CHANGES revision 276864
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 5276864Sjkim Changes between 1.0.1j and 1.0.1k [8 Jan 2015] 6276864Sjkim 7276864Sjkim *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS 8276864Sjkim message can cause a segmentation fault in OpenSSL due to a NULL pointer 9276864Sjkim dereference. This could lead to a Denial Of Service attack. Thanks to 10276864Sjkim Markus Stenberg of Cisco Systems, Inc. for reporting this issue. 11276864Sjkim (CVE-2014-3571) 12276864Sjkim [Steve Henson] 13276864Sjkim 14276864Sjkim *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the 15276864Sjkim dtls1_buffer_record function under certain conditions. In particular this 16276864Sjkim could occur if an attacker sent repeated DTLS records with the same 17276864Sjkim sequence number but for the next epoch. The memory leak could be exploited 18276864Sjkim by an attacker in a Denial of Service attack through memory exhaustion. 19276864Sjkim Thanks to Chris Mueller for reporting this issue. 20276864Sjkim (CVE-2015-0206) 21276864Sjkim [Matt Caswell] 22276864Sjkim 23276864Sjkim *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is 24276864Sjkim built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl 25276864Sjkim method would be set to NULL which could later result in a NULL pointer 26276864Sjkim dereference. Thanks to Frank Schmirler for reporting this issue. 27276864Sjkim (CVE-2014-3569) 28276864Sjkim [Kurt Roeckx] 29276864Sjkim 30276864Sjkim *) Abort handshake if server key exchange message is omitted for ephemeral 31276864Sjkim ECDH ciphersuites. 32276864Sjkim 33276864Sjkim Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for 34276864Sjkim reporting this issue. 35276864Sjkim (CVE-2014-3572) 36276864Sjkim [Steve Henson] 37276864Sjkim 38276864Sjkim *) Remove non-export ephemeral RSA code on client and server. This code 39276864Sjkim violated the TLS standard by allowing the use of temporary RSA keys in 40276864Sjkim non-export ciphersuites and could be used by a server to effectively 41276864Sjkim downgrade the RSA key length used to a value smaller than the server 42276864Sjkim certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at 43276864Sjkim INRIA or reporting this issue. 44276864Sjkim (CVE-2015-0204) 45276864Sjkim [Steve Henson] 46276864Sjkim 47276864Sjkim *) Fixed issue where DH client certificates are accepted without verification. 48276864Sjkim An OpenSSL server will accept a DH certificate for client authentication 49276864Sjkim without the certificate verify message. This effectively allows a client to 50276864Sjkim authenticate without the use of a private key. This only affects servers 51276864Sjkim which trust a client certificate authority which issues certificates 52276864Sjkim containing DH keys: these are extremely rare and hardly ever encountered. 53276864Sjkim Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting 54276864Sjkim this issue. 55276864Sjkim (CVE-2015-0205) 56276864Sjkim [Steve Henson] 57276864Sjkim 58276864Sjkim *) Ensure that the session ID context of an SSL is updated when its 59276864Sjkim SSL_CTX is updated via SSL_set_SSL_CTX. 60276864Sjkim 61276864Sjkim The session ID context is typically set from the parent SSL_CTX, 62276864Sjkim and can vary with the CTX. 63276864Sjkim [Adam Langley] 64276864Sjkim 65276864Sjkim *) Fix various certificate fingerprint issues. 66276864Sjkim 67276864Sjkim By using non-DER or invalid encodings outside the signed portion of a 68276864Sjkim certificate the fingerprint can be changed without breaking the signature. 69276864Sjkim Although no details of the signed portion of the certificate can be changed 70276864Sjkim this can cause problems with some applications: e.g. those using the 71276864Sjkim certificate fingerprint for blacklists. 72276864Sjkim 73276864Sjkim 1. Reject signatures with non zero unused bits. 74276864Sjkim 75276864Sjkim If the BIT STRING containing the signature has non zero unused bits reject 76276864Sjkim the signature. All current signature algorithms require zero unused bits. 77276864Sjkim 78276864Sjkim 2. Check certificate algorithm consistency. 79276864Sjkim 80276864Sjkim Check the AlgorithmIdentifier inside TBS matches the one in the 81276864Sjkim certificate signature. NB: this will result in signature failure 82276864Sjkim errors for some broken certificates. 83276864Sjkim 84276864Sjkim Thanks to Konrad Kraszewski from Google for reporting this issue. 85276864Sjkim 86276864Sjkim 3. Check DSA/ECDSA signatures use DER. 87276864Sjkim 88276864Sjkim Reencode DSA/ECDSA signatures and compare with the original received 89276864Sjkim signature. Return an error if there is a mismatch. 90276864Sjkim 91276864Sjkim This will reject various cases including garbage after signature 92276864Sjkim (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS 93276864Sjkim program for discovering this case) and use of BER or invalid ASN.1 INTEGERs 94276864Sjkim (negative or with leading zeroes). 95276864Sjkim 96276864Sjkim Further analysis was conducted and fixes were developed by Stephen Henson 97276864Sjkim of the OpenSSL core team. 98276864Sjkim 99276864Sjkim (CVE-2014-8275) 100276864Sjkim [Steve Henson] 101276864Sjkim 102276864Sjkim *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect 103276864Sjkim results on some platforms, including x86_64. This bug occurs at random 104276864Sjkim with a very low probability, and is not known to be exploitable in any 105276864Sjkim way, though its exact impact is difficult to determine. Thanks to Pieter 106276864Sjkim Wuille (Blockstream) who reported this issue and also suggested an initial 107276864Sjkim fix. Further analysis was conducted by the OpenSSL development team and 108276864Sjkim Adam Langley of Google. The final fix was developed by Andy Polyakov of 109276864Sjkim the OpenSSL core team. 110276864Sjkim (CVE-2014-3570) 111276864Sjkim [Andy Polyakov] 112276864Sjkim 113276864Sjkim *) Do not resume sessions on the server if the negotiated protocol 114276864Sjkim version does not match the session's version. Resuming with a different 115276864Sjkim version, while not strictly forbidden by the RFC, is of questionable 116276864Sjkim sanity and breaks all known clients. 117276864Sjkim [David Benjamin, Emilia K�sper] 118276864Sjkim 119276864Sjkim *) Tighten handling of the ChangeCipherSpec (CCS) message: reject 120276864Sjkim early CCS messages during renegotiation. (Note that because 121276864Sjkim renegotiation is encrypted, this early CCS was not exploitable.) 122276864Sjkim [Emilia K�sper] 123276864Sjkim 124276864Sjkim *) Tighten client-side session ticket handling during renegotiation: 125276864Sjkim ensure that the client only accepts a session ticket if the server sends 126276864Sjkim the extension anew in the ServerHello. Previously, a TLS client would 127276864Sjkim reuse the old extension state and thus accept a session ticket if one was 128276864Sjkim announced in the initial ServerHello. 129276864Sjkim 130276864Sjkim Similarly, ensure that the client requires a session ticket if one 131276864Sjkim was advertised in the ServerHello. Previously, a TLS client would 132276864Sjkim ignore a missing NewSessionTicket message. 133276864Sjkim [Emilia K�sper] 134276864Sjkim 135273149Sjkim Changes between 1.0.1i and 1.0.1j [15 Oct 2014] 136273149Sjkim 137273149Sjkim *) SRTP Memory Leak. 138273149Sjkim 139273149Sjkim A flaw in the DTLS SRTP extension parsing code allows an attacker, who 140273149Sjkim sends a carefully crafted handshake message, to cause OpenSSL to fail 141273149Sjkim to free up to 64k of memory causing a memory leak. This could be 142273149Sjkim exploited in a Denial Of Service attack. This issue affects OpenSSL 143273149Sjkim 1.0.1 server implementations for both SSL/TLS and DTLS regardless of 144273149Sjkim whether SRTP is used or configured. Implementations of OpenSSL that 145273149Sjkim have been compiled with OPENSSL_NO_SRTP defined are not affected. 146273149Sjkim 147273149Sjkim The fix was developed by the OpenSSL team. 148273149Sjkim (CVE-2014-3513) 149273149Sjkim [OpenSSL team] 150273149Sjkim 151273149Sjkim *) Session Ticket Memory Leak. 152273149Sjkim 153273149Sjkim When an OpenSSL SSL/TLS/DTLS server receives a session ticket the 154273149Sjkim integrity of that ticket is first verified. In the event of a session 155273149Sjkim ticket integrity check failing, OpenSSL will fail to free memory 156273149Sjkim causing a memory leak. By sending a large number of invalid session 157273149Sjkim tickets an attacker could exploit this issue in a Denial Of Service 158273149Sjkim attack. 159273149Sjkim (CVE-2014-3567) 160273149Sjkim [Steve Henson] 161273149Sjkim 162273149Sjkim *) Build option no-ssl3 is incomplete. 163273149Sjkim 164273149Sjkim When OpenSSL is configured with "no-ssl3" as a build option, servers 165273149Sjkim could accept and complete a SSL 3.0 handshake, and clients could be 166273149Sjkim configured to send them. 167273149Sjkim (CVE-2014-3568) 168273149Sjkim [Akamai and the OpenSSL team] 169273149Sjkim 170273149Sjkim *) Add support for TLS_FALLBACK_SCSV. 171273149Sjkim Client applications doing fallback retries should call 172273149Sjkim SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV). 173273149Sjkim (CVE-2014-3566) 174273149Sjkim [Adam Langley, Bodo Moeller] 175273149Sjkim 176273149Sjkim *) Add additional DigestInfo checks. 177273149Sjkim 178273149Sjkim Reencode DigestInto in DER and check against the original when 179273149Sjkim verifying RSA signature: this will reject any improperly encoded 180273149Sjkim DigestInfo structures. 181273149Sjkim 182273149Sjkim Note: this is a precautionary measure and no attacks are currently known. 183273149Sjkim 184273149Sjkim [Steve Henson] 185273149Sjkim 186269686Sjkim Changes between 1.0.1h and 1.0.1i [6 Aug 2014] 187269686Sjkim 188269686Sjkim *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the 189269686Sjkim SRP code can be overrun an internal buffer. Add sanity check that 190269686Sjkim g, A, B < N to SRP code. 191269686Sjkim 192269686Sjkim Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC 193269686Sjkim Group for discovering this issue. 194269686Sjkim (CVE-2014-3512) 195269686Sjkim [Steve Henson] 196269686Sjkim 197269686Sjkim *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate 198269686Sjkim TLS 1.0 instead of higher protocol versions when the ClientHello message 199269686Sjkim is badly fragmented. This allows a man-in-the-middle attacker to force a 200269686Sjkim downgrade to TLS 1.0 even if both the server and the client support a 201269686Sjkim higher protocol version, by modifying the client's TLS records. 202269686Sjkim 203269686Sjkim Thanks to David Benjamin and Adam Langley (Google) for discovering and 204269686Sjkim researching this issue. 205269686Sjkim (CVE-2014-3511) 206269686Sjkim [David Benjamin] 207269686Sjkim 208269686Sjkim *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject 209269686Sjkim to a denial of service attack. A malicious server can crash the client 210269686Sjkim with a null pointer dereference (read) by specifying an anonymous (EC)DH 211269686Sjkim ciphersuite and sending carefully crafted handshake messages. 212269686Sjkim 213269686Sjkim Thanks to Felix Gr�bert (Google) for discovering and researching this 214269686Sjkim issue. 215269686Sjkim (CVE-2014-3510) 216269686Sjkim [Emilia K�sper] 217269686Sjkim 218269686Sjkim *) By sending carefully crafted DTLS packets an attacker could cause openssl 219269686Sjkim to leak memory. This can be exploited through a Denial of Service attack. 