1010655714Skris 1010755714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 1010855714Skris X509_LOOKUP_by_alias(). 1010955714Skris [Ben Laurie] 1011055714Skris 1011155714Skris *) Sort openssl functions by name. 1011255714Skris [Ben Laurie] 1011355714Skris 1011455714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 1011555714Skris encryption from sample DSA keys (in case anyone is interested the password 1011655714Skris was "1234"). 1011755714Skris [Steve Henson] 1011855714Skris 1011955714Skris *) Make _all_ *_free functions accept a NULL pointer. 1012055714Skris [Frans Heymans <fheymans@isaserver.be>] 1012155714Skris 1012255714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 1012355714Skris NULL pointers. 1012455714Skris [Anonymous <nobody@replay.com>] 1012555714Skris 1012655714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 1012755714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 1012855714Skris 1012955714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 1013055714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 1013155714Skris 1013255714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 1013355714Skris [Anonymous <nobody@replay.com>] 1013455714Skris 1013555714Skris *) Add prototype for temp key callback functions 1013655714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 1013755714Skris [Ben Laurie] 1013855714Skris 1013955714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 1014055714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 1014155714Skris [Steve Henson] 1014255714Skris 1014355714Skris *) X509_name_add_entry() freed the wrong thing after an error. 1014455714Skris [Arne Ansper <arne@ats.cyber.ee>] 1014555714Skris 1014655714Skris *) rsa_eay.c would attempt to free a NULL context. 1014755714Skris [Arne Ansper <arne@ats.cyber.ee>] 1014855714Skris 1014955714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 1015055714Skris [Arne Ansper <arne@ats.cyber.ee>] 1015155714Skris 1015255714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 1015355714Skris [Arne Ansper <arne@ats.cyber.ee>] 1015455714Skris 1015555714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 1015655714Skris in X509_STORE_new(), but document the fact that this variable is still 1015755714Skris unused in the certificate verification process. 1015855714Skris [Ralf S. Engelschall] 1015955714Skris 1016055714Skris *) Fix the various library and apps files to free up pkeys obtained from 1016155714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 1016255714Skris [Steve Henson] 1016355714Skris 1016455714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 1016555714Skris demos/maurice/example2.c work, amongst others, probably. 1016655714Skris [Steve Henson and Ben Laurie] 1016755714Skris 1016855714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 1016955714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 1017055714Skris are no longer created. This way we have a single and consistent command 1017155714Skris line interface `openssl <command>', similar to `cvs <command>'. 1017255714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 1017355714Skris 1017455714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 1017555714Skris BIT STRING wrapper always have zero unused bits. 1017655714Skris [Steve Henson] 1017755714Skris 1017855714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 1017955714Skris [Steve Henson] 1018055714Skris 1018155714Skris *) Make the top-level INSTALL documentation easier to understand. 1018255714Skris [Paul Sutton] 1018355714Skris 1018455714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 1018555714Skris make (including if user presses ^C) [Paul Sutton] 1018655714Skris 1018755714Skris *) Make Montgomery context stuff explicit in RSA data structure. 1018855714Skris [Ben Laurie] 1018955714Skris 1019055714Skris *) Fix build order of pem and err to allow for generated pem.h. 1019155714Skris [Ben Laurie] 1019255714Skris 1019355714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 1019455714Skris [Ben Laurie] 1019555714Skris 1019655714Skris *) Enhanced the err-ins.pl script so it makes the error library number 1019755714Skris global and can add a library name. This is needed for external ASN1 and 1019855714Skris other error libraries. 1019955714Skris [Steve Henson] 1020055714Skris 1020155714Skris *) Fixed sk_insert which never worked properly. 1020255714Skris [Steve Henson] 1020355714Skris 1020455714Skris *) Fix ASN1 macros so they can handle indefinite length construted 1020555714Skris EXPLICIT tags. Some non standard certificates use these: they can now 1020655714Skris be read in. 1020755714Skris [Steve Henson] 1020855714Skris 1020955714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 1021055714Skris into a single doc/ssleay.txt bundle. This way the information is still 1021155714Skris preserved but no longer messes up this directory. Now it's new room for 1021255714Skris the new set of documenation files. 1021355714Skris [Ralf S. Engelschall] 1021455714Skris 1021555714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 1021655714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 1021755714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 1021855714Skris number of arguments. 1021955714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 1022055714Skris 1022155714Skris *) Fix test data to work with the above. 1022255714Skris [Ben Laurie] 1022355714Skris 1022455714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 1022555714Skris was already fixed by Eric for 0.9.1 it seems. 1022655714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 1022755714Skris 1022855714Skris *) Autodetect FreeBSD3. 1022955714Skris [Ben Laurie] 1023055714Skris 1023155714Skris *) Fix various bugs in Configure. This affects the following platforms: 1023255714Skris nextstep 1023355714Skris ncr-scde 1023455714Skris unixware-2.0 1023555714Skris unixware-2.0-pentium 1023655714Skris sco5-cc. 1023755714Skris [Ben Laurie] 1023855714Skris 1023955714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 1024055714Skris before they are needed. 1024155714Skris [Ben Laurie] 1024255714Skris 1024355714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 1024455714Skris [Ben Laurie] 1024555714Skris 1024655714Skris 1024755714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 1024855714Skris 1024955714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 1025055714Skris changed SSLeay to OpenSSL in version strings. 