220269686Sjkim Thanks to Adam Langley for discovering and researching this issue. 221269686Sjkim (CVE-2014-3507) 222269686Sjkim [Adam Langley] 223269686Sjkim 224269686Sjkim *) An attacker can force openssl to consume large amounts of memory whilst 225269686Sjkim processing DTLS handshake messages. This can be exploited through a 226269686Sjkim Denial of Service attack. 227269686Sjkim Thanks to Adam Langley for discovering and researching this issue. 228269686Sjkim (CVE-2014-3506) 229269686Sjkim [Adam Langley] 230269686Sjkim 231269686Sjkim *) An attacker can force an error condition which causes openssl to crash 232269686Sjkim whilst processing DTLS packets due to memory being freed twice. This 233269686Sjkim can be exploited through a Denial of Service attack. 234269686Sjkim Thanks to Adam Langley and Wan-Teh Chang for discovering and researching 235269686Sjkim this issue. 236269686Sjkim (CVE-2014-3505) 237269686Sjkim [Adam Langley] 238269686Sjkim 239269686Sjkim *) If a multithreaded client connects to a malicious server using a resumed 240269686Sjkim session and the server sends an ec point format extension it could write 241269686Sjkim up to 255 bytes to freed memory. 242269686Sjkim 243269686Sjkim Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this 244269686Sjkim issue. 245269686Sjkim (CVE-2014-3509) 246269686Sjkim [Gabor Tyukasz] 247269686Sjkim 248269686Sjkim *) A malicious server can crash an OpenSSL client with a null pointer 249269686Sjkim dereference (read) by specifying an SRP ciphersuite even though it was not 250269686Sjkim properly negotiated with the client. This can be exploited through a 251269686Sjkim Denial of Service attack. 252269686Sjkim 253269686Sjkim Thanks to Joonas Kuorilehto and Riku Hietam�ki (Codenomicon) for 254269686Sjkim discovering and researching this issue. 255269686Sjkim (CVE-2014-5139) 256269686Sjkim [Steve Henson] 257269686Sjkim 258269686Sjkim *) A flaw in OBJ_obj2txt may cause pretty printing functions such as 259269686Sjkim X509_name_oneline, X509_name_print_ex et al. to leak some information 260269686Sjkim from the stack. Applications may be affected if they echo pretty printing 261269686Sjkim output to the attacker. 262269686Sjkim 263269686Sjkim Thanks to Ivan Fratric (Google) for discovering this issue. 264269686Sjkim (CVE-2014-3508) 265269686Sjkim [Emilia K�sper, and Steve Henson] 266269686Sjkim 267269686Sjkim *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) 268269686Sjkim for corner cases. (Certain input points at infinity could lead to 269269686Sjkim bogus results, with non-infinity inputs mapped to infinity too.) 270269686Sjkim [Bodo Moeller] 271269686Sjkim 272267258Sjkim Changes between 1.0.1g and 1.0.1h [5 Jun 2014] 273267258Sjkim 274267258Sjkim *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted 275267258Sjkim handshake can force the use of weak keying material in OpenSSL 276267258Sjkim SSL/TLS clients and servers. 277267258Sjkim 278267258Sjkim Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and 279267258Sjkim researching this issue. (CVE-2014-0224) 280267258Sjkim [KIKUCHI Masashi, Steve Henson] 281267258Sjkim 282267258Sjkim *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an 283267258Sjkim OpenSSL DTLS client the code can be made to recurse eventually crashing 284267258Sjkim in a DoS attack. 285267258Sjkim 286267258Sjkim Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue. 287267258Sjkim (CVE-2014-0221) 288267258Sjkim [Imre Rad, Steve Henson] 289267258Sjkim 290267258Sjkim *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can 291267258Sjkim be triggered by sending invalid DTLS fragments to an OpenSSL DTLS 292267258Sjkim client or server. This is potentially exploitable to run arbitrary 293267258Sjkim code on a vulnerable client or server. 294267258Sjkim 295267258Sjkim Thanks to J�ri Aedla for reporting this issue. (CVE-2014-0195) 296267258Sjkim [J�ri Aedla, Steve Henson] 297267258Sjkim 298267258Sjkim *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites 299267258Sjkim are subject to a denial of service attack. 300267258Sjkim 301267258Sjkim Thanks to Felix Gr�bert and Ivan Fratric at Google for discovering 302267258Sjkim this issue. (CVE-2014-3470) 303267258Sjkim [Felix Gr�bert, Ivan Fratric, Steve Henson] 304267258Sjkim 305267258Sjkim *) Harmonize version and its documentation. -f flag is used to display 306267258Sjkim compilation flags. 307267258Sjkim [mancha <mancha1@zoho.com>] 308267258Sjkim 309267258Sjkim *) Fix eckey_priv_encode so it immediately returns an error upon a failure 310267258Sjkim in i2d_ECPrivateKey. 311267258Sjkim [mancha <mancha1@zoho.com>] 312267258Sjkim 313267258Sjkim *) Fix some double frees. These are not thought to be exploitable. 314267258Sjkim [mancha <mancha1@zoho.com>] 315267258Sjkim 316264331Sjkim Changes between 1.0.1f and 1.0.1g [7 Apr 2014] 317264331Sjkim 318264331Sjkim *) A missing bounds check in the handling of the TLS heartbeat extension 319264331Sjkim can be used to reveal up to 64k of memory to a connected client or 320264331Sjkim server. 321264331Sjkim 322264331Sjkim Thanks for Neel Mehta of Google Security for discovering this bug and to 323264331Sjkim Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for 324264331Sjkim preparing the fix (CVE-2014-0160) 325264331Sjkim [Adam Langley, Bodo Moeller] 326264331Sjkim 327264331Sjkim *) Fix for the attack described in the paper "Recovering OpenSSL 328264331Sjkim ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack" 329264331Sjkim by Yuval Yarom and Naomi Benger. Details can be obtained from: 330264331Sjkim http://eprint.iacr.org/2014/140 331264331Sjkim 332264331Sjkim Thanks to Yuval Yarom and Naomi Benger for discovering this 333264331Sjkim flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076) 334264331Sjkim [Yuval Yarom and Naomi Benger] 335264331Sjkim 336264331Sjkim *) TLS pad extension: draft-agl-tls-padding-03 337264331Sjkim 338264331Sjkim Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the 339264331Sjkim TLS client Hello record length value would otherwise be > 255 and 340264331Sjkim less that 512 pad with a dummy extension containing zeroes so it 341264331Sjkim is at least 512 bytes long. 342264331Sjkim 343264331Sjkim [Adam Langley, Steve Henson] 344264331Sjkim 345264331Sjkim Changes between 1.0.1e and 1.0.1f [6 Jan 2014] 346264331Sjkim 347264331Sjkim *) Fix for TLS record tampering bug. A carefully crafted invalid 348264331Sjkim handshake could crash OpenSSL with a NULL pointer exception. 349264331Sjkim Thanks to Anton Johansson for reporting this issues. 350264331Sjkim (CVE-2013-4353) 351264331Sjkim 352264331Sjkim *) Keep original DTLS digest and encryption contexts in retransmission 353264331Sjkim structures so we can use the previous session parameters if they need 354264331Sjkim to be resent. (CVE-2013-6450) 355264331Sjkim [Steve Henson] 356264331Sjkim 357264331Sjkim *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which 358264331Sjkim avoids preferring ECDHE-ECDSA ciphers when the client appears to be 359264331Sjkim Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for 360264331Sjkim several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug 361264331Sjkim is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing 362264331Sjkim 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer. 363264331Sjkim [Rob Stradling, Adam Langley] 364264331Sjkim 365246772Sjkim Changes between 1.0.1d and 1.0.1e [11 Feb 2013] 366246772Sjkim 367264331Sjkim *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI 368264331Sjkim supporting platforms or when small records were transferred. 369264331Sjkim [Andy Polyakov, Steve Henson] 370246772Sjkim 371246772Sjkim Changes between 1.0.1c and 1.0.1d [5 Feb 2013] 372246772Sjkim 373246772Sjkim *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 374246772Sjkim 375246772Sjkim This addresses the flaw in CBC record processing discovered by 376246772Sjkim Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 377246772Sjkim at: http://www.isg.rhul.ac.uk/tls/ 378246772Sjkim 379246772Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 380246772Sjkim Security Group at Royal Holloway, University of London 381246772Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 382246772Sjkim Emilia K�sper for the initial patch. 383246772Sjkim (CVE-2013-0169) 384246772Sjkim [Emilia K�sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 385246772Sjkim 386246772Sjkim *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode 387246772Sjkim ciphersuites which can be exploited in a denial of service attack. 388246772Sjkim Thanks go to and to Adam Langley <agl@chromium.org> for discovering 389246772Sjkim and detecting this bug and to Wolfgang Ettlinger 390246772Sjkim <wolfgang.ettlinger@gmail.com> for independently discovering this issue. 391246772Sjkim (CVE-2012-2686) 392246772Sjkim [Adam Langley] 393246772Sjkim 394246772Sjkim *) Return an error when checking OCSP signatures when key is NULL. 395246772Sjkim This fixes a DoS attack. (CVE-2013-0166) 396246772Sjkim [Steve Henson] 397246772Sjkim 398246772Sjkim *) Make openssl verify return errors. 399246772Sjkim [Chris Palmer <palmer@google.com> and Ben Laurie] 400246772Sjkim 401246772Sjkim *) Call OCSP Stapling callback after ciphersuite has been chosen, so 402246772Sjkim the right response is stapled. Also change SSL_get_certificate() 403246772Sjkim so it returns the certificate actually sent. 404246772Sjkim See http://rt.openssl.org/Ticket/Display.html?id=2836. 405246772Sjkim [Rob Stradling <rob.stradling@comodo.com>] 406246772Sjkim 407246772Sjkim *) Fix possible deadlock when decoding public keys. 408246772Sjkim [Steve Henson] 409246772Sjkim 410246772Sjkim *) Don't use TLS 1.0 record version number in initial client hello 411246772Sjkim if renegotiating. 412246772Sjkim [Steve Henson] 413246772Sjkim 414238405Sjkim Changes between 1.0.1b and 1.0.1c [10 May 2012] 415237657Sjkim 416238405Sjkim *) Sanity check record length before skipping explicit IV in TLS 417238405Sjkim 1.2, 1.1 and DTLS to fix DoS attack. 418237657Sjkim 419237657Sjkim Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 420237657Sjkim fuzzing as a service testing platform. 421237657Sjkim (CVE-2012-2333) 422237657Sjkim [Steve Henson] 423237657Sjkim 424237657Sjkim *) Initialise tkeylen properly when encrypting CMS messages. 425237657Sjkim Thanks to Solar Designer of Openwall for reporting this issue. 426237657Sjkim [Steve Henson] 427237657Sjkim 428238405Sjkim *) In FIPS mode don't try to use composite ciphers as they are not 429238405Sjkim approved. 430238405Sjkim [Steve Henson] 431237657Sjkim 432238405Sjkim Changes between 1.0.1a and 1.0.1b [26 Apr 2012] 433237657Sjkim 434238405Sjkim *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and 435238405Sjkim 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately 436238405Sjkim mean any application compiled against OpenSSL 1.0.0 headers setting 437238405Sjkim SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng 438238405Sjkim TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to 439238405Sjkim 0x10000000L Any application which was previously compiled against 440238405Sjkim OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1 441238405Sjkim will need to be recompiled as a result. Letting be results in 442238405Sjkim inability to disable specifically TLS 1.1 and in client context, 443238405Sjkim in unlike event, limit maximum offered version to TLS 1.0 [see below]. 444238405Sjkim [Steve Henson] 445237657Sjkim 446238405Sjkim *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not 447238405Sjkim disable just protocol X, but all protocols above X *if* there are 448238405Sjkim protocols *below* X still enabled. In more practical terms it means 449238405Sjkim that if application wants to disable TLS1.0 in favor of TLS1.1 and 450238405Sjkim above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass 451238405Sjkim SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to 452238405Sjkim client side. 453238405Sjkim [Andy Polyakov] 454238405Sjkim 455238405Sjkim Changes between 1.0.1 and 1.0.1a [19 Apr 2012] 456238405Sjkim 457237657Sjkim *) Check for potentially exploitable overflows in asn1_d2i_read_bio 458237657Sjkim BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 459237657Sjkim in CRYPTO_realloc_clean. 460237657Sjkim 461237657Sjkim Thanks to Tavis Ormandy, Google Security Team, for discovering this 462237657Sjkim issue and to Adam Langley <agl@chromium.org> for fixing it. 463237657Sjkim (CVE-2012-2110) 464237657Sjkim [Adam Langley (Google), Tavis Ormandy, Google Security Team] 465237657Sjkim 466238405Sjkim *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections. 467238405Sjkim [Adam Langley] 468237657Sjkim 469238405Sjkim *) Workarounds for some broken servers that "hang" if a client hello 470238405Sjkim record length exceeds 255 bytes. 471238405Sjkim 472238405Sjkim 1. Do not use record version number > TLS 1.0 in initial client 473238405Sjkim hello: some (but not all) hanging servers will now work. 474238405Sjkim 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate 475238405Sjkim the number of ciphers sent in the client hello. This should be 476238405Sjkim set to an even number, such as 50, for example by passing: 477238405Sjkim -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure. 478238405Sjkim Most broken servers should now work. 479238405Sjkim 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable 480238405Sjkim TLS 1.2 client support entirely. 481238405Sjkim [Steve Henson] 482238405Sjkim 483238405Sjkim *) Fix SEGV in Vector Permutation AES module observed in OpenSSH. 484238405Sjkim [Andy Polyakov] 485238405Sjkim 486238405Sjkim Changes between 1.0.0h and 1.0.1 [14 Mar 2012] 487238405Sjkim 488238405Sjkim *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET 489238405Sjkim STRING form instead of a DigestInfo. 490238405Sjkim [Steve Henson] 491238405Sjkim 492238405Sjkim *) The format used for MDC2 RSA signatures is inconsistent between EVP 493238405Sjkim and the RSA_sign/RSA_verify functions. This was made more apparent when 494238405Sjkim OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular 495238405Sjkim those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect 496238405Sjkim the correct format in RSA_verify so both forms transparently work. 497238405Sjkim [Steve Henson] 498238405Sjkim 499238405Sjkim *) Some servers which support TLS 1.0 can choke if we initially indicate 500238405Sjkim support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA 501238405Sjkim encrypted premaster secret. As a workaround use the maximum pemitted 502238405Sjkim client version in client hello, this should keep such servers happy 503238405Sjkim and still work with previous versions of OpenSSL. 504238405Sjkim [Steve Henson] 505238405Sjkim 506238405Sjkim *) Add support for TLS/DTLS heartbeats. 507238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 508238405Sjkim 509238405Sjkim *) Add support for SCTP. 510238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 511238405Sjkim 512238405Sjkim *) Improved PRNG seeding for VOS. 513238405Sjkim [Paul Green <Paul.Green@stratus.com>] 514238405Sjkim 515238405Sjkim *) Extensive assembler packs updates, most notably: 516238405Sjkim 517238405Sjkim - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support; 518238405Sjkim - x86[_64]: SSSE3 support (SHA1, vector-permutation AES); 519238405Sjkim - x86_64: bit-sliced AES implementation; 520238405Sjkim - ARM: NEON support, contemporary platforms optimizations; 521238405Sjkim - s390x: z196 support; 522238405Sjkim - *: GHASH and GF(2^m) multiplication implementations; 523238405Sjkim 524238405Sjkim [Andy Polyakov] 525238405Sjkim 526238405Sjkim *) Make TLS-SRP code conformant with RFC 5054 API cleanup 527238405Sjkim (removal of unnecessary code) 528238405Sjkim [Peter Sylvester <peter.sylvester@edelweb.fr>] 529238405Sjkim 530238405Sjkim *) Add TLS key material exporter from RFC 5705. 531238405Sjkim [Eric Rescorla] 532238405Sjkim 533238405Sjkim *) Add DTLS-SRTP negotiation from RFC 5764. 534238405Sjkim [Eric Rescorla] 535238405Sjkim 536238405Sjkim *) Add Next Protocol Negotiation, 537238405Sjkim http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be 538238405Sjkim disabled with a no-npn flag to config or Configure. Code donated 539238405Sjkim by Google. 540238405Sjkim [Adam Langley <agl@google.com> and Ben Laurie] 541238405Sjkim 542238405Sjkim *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224, 543238405Sjkim NIST-P256, NIST-P521, with constant-time single point multiplication on 544238405Sjkim typical inputs. Compiler support for the nonstandard type __uint128_t is 545238405Sjkim required to use this (present in gcc 4.4 and later, for 64-bit builds). 546238405Sjkim Code made available under Apache License version 2.0. 547238405Sjkim 548238405Sjkim Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command 549238405Sjkim line to include this in your build of OpenSSL, and run "make depend" (or 550238405Sjkim "make update"). This enables the following EC_METHODs: 551238405Sjkim 552238405Sjkim EC_GFp_nistp224_method() 553238405Sjkim EC_GFp_nistp256_method() 554238405Sjkim EC_GFp_nistp521_method() 555238405Sjkim 556238405Sjkim EC_GROUP_new_by_curve_name() will automatically use these (while 557238405Sjkim EC_GROUP_new_curve_GFp() currently prefers the more flexible 558238405Sjkim implementations). 559238405Sjkim [Emilia K�sper, Adam Langley, Bodo Moeller (Google)] 560238405Sjkim 561238405Sjkim *) Use type ossl_ssize_t instad of ssize_t which isn't available on 562238405Sjkim all platforms. Move ssize_t definition from e_os.h to the public 563238405Sjkim header file e_os2.h as it now appears in public header file cms.h 564238405Sjkim [Steve Henson] 565238405Sjkim 566238405Sjkim *) New -sigopt option to the ca, req and x509 utilities. Additional 567238405Sjkim signature parameters can be passed using this option and in 568238405Sjkim particular PSS. 569238405Sjkim [Steve Henson] 570238405Sjkim 571238405Sjkim *) Add RSA PSS signing function. This will generate and set the 572238405Sjkim appropriate AlgorithmIdentifiers for PSS based on those in the 573238405Sjkim corresponding EVP_MD_CTX structure. No application support yet. 574238405Sjkim [Steve Henson] 575238405Sjkim 576238405Sjkim *) Support for companion algorithm specific ASN1 signing routines. 577238405Sjkim New function ASN1_item_sign_ctx() signs a pre-initialised 578238405Sjkim EVP_MD_CTX structure and sets AlgorithmIdentifiers based on 579238405Sjkim the appropriate parameters. 580238405Sjkim [Steve Henson] 581238405Sjkim 582238405Sjkim *) Add new algorithm specific ASN1 verification initialisation function 583238405Sjkim to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1 584238405Sjkim handling will be the same no matter what EVP_PKEY_METHOD is used. 585238405Sjkim Add a PSS handler to support verification of PSS signatures: checked 586238405Sjkim against a number of sample certificates. 587238405Sjkim [Steve Henson] 588238405Sjkim 589238405Sjkim *) Add signature printing for PSS. Add PSS OIDs. 590238405Sjkim [Steve Henson, Martin Kaiser <lists@kaiser.cx>] 591238405Sjkim 592238405Sjkim *) Add algorithm specific signature printing. An individual ASN1 method 593238405Sjkim can now print out signatures instead of the standard hex dump. 594238405Sjkim 595238405Sjkim More complex signatures (e.g. PSS) can print out more meaningful 596238405Sjkim information. Include DSA version that prints out the signature 597238405Sjkim parameters r, s. 598238405Sjkim [Steve Henson] 599238405Sjkim 600238405Sjkim *) Password based recipient info support for CMS library: implementing 601238405Sjkim RFC3211. 602238405Sjkim [Steve Henson] 603238405Sjkim 604238405Sjkim *) Split password based encryption into PBES2 and PBKDF2 functions. This 605238405Sjkim neatly separates the code into cipher and PBE sections and is required 606238405Sjkim for some algorithms that split PBES2 into separate pieces (such as 607238405Sjkim password based CMS). 608238405Sjkim [Steve Henson] 609238405Sjkim 610238405Sjkim *) Session-handling fixes: 611238405Sjkim - Fix handling of connections that are resuming with a session ID, 612238405Sjkim but also support Session Tickets. 613238405Sjkim - Fix a bug that suppressed issuing of a new ticket if the client 614238405Sjkim presented a ticket with an expired session. 615238405Sjkim - Try to set the ticket lifetime hint to something reasonable. 616238405Sjkim - Make tickets shorter by excluding irrelevant information. 617238405Sjkim - On the client side, don't ignore renewed tickets. 618238405Sjkim [Adam Langley, Bodo Moeller (Google)] 619238405Sjkim 620238405Sjkim *) Fix PSK session representation. 621238405Sjkim [Bodo Moeller] 622238405Sjkim 623238405Sjkim *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations. 624238405Sjkim 625238405Sjkim This work was sponsored by Intel. 626238405Sjkim [Andy Polyakov] 627238405Sjkim 628238405Sjkim *) Add GCM support to TLS library. Some custom code is needed to split 629238405Sjkim the IV between the fixed (from PRF) and explicit (from TLS record) 630238405Sjkim portions. This adds all GCM ciphersuites supported by RFC5288 and 631238405Sjkim RFC5289. Generalise some AES* cipherstrings to inlclude GCM and 632238405Sjkim add a special AESGCM string for GCM only. 633238405Sjkim [Steve Henson] 634238405Sjkim 635238405Sjkim *) Expand range of ctrls for AES GCM. Permit setting invocation 636238405Sjkim field on decrypt and retrieval of invocation field only on encrypt. 637238405Sjkim [Steve Henson] 638238405Sjkim 639238405Sjkim *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support. 640238405Sjkim As required by RFC5289 these ciphersuites cannot be used if for 641238405Sjkim versions of TLS earlier than 1.2. 642238405Sjkim [Steve Henson] 643238405Sjkim 644238405Sjkim *) For FIPS capable OpenSSL interpret a NULL default public key method 645238405Sjkim as unset and return the appopriate default but do *not* set the default. 646238405Sjkim This means we can return the appopriate method in applications that 647238405Sjkim swicth between FIPS and non-FIPS modes. 648238405Sjkim [Steve Henson] 649238405Sjkim 650238405Sjkim *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an 651238405Sjkim ENGINE is used then we cannot handle that in the FIPS module so we 652238405Sjkim keep original code iff non-FIPS operations are allowed. 653238405Sjkim [Steve Henson] 654238405Sjkim 655238405Sjkim *) Add -attime option to openssl utilities. 656238405Sjkim [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson] 657238405Sjkim 658238405Sjkim *) Redirect DSA and DH operations to FIPS module in FIPS mode. 659238405Sjkim [Steve Henson] 660238405Sjkim 661238405Sjkim *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use 662238405Sjkim FIPS EC methods unconditionally for now. 663238405Sjkim [Steve Henson] 664238405Sjkim 665238405Sjkim *) New build option no-ec2m to disable characteristic 2 code. 666238405Sjkim [Steve Henson] 667238405Sjkim 668238405Sjkim *) Backport libcrypto audit of return value checking from 1.1.0-dev; not 669238405Sjkim all cases can be covered as some introduce binary incompatibilities. 