1025155714Skris [Ralf S. Engelschall] 1025255714Skris 1025355714Skris *) Some fixups to the top-level documents. 1025455714Skris [Paul Sutton] 1025555714Skris 1025655714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 1025755714Skris because the symlink to include/ was missing. 1025855714Skris [Ralf S. Engelschall] 1025955714Skris 1026055714Skris *) Incorporated the popular no-RSA/DSA-only patches 1026155714Skris which allow to compile a RSA-free SSLeay. 1026255714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 1026355714Skris 1026455714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 1026555714Skris when "ssleay" is still not found. 1026655714Skris [Ralf S. Engelschall] 1026755714Skris 1026855714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 1026955714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 1027055714Skris 1027155714Skris *) Updated the README file. 1027255714Skris [Ralf S. Engelschall] 1027355714Skris 1027455714Skris *) Added various .cvsignore files in the CVS repository subdirs 1027555714Skris to make a "cvs update" really silent. 1027655714Skris [Ralf S. Engelschall] 1027755714Skris 1027855714Skris *) Recompiled the error-definition header files and added 1027955714Skris missing symbols to the Win32 linker tables. 1028055714Skris [Ralf S. Engelschall] 1028155714Skris 1028255714Skris *) Cleaned up the top-level documents; 1028355714Skris o new files: CHANGES and LICENSE 1028455714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 1028555714Skris o merged COPYRIGHT into LICENSE 1028655714Skris o removed obsolete TODO file 1028755714Skris o renamed MICROSOFT to INSTALL.W32 1028855714Skris [Ralf S. Engelschall] 1028955714Skris 1029055714Skris *) Removed dummy files from the 0.9.1b source tree: 1029155714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 1029255714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 1029355714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 1029455714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 1029555714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 1029655714Skris [Ralf S. Engelschall] 1029755714Skris 1029855714Skris *) Added various platform portability fixes. 1029955714Skris [Mark J. Cox] 1030055714Skris 1030155714Skris *) The Genesis of the OpenSSL rpject: 1030255714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 1030355714Skris Young and Tim J. Hudson created while they were working for C2Net until 1030455714Skris summer 1998. 1030555714Skris [The OpenSSL Project] 1030655714Skris 1030755714Skris 1030855714Skris Changes between 0.9.0b and 0.9.1b [not released] 1030955714Skris 1031055714Skris *) Updated a few CA certificates under certs/ 1031155714Skris [Eric A. Young] 1031255714Skris 1031355714Skris *) Changed some BIGNUM api stuff. 1031455714Skris [Eric A. Young] 1031555714Skris 1031655714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 1031755714Skris DGUX x86, Linux Alpha, etc. 1031855714Skris [Eric A. Young] 1031955714Skris 1032055714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 1032155714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 1032255714Skris available). 1032355714Skris [Eric A. Young] 1032455714Skris 1032555714Skris *) Add -strparse option to asn1pars program which parses nested 1032655714Skris binary structures 1032755714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 1032855714Skris 1032955714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 1033055714Skris [Eric A. Young] 1033155714Skris 1033255714Skris *) DSA fix for "ca" program. 1033355714Skris [Eric A. Young] 1033455714Skris 1033555714Skris *) Added "-genkey" option to "dsaparam" program. 1033655714Skris [Eric A. Young] 1033755714Skris 1033855714Skris *) Added RIPE MD160 (rmd160) message digest. 1033955714Skris [Eric A. Young] 1034055714Skris 1034155714Skris *) Added -a (all) option to "ssleay version" command. 1034255714Skris [Eric A. Young] 1034355714Skris 1034455714Skris *) Added PLATFORM define which is the id given to Configure. 1034555714Skris [Eric A. Young] 1034655714Skris 1034755714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 1034855714Skris [Eric A. Young] 1034955714Skris 1035055714Skris *) Extended the ASN.1 parser routines. 1035155714Skris [Eric A. Young] 1035255714Skris 1035355714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 1035455714Skris [Eric A. Young] 1035555714Skris 1035655714Skris *) Added a BN_CTX to the BN library. 1035755714Skris [Eric A. Young] 1035855714Skris 1035955714Skris *) Fixed the weak key values in DES library 1036055714Skris [Eric A. Young] 1036155714Skris 1036255714Skris *) Changed API in EVP library for cipher aliases. 1036355714Skris [Eric A. Young] 1036455714Skris 1036555714Skris *) Added support for RC2/64bit cipher. 1036655714Skris [Eric A. Young] 1036755714Skris 1036855714Skris *) Converted the lhash library to the crypto/mem.c functions. 1036955714Skris [Eric A. Young] 1037055714Skris 1037155714Skris *) Added more recognized ASN.1 object ids. 1037255714Skris [Eric A. Young] 1037355714Skris 1037455714Skris *) Added more RSA padding checks for SSL/TLS. 1037555714Skris [Eric A. Young] 1037655714Skris 1037755714Skris *) Added BIO proxy/filter functionality. 1037855714Skris [Eric A. Young] 1037955714Skris 1038055714Skris *) Added extra_certs to SSL_CTX which can be used 1038155714Skris send extra CA certificates to the client in the CA cert chain sending 1038255714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 1038355714Skris [Eric A. Young] 1038455714Skris 1038555714Skris *) Now Fortezza is denied in the authentication phase because 1038655714Skris this is key exchange mechanism is not supported by SSLeay at all. 1038755714Skris [Eric A. Young] 1038855714Skris 1038955714Skris *) Additional PKCS1 checks. 1039055714Skris [Eric A. Young] 1039155714Skris 1039255714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 1039355714Skris [Eric A. Young] 1039455714Skris 1039555714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 1039655714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 1039755714Skris [Eric A. Young] 1039855714Skris 1039955714Skris *) Fixed a few memory leaks. 1040055714Skris [Eric A. Young] 1040155714Skris 1040255714Skris *) Fixed various code and comment typos. 1040355714Skris [Eric A. Young] 1040455714Skris 1040555714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 1040655714Skris bytes sent in the client random. 1040755714Skris [Edward Bishop <ebishop@spyglass.com>] 1040855714Skris 10409