670238405Sjkim [Steve Henson] 671238405Sjkim 672238405Sjkim *) Redirect RSA operations to FIPS module including keygen, 673238405Sjkim encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods. 674238405Sjkim [Steve Henson] 675238405Sjkim 676238405Sjkim *) Add similar low level API blocking to ciphers. 677238405Sjkim [Steve Henson] 678238405Sjkim 679238405Sjkim *) Low level digest APIs are not approved in FIPS mode: any attempt 680238405Sjkim to use these will cause a fatal error. Applications that *really* want 681238405Sjkim to use them can use the private_* version instead. 682238405Sjkim [Steve Henson] 683238405Sjkim 684238405Sjkim *) Redirect cipher operations to FIPS module for FIPS builds. 685238405Sjkim [Steve Henson] 686238405Sjkim 687238405Sjkim *) Redirect digest operations to FIPS module for FIPS builds. 688238405Sjkim [Steve Henson] 689238405Sjkim 690238405Sjkim *) Update build system to add "fips" flag which will link in fipscanister.o 691238405Sjkim for static and shared library builds embedding a signature if needed. 692238405Sjkim [Steve Henson] 693238405Sjkim 694238405Sjkim *) Output TLS supported curves in preference order instead of numerical 695238405Sjkim order. This is currently hardcoded for the highest order curves first. 696238405Sjkim This should be configurable so applications can judge speed vs strength. 697238405Sjkim [Steve Henson] 698238405Sjkim 699238405Sjkim *) Add TLS v1.2 server support for client authentication. 700238405Sjkim [Steve Henson] 701238405Sjkim 702238405Sjkim *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers 703238405Sjkim and enable MD5. 704238405Sjkim [Steve Henson] 705238405Sjkim 706238405Sjkim *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying 707238405Sjkim FIPS modules versions. 708238405Sjkim [Steve Henson] 709238405Sjkim 710238405Sjkim *) Add TLS v1.2 client side support for client authentication. Keep cache 711238405Sjkim of handshake records longer as we don't know the hash algorithm to use 712238405Sjkim until after the certificate request message is received. 713238405Sjkim [Steve Henson] 714238405Sjkim 715238405Sjkim *) Initial TLS v1.2 client support. Add a default signature algorithms 716238405Sjkim extension including all the algorithms we support. Parse new signature 717238405Sjkim format in client key exchange. Relax some ECC signing restrictions for 718238405Sjkim TLS v1.2 as indicated in RFC5246. 719238405Sjkim [Steve Henson] 720238405Sjkim 721238405Sjkim *) Add server support for TLS v1.2 signature algorithms extension. Switch 722238405Sjkim to new signature format when needed using client digest preference. 723238405Sjkim All server ciphersuites should now work correctly in TLS v1.2. No client 724238405Sjkim support yet and no support for client certificates. 725238405Sjkim [Steve Henson] 726238405Sjkim 727238405Sjkim *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch 728238405Sjkim to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based 729238405Sjkim ciphersuites. At present only RSA key exchange ciphersuites work with 730238405Sjkim TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete 731238405Sjkim SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods 732238405Sjkim and version checking. 733238405Sjkim [Steve Henson] 734238405Sjkim 735238405Sjkim *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled 736238405Sjkim with this defined it will not be affected by any changes to ssl internal 737238405Sjkim structures. Add several utility functions to allow openssl application 738238405Sjkim to work with OPENSSL_NO_SSL_INTERN defined. 739238405Sjkim [Steve Henson] 740238405Sjkim 741238405Sjkim *) Add SRP support. 742238405Sjkim [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie] 743238405Sjkim 744238405Sjkim *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id. 745238405Sjkim [Steve Henson] 746238405Sjkim 747238405Sjkim *) Permit abbreviated handshakes when renegotiating using the function 748238405Sjkim SSL_renegotiate_abbreviated(). 749238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 750238405Sjkim 751238405Sjkim *) Add call to ENGINE_register_all_complete() to 752238405Sjkim ENGINE_load_builtin_engines(), so some implementations get used 753238405Sjkim automatically instead of needing explicit application support. 754238405Sjkim [Steve Henson] 755238405Sjkim 756238405Sjkim *) Add support for TLS key exporter as described in RFC5705. 757238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson] 758238405Sjkim 759238405Sjkim *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only 760238405Sjkim a few changes are required: 761238405Sjkim 762238405Sjkim Add SSL_OP_NO_TLSv1_1 flag. 763238405Sjkim Add TLSv1_1 methods. 764238405Sjkim Update version checking logic to handle version 1.1. 765238405Sjkim Add explicit IV handling (ported from DTLS code). 766238405Sjkim Add command line options to s_client/s_server. 767238405Sjkim [Steve Henson] 768238405Sjkim 769264331Sjkim Changes between 1.0.0j and 1.0.0k [5 Feb 2013] 770264331Sjkim 771264331Sjkim *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 772264331Sjkim 773264331Sjkim This addresses the flaw in CBC record processing discovered by 774264331Sjkim Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 775264331Sjkim at: http://www.isg.rhul.ac.uk/tls/ 776264331Sjkim 777264331Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 778264331Sjkim Security Group at Royal Holloway, University of London 779264331Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 780264331Sjkim Emilia K�sper for the initial patch. 781264331Sjkim (CVE-2013-0169) 782264331Sjkim [Emilia K�sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 783264331Sjkim 784264331Sjkim *) Return an error when checking OCSP signatures when key is NULL. 785264331Sjkim This fixes a DoS attack. (CVE-2013-0166) 786264331Sjkim [Steve Henson] 787264331Sjkim 788264331Sjkim *) Call OCSP Stapling callback after ciphersuite has been chosen, so 789264331Sjkim the right response is stapled. Also change SSL_get_certificate() 790264331Sjkim so it returns the certificate actually sent. 791264331Sjkim See http://rt.openssl.org/Ticket/Display.html?id=2836. 792264331Sjkim (This is a backport) 793264331Sjkim [Rob Stradling <rob.stradling@comodo.com>] 794264331Sjkim 795264331Sjkim *) Fix possible deadlock when decoding public keys. 796264331Sjkim [Steve Henson] 797264331Sjkim 798264331Sjkim Changes between 1.0.0i and 1.0.0j [10 May 2012] 799264331Sjkim 800264331Sjkim [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after 801264331Sjkim OpenSSL 1.0.1.] 802264331Sjkim 803264331Sjkim *) Sanity check record length before skipping explicit IV in DTLS 804264331Sjkim to fix DoS attack. 805264331Sjkim 806264331Sjkim Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 807264331Sjkim fuzzing as a service testing platform. 808264331Sjkim (CVE-2012-2333) 809264331Sjkim [Steve Henson] 810264331Sjkim 811264331Sjkim *) Initialise tkeylen properly when encrypting CMS messages. 812264331Sjkim Thanks to Solar Designer of Openwall for reporting this issue. 813264331Sjkim [Steve Henson] 814264331Sjkim 815264331Sjkim Changes between 1.0.0h and 1.0.0i [19 Apr 2012] 816264331Sjkim 817264331Sjkim *) Check for potentially exploitable overflows in asn1_d2i_read_bio 818264331Sjkim BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 819264331Sjkim in CRYPTO_realloc_clean. 820264331Sjkim 821264331Sjkim Thanks to Tavis Ormandy, Google Security Team, for discovering this 822264331Sjkim issue and to Adam Langley <agl@chromium.org> for fixing it. 823264331Sjkim (CVE-2012-2110) 824264331Sjkim [Adam Langley (Google), Tavis Ormandy, Google Security Team] 825264331Sjkim 826238405Sjkim Changes between 1.0.0g and 1.0.0h [12 Mar 2012] 827238405Sjkim 828237657Sjkim *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 829237657Sjkim in CMS and PKCS7 code. When RSA decryption fails use a random key for 830237657Sjkim content decryption and always return the same error. Note: this attack 831237657Sjkim needs on average 2^20 messages so it only affects automated senders. The 832237657Sjkim old behaviour can be reenabled in the CMS code by setting the 833237657Sjkim CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 834237657Sjkim an MMA defence is not necessary. 835237657Sjkim Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 836237657Sjkim this issue. (CVE-2012-0884) 837237657Sjkim [Steve Henson] 838237657Sjkim 839237657Sjkim *) Fix CVE-2011-4619: make sure we really are receiving a 840237657Sjkim client hello before rejecting multiple SGC restarts. Thanks to 841237657Sjkim Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 842237657Sjkim [Steve Henson] 843237657Sjkim 844238405Sjkim Changes between 1.0.0f and 1.0.0g [18 Jan 2012] 845238405Sjkim 846238405Sjkim *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 847238405Sjkim Thanks to Antonio Martin, Enterprise Secure Access Research and 848238405Sjkim Development, Cisco Systems, Inc. for discovering this bug and 849238405Sjkim preparing a fix. (CVE-2012-0050) 850238405Sjkim [Antonio Martin] 851238405Sjkim 852238405Sjkim Changes between 1.0.0e and 1.0.0f [4 Jan 2012] 853238405Sjkim 854238405Sjkim *) Nadhem Alfardan and Kenny Paterson have discovered an extension 855238405Sjkim of the Vaudenay padding oracle attack on CBC mode encryption 856238405Sjkim which enables an efficient plaintext recovery attack against 857238405Sjkim the OpenSSL implementation of DTLS. Their attack exploits timing 858238405Sjkim differences arising during decryption processing. A research 859238405Sjkim paper describing this attack can be found at: 860238405Sjkim http://www.isg.rhul.ac.uk/~kp/dtls.pdf 861238405Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 862238405Sjkim Security Group at Royal Holloway, University of London 863238405Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 864238405Sjkim <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 865238405Sjkim for preparing the fix. (CVE-2011-4108) 866238405Sjkim [Robin Seggelmann, Michael Tuexen] 867238405Sjkim 868238405Sjkim *) Clear bytes used for block padding of SSL 3.0 records. 869238405Sjkim (CVE-2011-4576) 870238405Sjkim [Adam Langley (Google)] 871238405Sjkim 872238405Sjkim *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 873238405Sjkim Kadianakis <desnacked@gmail.com> for discovering this issue and 874238405Sjkim Adam Langley for preparing the fix. (CVE-2011-4619) 875238405Sjkim [Adam Langley (Google)] 876238405Sjkim 877238405Sjkim *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027) 878238405Sjkim [Andrey Kulikov <amdeich@gmail.com>] 879238405Sjkim 880238405Sjkim *) Prevent malformed RFC3779 data triggering an assertion failure. 881238405Sjkim Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 882238405Sjkim and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 883238405Sjkim [Rob Austein <sra@hactrn.net>] 884238405Sjkim 885238405Sjkim *) Improved PRNG seeding for VOS. 886238405Sjkim [Paul Green <Paul.Green@stratus.com>] 887238405Sjkim 888238405Sjkim *) Fix ssl_ciph.c set-up race. 889238405Sjkim [Adam Langley (Google)] 890238405Sjkim 891238405Sjkim *) Fix spurious failures in ecdsatest.c. 892238405Sjkim [Emilia K�sper (Google)] 893238405Sjkim 894238405Sjkim *) Fix the BIO_f_buffer() implementation (which was mixing different 895238405Sjkim interpretations of the '..._len' fields). 896238405Sjkim [Adam Langley (Google)] 897238405Sjkim 898238405Sjkim *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 899238405Sjkim BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 900238405Sjkim threads won't reuse the same blinding coefficients. 901238405Sjkim 902238405Sjkim This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 903238405Sjkim lock to call BN_BLINDING_invert_ex, and avoids one use of 904238405Sjkim BN_BLINDING_update for each BN_BLINDING structure (previously, 905238405Sjkim the last update always remained unused). 906238405Sjkim [Emilia K�sper (Google)] 907238405Sjkim 908238405Sjkim *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf. 909238405Sjkim [Bob Buckholz (Google)] 910238405Sjkim 911238405Sjkim Changes between 1.0.0d and 1.0.0e [6 Sep 2011] 912238405Sjkim 913238405Sjkim *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted 914238405Sjkim by initialising X509_STORE_CTX properly. (CVE-2011-3207) 915238405Sjkim [Kaspar Brand <ossl@velox.ch>] 916238405Sjkim 917238405Sjkim *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 918238405Sjkim for multi-threaded use of ECDH. (CVE-2011-3210) 919238405Sjkim [Adam Langley (Google)] 920238405Sjkim 921238405Sjkim *) Fix x509_name_ex_d2i memory leak on bad inputs. 922238405Sjkim [Bodo Moeller] 923238405Sjkim 924238405Sjkim *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check 925238405Sjkim signature public key algorithm by using OID xref utilities instead. 926238405Sjkim Before this you could only use some ECC ciphersuites with SHA1 only. 927238405Sjkim [Steve Henson] 928238405Sjkim 929238405Sjkim *) Add protection against ECDSA timing attacks as mentioned in the paper 930238405Sjkim by Billy Bob Brumley and Nicola Tuveri, see: 931238405Sjkim 932238405Sjkim http://eprint.iacr.org/2011/232.pdf 933238405Sjkim 934238405Sjkim [Billy Bob Brumley and Nicola Tuveri] 935238405Sjkim 936238405Sjkim Changes between 1.0.0c and 1.0.0d [8 Feb 2011] 937238405Sjkim 938238405Sjkim *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 939238405Sjkim [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 940238405Sjkim 941238405Sjkim *) Fix bug in string printing code: if *any* escaping is enabled we must 942238405Sjkim escape the escape character (backslash) or the resulting string is 943238405Sjkim ambiguous. 944238405Sjkim [Steve Henson] 945238405Sjkim 946238405Sjkim Changes between 1.0.0b and 1.0.0c [2 Dec 2010] 947238405Sjkim 948238405Sjkim *) Disable code workaround for ancient and obsolete Netscape browsers 949238405Sjkim and servers: an attacker can use it in a ciphersuite downgrade attack. 950238405Sjkim Thanks to Martin Rex for discovering this bug. CVE-2010-4180 951238405Sjkim [Steve Henson] 952238405Sjkim 953238405Sjkim *) Fixed J-PAKE implementation error, originally discovered by 954238405Sjkim Sebastien Martini, further info and confirmation from Stefan 955238405Sjkim Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 956238405Sjkim [Ben Laurie] 957238405Sjkim 958238405Sjkim Changes between 1.0.0a and 1.0.0b [16 Nov 2010] 959238405Sjkim 960238405Sjkim *) Fix extension code to avoid race conditions which can result in a buffer 961238405Sjkim overrun vulnerability: resumed sessions must not be modified as they can 962238405Sjkim be shared by multiple threads. CVE-2010-3864 963238405Sjkim [Steve Henson] 964238405Sjkim 965238405Sjkim *) Fix WIN32 build system to correctly link an ENGINE directory into 966238405Sjkim a DLL. 967238405Sjkim [Steve Henson] 968238405Sjkim 969238405Sjkim Changes between 1.0.0 and 1.0.0a [01 Jun 2010] 970238405Sjkim 971238405Sjkim *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover 972238405Sjkim (CVE-2010-1633) 973238405Sjkim [Steve Henson, Peter-Michael Hager <hager@dortmund.net>] 974238405Sjkim 975238405Sjkim Changes between 0.9.8n and 1.0.0 [29 Mar 2010] 976238405Sjkim 977238405Sjkim *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher 978238405Sjkim context. The operation can be customised via the ctrl mechanism in 979238405Sjkim case ENGINEs want to include additional functionality. 980238405Sjkim [Steve Henson] 981238405Sjkim 982238405Sjkim *) Tolerate yet another broken PKCS#8 key format: private key value negative. 983238405Sjkim [Steve Henson] 984238405Sjkim 985238405Sjkim *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to 986238405Sjkim output hashes compatible with older versions of OpenSSL. 987238405Sjkim [Willy Weisz <weisz@vcpc.univie.ac.at>] 988238405Sjkim 989238405Sjkim *) Fix compression algorithm handling: if resuming a session use the 990238405Sjkim compression algorithm of the resumed session instead of determining 991238405Sjkim it from client hello again. Don't allow server to change algorithm. 992238405Sjkim [Steve Henson] 993238405Sjkim 994238405Sjkim *) Add load_crls() function to apps tidying load_certs() too. Add option 995238405Sjkim to verify utility to allow additional CRLs to be included. 996238405Sjkim [Steve Henson] 997238405Sjkim 998238405Sjkim *) Update OCSP request code to permit adding custom headers to the request: 999238405Sjkim some responders need this. 1000238405Sjkim [Steve Henson] 1001238405Sjkim 1002238405Sjkim *) The function EVP_PKEY_sign() returns <=0 on error: check return code 1003238405Sjkim correctly. 1004238405Sjkim [Julia Lawall <julia@diku.dk>] 1005238405Sjkim 1006238405Sjkim *) Update verify callback code in apps/s_cb.c and apps/verify.c, it 1007238405Sjkim needlessly dereferenced structures, used obsolete functions and 1008238405Sjkim didn't handle all updated verify codes correctly. 1009238405Sjkim [Steve Henson] 1010238405Sjkim 1011238405Sjkim *) Disable MD2 in the default configuration. 1012238405Sjkim [Steve Henson] 1013238405Sjkim 1014238405Sjkim *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to 1015238405Sjkim indicate the initial BIO being pushed or popped. This makes it possible 1016238405Sjkim to determine whether the BIO is the one explicitly called or as a result 1017238405Sjkim of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so 1018238405Sjkim it handles reference counts correctly and doesn't zero out the I/O bio 1019238405Sjkim when it is not being explicitly popped. WARNING: applications which 1020238405Sjkim included workarounds for the old buggy behaviour will need to be modified 1021238405Sjkim or they could free up already freed BIOs. 1022238405Sjkim [Steve Henson] 1023238405Sjkim 1024238405Sjkim *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni 1025238405Sjkim renaming to all platforms (within the 0.9.8 branch, this was 1026238405Sjkim done conditionally on Netware platforms to avoid a name clash). 1027238405Sjkim [Guenter <lists@gknw.net>] 1028238405Sjkim 1029238405Sjkim *) Add ECDHE and PSK support to DTLS. 1030238405Sjkim [Michael Tuexen <tuexen@fh-muenster.de>] 1031238405Sjkim 1032238405Sjkim *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't 1033238405Sjkim be used on C++. 1034238405Sjkim [Steve Henson] 1035238405Sjkim 1036238405Sjkim *) Add "missing" function EVP_MD_flags() (without this the only way to 1037238405Sjkim retrieve a digest flags is by accessing the structure directly. Update 1038238405Sjkim EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest 1039238405Sjkim or cipher is registered as in the "from" argument. Print out all 1040238405Sjkim registered digests in the dgst usage message instead of manually 1041238405Sjkim attempting to work them out. 1042238405Sjkim [Steve Henson] 1043238405Sjkim 1044238405Sjkim *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello: 1045238405Sjkim this allows the use of compression and extensions. Change default cipher 1046238405Sjkim string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2 1047238405Sjkim by default unless an application cipher string requests it. 1048238405Sjkim [Steve Henson] 1049238405Sjkim 1050238405Sjkim *) Alter match criteria in PKCS12_parse(). It used to try to use local 1051238405Sjkim key ids to find matching certificates and keys but some PKCS#12 files 1052238405Sjkim don't follow the (somewhat unwritten) rules and this strategy fails. 1053238405Sjkim Now just gather all certificates together and the first private key 1054238405Sjkim then look for the first certificate that matches the key. 1055238405Sjkim [Steve Henson] 1056238405Sjkim 1057238405Sjkim *) Support use of registered digest and cipher names for dgst and cipher 1058238405Sjkim commands instead of having to add each one as a special case. So now 1059238405Sjkim you can do: 1060238405Sjkim 1061238405Sjkim openssl sha256 foo 1062238405Sjkim 1063238405Sjkim as well as: 1064238405Sjkim 1065238405Sjkim openssl dgst -sha256 foo 1066238405Sjkim 1067238405Sjkim and this works for ENGINE based algorithms too. 1068238405Sjkim 1069238405Sjkim [Steve Henson] 1070238405Sjkim 1071238405Sjkim *) Update Gost ENGINE to support parameter files. 1072238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 1073238405Sjkim 1074238405Sjkim *) Support GeneralizedTime in ca utility. 1075238405Sjkim [Oliver Martin <oliver@volatilevoid.net>, Steve Henson] 1076238405Sjkim 1077238405Sjkim *) Enhance the hash format used for certificate directory links. The new 1078238405Sjkim form uses the canonical encoding (meaning equivalent names will work 1079238405Sjkim even if they aren't identical) and uses SHA1 instead of MD5. This form 1080238405Sjkim is incompatible with the older format and as a result c_rehash should 1081238405Sjkim be used to rebuild symbolic links. 1082238405Sjkim [Steve Henson] 1083238405Sjkim 1084238405Sjkim *) Make PKCS#8 the default write format for private keys, replacing the 1085238405Sjkim traditional format. This form is standardised, more secure and doesn't 1086238405Sjkim include an implicit MD5 dependency. 1087238405Sjkim [Steve Henson] 1088238405Sjkim 1089238405Sjkim *) Add a $gcc_devteam_warn option to Configure. The idea is that any code 1090238405Sjkim committed to OpenSSL should pass this lot as a minimum. 1091238405Sjkim [Steve Henson] 1092238405Sjkim 1093238405Sjkim *) Add session ticket override functionality for use by EAP-FAST. 1094238405Sjkim [Jouni Malinen <j@w1.fi>] 1095238405Sjkim 1096238405Sjkim *) Modify HMAC functions to return a value. Since these can be implemented 1097238405Sjkim in an ENGINE errors can occur. 1098238405Sjkim [Steve Henson] 1099238405Sjkim 1100238405Sjkim *) Type-checked OBJ_bsearch_ex. 1101238405Sjkim [Ben Laurie] 1102238405Sjkim 1103238405Sjkim *) Type-checked OBJ_bsearch. Also some constification necessitated 1104238405Sjkim by type-checking. Still to come: TXT_DB, bsearch(?), 1105238405Sjkim OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING, 1106238405Sjkim CONF_VALUE. 1107238405Sjkim [Ben Laurie] 1108238405Sjkim 1109238405Sjkim *) New function OPENSSL_gmtime_adj() to add a specific number of days and 1110238405Sjkim seconds to a tm structure directly, instead of going through OS 1111238405Sjkim specific date routines. This avoids any issues with OS routines such 1112238405Sjkim as the year 2038 bug. New *_adj() functions for ASN1 time structures 1113238405Sjkim and X509_time_adj_ex() to cover the extended range. The existing 1114238405Sjkim X509_time_adj() is still usable and will no longer have any date issues. 1115238405Sjkim [Steve Henson] 1116238405Sjkim 1117238405Sjkim *) Delta CRL support. New use deltas option which will attempt to locate 1118238405Sjkim and search any appropriate delta CRLs available. 1119238405Sjkim 1120238405Sjkim This work was sponsored by Google. 1121238405Sjkim [Steve Henson] 1122238405Sjkim 1123238405Sjkim *) Support for CRLs partitioned by reason code. Reorganise CRL processing 1124238405Sjkim code and add additional score elements. Validate alternate CRL paths 1125238405Sjkim as part of the CRL checking and indicate a new error "CRL path validation 1126238405Sjkim error" in this case. Applications wanting additional details can use 1127238405Sjkim the verify callback and check the new "parent" field. If this is not 1128238405Sjkim NULL CRL path validation is taking place. Existing applications wont 1129238405Sjkim see this because it requires extended CRL support which is off by 1130238405Sjkim default. 1131238405Sjkim 1132238405Sjkim This work was sponsored by Google. 1133238405Sjkim [Steve Henson] 1134238405Sjkim 1135238405Sjkim *) Support for freshest CRL extension. 1136238405Sjkim 1137238405Sjkim This work was sponsored by Google. 1138238405Sjkim [Steve Henson] 1139238405Sjkim 1140238405Sjkim *) Initial indirect CRL support. Currently only supported in the CRLs 1141238405Sjkim passed directly and not via lookup. Process certificate issuer 1142238405Sjkim CRL entry extension and lookup CRL entries by bother issuer name 1143238405Sjkim and serial number. Check and process CRL issuer entry in IDP extension. 1144238405Sjkim 1145238405Sjkim This work was sponsored by Google. 1146238405Sjkim [Steve Henson] 1147238405Sjkim 1148238405Sjkim *) Add support for distinct certificate and CRL paths. The CRL issuer 1149238405Sjkim certificate is validated separately in this case. Only enabled if 1150238405Sjkim an extended CRL support flag is set: this flag will enable additional 1151238405Sjkim CRL functionality in future. 1152238405Sjkim 1153238405Sjkim This work was sponsored by Google. 1154238405Sjkim [Steve Henson] 1155238405Sjkim 1156238405Sjkim *) Add support for policy mappings extension. 1157238405Sjkim 1158238405Sjkim This work was sponsored by Google. 1159238405Sjkim [Steve Henson] 1160238405Sjkim 1161238405Sjkim *) Fixes to pathlength constraint, self issued certificate handling, 1162238405Sjkim policy processing to align with RFC3280 and PKITS tests. 1163238405Sjkim 1164238405Sjkim This work was sponsored by Google. 1165238405Sjkim [Steve Henson] 1166238405Sjkim 1167238405Sjkim *) Support for name constraints certificate extension. DN, email, DNS 1168238405Sjkim and URI types are currently supported. 1169238405Sjkim 1170238405Sjkim This work was sponsored by Google. 1171238405Sjkim [Steve Henson] 1172238405Sjkim 1173238405Sjkim *) To cater for systems that provide a pointer-based thread ID rather 1174238405Sjkim than numeric, deprecate the current numeric thread ID mechanism and 1175238405Sjkim replace it with a structure and associated callback type. This 1176238405Sjkim mechanism allows a numeric "hash" to be extracted from a thread ID in 1177238405Sjkim either case, and on platforms where pointers are larger than 'long', 1178238405Sjkim mixing is done to help ensure the numeric 'hash' is usable even if it 1179238405Sjkim can't be guaranteed unique. The default mechanism is to use "&errno" 1180238405Sjkim as a pointer-based thread ID to distinguish between threads. 1181238405Sjkim 1182238405Sjkim Applications that want to provide their own thread IDs should now use 1183238405Sjkim CRYPTO_THREADID_set_callback() to register a callback that will call 1184238405Sjkim either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer(). 1185238405Sjkim 1186238405Sjkim Note that ERR_remove_state() is now deprecated, because it is tied 1187238405Sjkim to the assumption that thread IDs are numeric. ERR_remove_state(0) 1188238405Sjkim to free the current thread's error state should be replaced by 1189238405Sjkim ERR_remove_thread_state(NULL). 1190238405Sjkim 1191238405Sjkim (This new approach replaces the functions CRYPTO_set_idptr_callback(), 1192238405Sjkim CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in 1193238405Sjkim OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an 1194238405Sjkim application was previously providing a numeric thread callback that 1195238405Sjkim was inappropriate for distinguishing threads, then uniqueness might 1196238405Sjkim have been obtained with &errno that happened immediately in the 1197238405Sjkim intermediate development versions of OpenSSL; this is no longer the 1198238405Sjkim case, the numeric thread callback will now override the automatic use 1199238405Sjkim of &errno.) 1200238405Sjkim [Geoff Thorpe, with help from Bodo Moeller] 1201238405Sjkim 1202238405Sjkim *) Initial support for different CRL issuing certificates. This covers a 1203238405Sjkim simple case where the self issued certificates in the chain exist and 1204238405Sjkim the real CRL issuer is higher in the existing chain. 1205238405Sjkim 1206238405Sjkim This work was sponsored by Google. 1207238405Sjkim [Steve Henson] 1208238405Sjkim 1209238405Sjkim *) Removed effectively defunct crypto/store from the build. 1210238405Sjkim [Ben Laurie] 1211238405Sjkim 1212238405Sjkim *) Revamp of STACK to provide stronger type-checking. Still to come: 1213238405Sjkim TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE, 1214238405Sjkim ASN1_STRING, CONF_VALUE. 1215238405Sjkim [Ben Laurie] 1216238405Sjkim 1217238405Sjkim *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer 1218238405Sjkim RAM on SSL connections. This option can save about 34k per idle SSL. 1219238405Sjkim [Nick Mathewson] 1220238405Sjkim 1221238405Sjkim *) Revamp of LHASH to provide stronger type-checking. Still to come: 1222238405Sjkim STACK, TXT_DB, bsearch, qsort. 1223238405Sjkim [Ben Laurie] 1224238405Sjkim 1225238405Sjkim *) Initial support for Cryptographic Message Syntax (aka CMS) based 1226238405Sjkim on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility, 1227238405Sjkim support for data, signedData, compressedData, digestedData and 1228238405Sjkim encryptedData, envelopedData types included. Scripts to check against 1229238405Sjkim RFC4134 examples draft and interop and consistency checks of many 1230238405Sjkim content types and variants. 1231238405Sjkim [Steve Henson] 1232238405Sjkim 1233238405Sjkim *) Add options to enc utility to support use of zlib compression BIO. 1234238405Sjkim [Steve Henson] 1235238405Sjkim 1236238405Sjkim *) Extend mk1mf to support importing of options and assembly language 1237238405Sjkim files from Configure script, currently only included in VC-WIN32. 1238238405Sjkim The assembly language rules can now optionally generate the source 1239238405Sjkim files from the associated perl scripts. 1240238405Sjkim [Steve Henson] 1241238405Sjkim 1242238405Sjkim *) Implement remaining functionality needed to support GOST ciphersuites. 1243238405Sjkim Interop testing has been performed using CryptoPro implementations. 1244238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 1245238405Sjkim 1246238405Sjkim *) s390x assembler pack. 1247238405Sjkim [Andy Polyakov] 1248238405Sjkim 1249238405Sjkim *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU 1250238405Sjkim "family." 1251238405Sjkim [Andy Polyakov] 1252238405Sjkim 1253238405Sjkim *) Implement Opaque PRF Input TLS extension as specified in 1254238405Sjkim draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an 1255238405Sjkim official specification yet and no extension type assignment by 1256238405Sjkim IANA exists, this extension (for now) will have to be explicitly 1257238405Sjkim enabled when building OpenSSL by providing the extension number 1258238405Sjkim to use. For example, specify an option 1259238405Sjkim 1260238405Sjkim -DTLSEXT_TYPE_opaque_prf_input=0x9527 1261238405Sjkim 1262238405Sjkim to the "config" or "Configure" script to enable the extension, 1263238405Sjkim assuming extension number 0x9527 (which is a completely arbitrary 1264238405Sjkim and unofficial assignment based on the MD5 hash of the Internet 1265238405Sjkim Draft). Note that by doing so, you potentially lose 1266238405Sjkim interoperability with other TLS implementations since these might 1267238405Sjkim be using the same extension number for other purposes. 1268238405Sjkim 1269238405Sjkim SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the 1270238405Sjkim opaque PRF input value to use in the handshake. This will create 1271238405Sjkim an interal copy of the length-'len' string at 'src', and will 1272238405Sjkim return non-zero for success. 1273238405Sjkim 1274238405Sjkim To get more control and flexibility, provide a callback function 1275238405Sjkim by using 1276238405Sjkim 1277238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb) 1278238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg) 1279238405Sjkim 1280238405Sjkim where 1281238405Sjkim 1282238405Sjkim int (*cb)(SSL *, void *peerinput, size_t len, void *arg); 1283238405Sjkim void *arg; 1284238405Sjkim 1285238405Sjkim Callback function 'cb' will be called in handshakes, and is 1286238405Sjkim expected to use SSL_set_tlsext_opaque_prf_input() as appropriate. 1287238405Sjkim Argument 'arg' is for application purposes (the value as given to 1288238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly 1289238405Sjkim be provided to the callback function). The callback function 1290238405Sjkim has to return non-zero to report success: usually 1 to use opaque 1291238405Sjkim PRF input just if possible, or 2 to enforce use of the opaque PRF 1292238405Sjkim input. In the latter case, the library will abort the handshake 1293238405Sjkim if opaque PRF input is not successfully negotiated. 1294238405Sjkim 1295238405Sjkim Arguments 'peerinput' and 'len' given to the callback function 1296238405Sjkim will always be NULL and 0 in the case of a client. A server will 1297238405Sjkim see the client's opaque PRF input through these variables if 1298238405Sjkim available (NULL and 0 otherwise). Note that if the server 1299238405Sjkim provides an opaque PRF input, the length must be the same as the 1300238405Sjkim length of the client's opaque PRF input. 1301238405Sjkim 1302238405Sjkim Note that the callback function will only be called when creating 1303238405Sjkim a new session (session resumption can resume whatever was 1304238405Sjkim previously negotiated), and will not be called in SSL 2.0 1305238405Sjkim handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or 1306238405Sjkim SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended 1307238405Sjkim for applications that need to enforce opaque PRF input. 1308238405Sjkim 1309238405Sjkim [Bodo Moeller] 1310238405Sjkim 1311238405Sjkim *) Update ssl code to support digests other than SHA1+MD5 for handshake 1312238405Sjkim MAC. 1313238405Sjkim 1314238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 1315238405Sjkim 1316238405Sjkim *) Add RFC4507 support to OpenSSL. This includes the corrections in 1317238405Sjkim RFC4507bis. The encrypted ticket format is an encrypted encoded 1318238405Sjkim SSL_SESSION structure, that way new session features are automatically 1319238405Sjkim supported. 1320238405Sjkim 1321238405Sjkim If a client application caches session in an SSL_SESSION structure 1322238405Sjkim support is transparent because tickets are now stored in the encoded 1323238405Sjkim SSL_SESSION. 1324238405Sjkim 1325238405Sjkim The SSL_CTX structure automatically generates keys for ticket 1326238405Sjkim protection in servers so again support should be possible 1327238405Sjkim with no application modification. 1328238405Sjkim 1329238405Sjkim If a client or server wishes to disable RFC4507 support then the option 1330238405Sjkim SSL_OP_NO_TICKET can be set. 1331238405Sjkim 1332238405Sjkim Add a TLS extension debugging callback to allow the contents of any client 1333238405Sjkim or server extensions to be examined. 1334238405Sjkim 1335238405Sjkim This work was sponsored by Google. 1336238405Sjkim [Steve Henson] 1337238405Sjkim 1338238405Sjkim *) Final changes to avoid use of pointer pointer casts in OpenSSL. 1339238405Sjkim OpenSSL should now compile cleanly on gcc 4.2 1340238405Sjkim [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson] 1341238405Sjkim 1342238405Sjkim *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC 1343238405Sjkim support including streaming MAC support: this is required for GOST 1344238405Sjkim ciphersuite support. 1345238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson] 1346238405Sjkim 1347238405Sjkim *) Add option -stream to use PKCS#7 streaming in smime utility. New 1348238405Sjkim function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream() 1349238405Sjkim to output in BER and PEM format. 1350238405Sjkim [Steve Henson] 1351238405Sjkim 1352238405Sjkim *) Experimental support for use of HMAC via EVP_PKEY interface. This 1353238405Sjkim allows HMAC to be handled via the EVP_DigestSign*() interface. The 1354238405Sjkim EVP_PKEY "key" in this case is the HMAC key, potentially allowing 1355238405Sjkim ENGINE support for HMAC keys which are unextractable. New -mac and 1356238405Sjkim -macopt options to dgst utility. 1357238405Sjkim [Steve Henson] 1358238405Sjkim 1359238405Sjkim *) New option -sigopt to dgst utility. Update dgst to use 1360238405Sjkim EVP_Digest{Sign,Verify}*. These two changes make it possible to use 1361238405Sjkim alternative signing paramaters such as X9.31 or PSS in the dgst 1362238405Sjkim utility. 1363238405Sjkim [Steve Henson] 1364238405Sjkim 1365238405Sjkim *) Change ssl_cipher_apply_rule(), the internal function that does 1366238405Sjkim the work each time a ciphersuite string requests enabling 1367238405Sjkim ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or 1368238405Sjkim removing ("!foo+bar") a class of ciphersuites: Now it maintains 1369238405Sjkim the order of disabled ciphersuites such that those ciphersuites 1370238405Sjkim that most recently went from enabled to disabled not only stay 1371238405Sjkim in order with respect to each other, but also have higher priority 1372238405Sjkim than other disabled ciphersuites the next time ciphersuites are 1373238405Sjkim enabled again. 1374238405Sjkim 1375238405Sjkim This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable 1376238405Sjkim the same ciphersuites as with "HIGH" alone, but in a specific 1377238405Sjkim order where the PSK ciphersuites come first (since they are the 1378238405Sjkim most recently disabled ciphersuites when "HIGH" is parsed). 1379238405Sjkim 1380238405Sjkim Also, change ssl_create_cipher_list() (using this new 1381238405Sjkim funcionality) such that between otherwise identical 1382238405Sjkim cihpersuites, ephemeral ECDH is preferred over ephemeral DH in 1383238405Sjkim the default order. 1384238405Sjkim [Bodo Moeller] 1385238405Sjkim 1386238405Sjkim *) Change ssl_create_cipher_list() so that it automatically 1387238405Sjkim arranges the ciphersuites in reasonable order before starting 1388238405Sjkim to process the rule string. Thus, the definition for "DEFAULT" 1389238405Sjkim (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but 1390238405Sjkim remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH". 1391238405Sjkim This makes it much easier to arrive at a reasonable default order 1392238405Sjkim in applications for which anonymous ciphers are OK (meaning 1393238405Sjkim that you can't actually use DEFAULT). 1394238405Sjkim [Bodo Moeller; suggested by Victor Duchovni] 1395238405Sjkim 1396238405Sjkim *) Split the SSL/TLS algorithm mask (as used for ciphersuite string 1397238405Sjkim processing) into multiple integers instead of setting 1398238405Sjkim "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK", 1399238405Sjkim "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer. 1400238405Sjkim (These masks as well as the individual bit definitions are hidden 1401238405Sjkim away into the non-exported interface ssl/ssl_locl.h, so this 1402238405Sjkim change to the definition of the SSL_CIPHER structure shouldn't 1403238405Sjkim affect applications.) This give us more bits for each of these 1404238405Sjkim categories, so there is no longer a need to coagulate AES128 and 1405238405Sjkim AES256 into a single algorithm bit, and to coagulate Camellia128 1406238405Sjkim and Camellia256 into a single algorithm bit, which has led to all 1407238405Sjkim kinds of kludges. 1408238405Sjkim 1409238405Sjkim Thus, among other things, the kludge introduced in 0.9.7m and 1410238405Sjkim 0.9.8e for masking out AES256 independently of AES128 or masking 1411238405Sjkim out Camellia256 independently of AES256 is not needed here in 0.9.9. 1412238405Sjkim 1413238405Sjkim With the change, we also introduce new ciphersuite aliases that 1414238405Sjkim so far were missing: "AES128", "AES256", "CAMELLIA128", and 1415238405Sjkim "CAMELLIA256". 1416238405Sjkim [Bodo Moeller] 1417238405Sjkim 1418238405Sjkim *) Add support for dsa-with-SHA224 and dsa-with-SHA256. 1419238405Sjkim Use the leftmost N bytes of the signature input if the input is 1420238405Sjkim larger than the prime q (with N being the size in bytes of q). 1421238405Sjkim [Nils Larsch] 1422238405Sjkim 1423238405Sjkim *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses 1424238405Sjkim it yet and it is largely untested. 1425238405Sjkim [Steve Henson] 1426238405Sjkim 1427238405Sjkim *) Add support for the ecdsa-with-SHA224/256/384/512 signature types. 1428238405Sjkim [Nils Larsch] 1429238405Sjkim 1430238405Sjkim *) Initial incomplete changes to avoid need for function casts in OpenSSL 1431238405Sjkim some compilers (gcc 4.2 and later) reject their use. Safestack is 1432238405Sjkim reimplemented. Update ASN1 to avoid use of legacy functions. 1433238405Sjkim [Steve Henson] 1434238405Sjkim 1435238405Sjkim *) Win32/64 targets are linked with Winsock2. 1436238405Sjkim [Andy Polyakov] 1437238405Sjkim 1438238405Sjkim *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected 1439238405Sjkim to external functions. This can be used to increase CRL handling 1440238405Sjkim efficiency especially when CRLs are very large by (for example) storing 1441238405Sjkim the CRL revoked certificates in a database. 1442238405Sjkim [Steve Henson] 1443238405Sjkim 1444238405Sjkim *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so 1445238405Sjkim new CRLs added to a directory can be used. New command line option 1446238405Sjkim -verify_return_error to s_client and s_server. This causes real errors 1447238405Sjkim to be returned by the verify callback instead of carrying on no matter 1448238405Sjkim what. This reflects the way a "real world" verify callback would behave. 1449238405Sjkim [Steve Henson] 1450238405Sjkim 1451238405Sjkim *) GOST engine, supporting several GOST algorithms and public key formats. 1452238405Sjkim Kindly donated by Cryptocom. 1453238405Sjkim [Cryptocom] 1454238405Sjkim 1455238405Sjkim *) Partial support for Issuing Distribution Point CRL extension. CRLs 1456238405Sjkim partitioned by DP are handled but no indirect CRL or reason partitioning 1457238405Sjkim (yet). Complete overhaul of CRL handling: now the most suitable CRL is 1458238405Sjkim selected via a scoring technique which handles IDP and AKID in CRLs. 1459238405Sjkim [Steve Henson] 1460238405Sjkim 1461238405Sjkim *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which 1462238405Sjkim will ultimately be used for all verify operations: this will remove the 1463238405Sjkim X509_STORE dependency on certificate verification and allow alternative 1464238405Sjkim lookup methods. X509_STORE based implementations of these two callbacks. 1465238405Sjkim [Steve Henson] 1466238405Sjkim 1467238405Sjkim *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names. 1468238405Sjkim Modify get_crl() to find a valid (unexpired) CRL if possible. 1469238405Sjkim [Steve Henson] 1470238405Sjkim 1471238405Sjkim *) New function X509_CRL_match() to check if two CRLs are identical. Normally 1472238405Sjkim this would be called X509_CRL_cmp() but that name is already used by 1473238405Sjkim a function that just compares CRL issuer names. Cache several CRL 1474238405Sjkim extensions in X509_CRL structure and cache CRLDP in X509. 1475238405Sjkim [Steve Henson] 1476238405Sjkim 1477238405Sjkim *) Store a "canonical" representation of X509_NAME structure (ASN1 Name) 1478238405Sjkim this maps equivalent X509_NAME structures into a consistent structure. 1479238405Sjkim Name comparison can then be performed rapidly using memcmp(). 1480238405Sjkim [Steve Henson] 1481238405Sjkim 1482238405Sjkim *) Non-blocking OCSP request processing. Add -timeout option to ocsp 1483238405Sjkim utility. 1484238405Sjkim [Steve Henson] 1485238405Sjkim 1486238405Sjkim *) Allow digests to supply their own micalg string for S/MIME type using 1487238405Sjkim the ctrl EVP_MD_CTRL_MICALG. 1488238405Sjkim [Steve Henson] 1489238405Sjkim 1490238405Sjkim *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the 1491238405Sjkim EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN 1492238405Sjkim ctrl. It can then customise the structure before and/or after signing 1493238405Sjkim if necessary. 1494238405Sjkim [Steve Henson] 1495238405Sjkim 1496238405Sjkim *) New function OBJ_add_sigid() to allow application defined signature OIDs 1497238405Sjkim to be added to OpenSSLs internal tables. New function OBJ_sigid_free() 1498238405Sjkim to free up any added signature OIDs. 1499238405Sjkim [Steve Henson] 1500238405Sjkim 1501238405Sjkim *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(), 1502238405Sjkim EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal 1503238405Sjkim digest and cipher tables. New options added to openssl utility: 1504238405Sjkim list-message-digest-algorithms and list-cipher-algorithms. 1505238405Sjkim [Steve Henson] 1506238405Sjkim 1507238405Sjkim *) Change the array representation of binary polynomials: the list 1508238405Sjkim of degrees of non-zero coefficients is now terminated with -1. 1509238405Sjkim Previously it was terminated with 0, which was also part of the 1510238405Sjkim value; thus, the array representation was not applicable to 1511238405Sjkim polynomials where t^0 has coefficient zero. This change makes 1512238405Sjkim the array representation useful in a more general context. 1513238405Sjkim [Douglas Stebila] 1514238405Sjkim 1515238405Sjkim *) Various modifications and fixes to SSL/TLS cipher string 1516238405Sjkim handling. For ECC, the code now distinguishes between fixed ECDH 1517238405Sjkim with RSA certificates on the one hand and with ECDSA certificates 1518238405Sjkim on the other hand, since these are separate ciphersuites. The 1519238405Sjkim unused code for Fortezza ciphersuites has been removed. 1520238405Sjkim 1521238405Sjkim For consistency with EDH, ephemeral ECDH is now called "EECDH" 1522238405Sjkim (not "ECDHE"). For consistency with the code for DH 1523238405Sjkim certificates, use of ECDH certificates is now considered ECDH 1524238405Sjkim authentication, not RSA or ECDSA authentication (the latter is 1525238405Sjkim merely the CA's signing algorithm and not actively used in the 1526238405Sjkim protocol). 1527238405Sjkim 1528238405Sjkim The temporary ciphersuite alias "ECCdraft" is no longer 1529238405Sjkim available, and ECC ciphersuites are no longer excluded from "ALL" 1530238405Sjkim and "DEFAULT". The following aliases now exist for RFC 4492 1531238405Sjkim ciphersuites, most of these by analogy with the DH case: 1532238405Sjkim 1533238405Sjkim kECDHr - ECDH cert, signed with RSA 1534238405Sjkim kECDHe - ECDH cert, signed with ECDSA 1535238405Sjkim kECDH - ECDH cert (signed with either RSA or ECDSA) 1536238405Sjkim kEECDH - ephemeral ECDH 1537238405Sjkim ECDH - ECDH cert or ephemeral ECDH 1538238405Sjkim 1539238405Sjkim aECDH - ECDH cert 1540238405Sjkim aECDSA - ECDSA cert 1541238405Sjkim ECDSA - ECDSA cert 1542238405Sjkim 1543238405Sjkim AECDH - anonymous ECDH 1544238405Sjkim EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH") 1545238405Sjkim 1546238405Sjkim [Bodo Moeller] 1547238405Sjkim 1548238405Sjkim *) Add additional S/MIME capabilities for AES and GOST ciphers if supported. 1549238405Sjkim Use correct micalg parameters depending on digest(s) in signed message. 1550238405Sjkim [Steve Henson] 1551238405Sjkim 1552238405Sjkim *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process 1553238405Sjkim an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code. 1554238405Sjkim [Steve Henson] 1555238405Sjkim 1556238405Sjkim *) Initial engine support for EVP_PKEY_METHOD. New functions to permit 1557238405Sjkim an engine to register a method. Add ENGINE lookups for methods and 1558238405Sjkim functional reference processing. 1559238405Sjkim [Steve Henson] 1560238405Sjkim 1561238405Sjkim *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of 1562238405Sjkim EVP_{Sign,Verify}* which allow an application to customise the signature 1563238405Sjkim process. 1564238405Sjkim [Steve Henson] 1565238405Sjkim 1566238405Sjkim *) New -resign option to smime utility. This adds one or more signers 1567238405Sjkim to an existing PKCS#7 signedData structure. Also -md option to use an 1568238405Sjkim alternative message digest algorithm for signing. 1569238405Sjkim [Steve Henson] 1570238405Sjkim 1571238405Sjkim *) Tidy up PKCS#7 routines and add new functions to make it easier to 1572238405Sjkim create PKCS7 structures containing multiple signers. Update smime 1573238405Sjkim application to support multiple signers. 1574238405Sjkim [Steve Henson] 1575238405Sjkim 1576238405Sjkim *) New -macalg option to pkcs12 utility to allow setting of an alternative 1577238405Sjkim digest MAC. 1578238405Sjkim [Steve Henson] 1579238405Sjkim 1580238405Sjkim *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC. 1581238405Sjkim Reorganize PBE internals to lookup from a static table using NIDs, 1582238405Sjkim add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl: 1583238405Sjkim EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative 1584238405Sjkim PRF which will be automatically used with PBES2. 1585238405Sjkim [Steve Henson] 1586238405Sjkim 1587238405Sjkim *) Replace the algorithm specific calls to generate keys in "req" with the 1588238405Sjkim new API. 1589238405Sjkim [Steve Henson] 1590238405Sjkim 1591238405Sjkim *) Update PKCS#7 enveloped data routines to use new API. This is now 1592238405Sjkim supported by any public key method supporting the encrypt operation. A 1593238405Sjkim ctrl is added to allow the public key algorithm to examine or modify 1594238405Sjkim the PKCS#7 RecipientInfo structure if it needs to: for RSA this is 1595238405Sjkim a no op. 1596238405Sjkim [Steve Henson] 1597238405Sjkim 1598238405Sjkim *) Add a ctrl to asn1 method to allow a public key algorithm to express 1599238405Sjkim a default digest type to use. In most cases this will be SHA1 but some 1600238405Sjkim algorithms (such as GOST) need to specify an alternative digest. The 1601238405Sjkim return value indicates how strong the prefernce is 1 means optional and 1602238405Sjkim 2 is mandatory (that is it is the only supported type). Modify 1603238405Sjkim ASN1_item_sign() to accept a NULL digest argument to indicate it should 1604238405Sjkim use the default md. Update openssl utilities to use the default digest 1605238405Sjkim type for signing if it is not explicitly indicated. 1606238405Sjkim [Steve Henson] 1607238405Sjkim 1608238405Sjkim *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New 1609238405Sjkim EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant 1610238405Sjkim signing method from the key type. This effectively removes the link 1611238405Sjkim between digests and public key types. 1612238405Sjkim [Steve Henson] 1613238405Sjkim 1614238405Sjkim *) Add an OID cross reference table and utility functions. Its purpose is to 1615238405Sjkim translate between signature OIDs such as SHA1WithrsaEncryption and SHA1, 1616238405Sjkim rsaEncryption. This will allow some of the algorithm specific hackery 1617238405Sjkim needed to use the correct OID to be removed. 1618238405Sjkim [Steve Henson] 1619238405Sjkim 1620238405Sjkim *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO 1621238405Sjkim structures for PKCS7_sign(). They are now set up by the relevant public 1622238405Sjkim key ASN1 method. 1623238405Sjkim [Steve Henson] 1624238405Sjkim 1625238405Sjkim *) Add provisional EC pkey method with support for ECDSA and ECDH. 1626238405Sjkim [Steve Henson] 1627238405Sjkim 1628238405Sjkim *) Add support for key derivation (agreement) in the API, DH method and 1629238405Sjkim pkeyutl. 1630238405Sjkim [Steve Henson] 1631238405Sjkim 1632238405Sjkim *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support 1633238405Sjkim public and private key formats. As a side effect these add additional 1634238405Sjkim command line functionality not previously available: DSA signatures can be 1635238405Sjkim generated and verified using pkeyutl and DH key support and generation in 1636238405Sjkim pkey, genpkey. 1637238405Sjkim [Steve Henson] 1638238405Sjkim 1639238405Sjkim *) BeOS support. 1640238405Sjkim [Oliver Tappe <zooey@hirschkaefer.de>] 1641238405Sjkim 1642238405Sjkim *) New make target "install_html_docs" installs HTML renditions of the 1643238405Sjkim manual pages. 1644238405Sjkim [Oliver Tappe <zooey@hirschkaefer.de>] 1645238405Sjkim 1646238405Sjkim *) New utility "genpkey" this is analagous to "genrsa" etc except it can 1647238405Sjkim generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to 1648238405Sjkim support key and parameter generation and add initial key generation 1649238405Sjkim functionality for RSA. 1650238405Sjkim [Steve Henson] 1651238405Sjkim 1652238405Sjkim *) Add functions for main EVP_PKEY_method operations. The undocumented 1653238405Sjkim functions EVP_PKEY_{encrypt,decrypt} have been renamed to 1654238405Sjkim EVP_PKEY_{encrypt,decrypt}_old. 1655238405Sjkim [Steve Henson] 1656238405Sjkim 1657238405Sjkim *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public 1658238405Sjkim key API, doesn't do much yet. 1659238405Sjkim [Steve Henson] 1660238405Sjkim 1661238405Sjkim *) New function EVP_PKEY_asn1_get0_info() to retrieve information about 1662238405Sjkim public key algorithms. New option to openssl utility: 1663238405Sjkim "list-public-key-algorithms" to print out info. 1664238405Sjkim [Steve Henson] 1665238405Sjkim 1666238405Sjkim *) Implement the Supported Elliptic Curves Extension for 1667238405Sjkim ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1668238405Sjkim [Douglas Stebila] 1669238405Sjkim 1670238405Sjkim *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or 1671238405Sjkim EVP_CIPHER structures to avoid later problems in EVP_cleanup(). 1672238405Sjkim [Steve Henson] 1673238405Sjkim 1674238405Sjkim *) New utilities pkey and pkeyparam. These are similar to algorithm specific 1675238405Sjkim utilities such as rsa, dsa, dsaparam etc except they process any key 1676238405Sjkim type. 1677238405Sjkim [Steve Henson] 1678238405Sjkim 1679238405Sjkim *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New 1680238405Sjkim functions EVP_PKEY_print_public(), EVP_PKEY_print_private(), 1681238405Sjkim EVP_PKEY_print_param() to print public key data from an EVP_PKEY 1682238405Sjkim structure. 1683238405Sjkim [Steve Henson] 1684238405Sjkim 1685238405Sjkim *) Initial support for pluggable public key ASN1. 1686238405Sjkim De-spaghettify the public key ASN1 handling. Move public and private 1687238405Sjkim key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate 1688238405Sjkim algorithm specific handling to a single module within the relevant 1689238405Sjkim algorithm directory. Add functions to allow (near) opaque processing 1690238405Sjkim of public and private key structures. 1691238405Sjkim [Steve Henson] 1692238405Sjkim 1693238405Sjkim *) Implement the Supported Point Formats Extension for 1694238405Sjkim ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1695238405Sjkim [Douglas Stebila] 1696238405Sjkim 1697238405Sjkim *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members 1698238405Sjkim for the psk identity [hint] and the psk callback functions to the 1699238405Sjkim SSL_SESSION, SSL and SSL_CTX structure. 1700238405Sjkim 1701238405Sjkim New ciphersuites: 1702238405Sjkim PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA, 1703238405Sjkim PSK-AES256-CBC-SHA 1704238405Sjkim 1705238405Sjkim New functions: 1706238405Sjkim SSL_CTX_use_psk_identity_hint 1707238405Sjkim SSL_get_psk_identity_hint 1708238405Sjkim SSL_get_psk_identity 1709238405Sjkim SSL_use_psk_identity_hint 1710238405Sjkim 1711238405Sjkim [Mika Kousa and Pasi Eronen of Nokia Corporation] 1712238405Sjkim 1713238405Sjkim *) Add RFC 3161 compliant time stamp request creation, response generation 1714238405Sjkim and response verification functionality. 1715238405Sjkim [Zolt�n Gl�zik <zglozik@opentsa.org>, The OpenTSA Project] 1716238405Sjkim 1717238405Sjkim *) Add initial support for TLS extensions, specifically for the server_name 1718238405Sjkim extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 1719238405Sjkim have new members for a host name. The SSL data structure has an 1720238405Sjkim additional member SSL_CTX *initial_ctx so that new sessions can be 1721238405Sjkim stored in that context to allow for session resumption, even after the 1722238405Sjkim SSL has been switched to a new SSL_CTX in reaction to a client's 1723238405Sjkim server_name extension. 1724238405Sjkim 1725238405Sjkim New functions (subject to change): 1726238405Sjkim 1727238405Sjkim SSL_get_servername() 1728238405Sjkim SSL_get_servername_type() 1729238405Sjkim SSL_set_SSL_CTX() 1730238405Sjkim 1731238405Sjkim New CTRL codes and macros (subject to change): 1732238405Sjkim 1733238405Sjkim SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 1734238405Sjkim - SSL_CTX_set_tlsext_servername_callback() 1735238405Sjkim SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 1736238405Sjkim - SSL_CTX_set_tlsext_servername_arg() 1737238405Sjkim SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 1738238405Sjkim 1739238405Sjkim openssl s_client has a new '-servername ...' option. 1740238405Sjkim 1741238405Sjkim openssl s_server has new options '-servername_host ...', '-cert2 ...', 1742238405Sjkim '-key2 ...', '-servername_fatal' (subject to change). This allows 1743238405Sjkim testing the HostName extension for a specific single host name ('-cert' 1744238405Sjkim and '-key' remain fallbacks for handshakes without HostName 1745238405Sjkim negotiation). If the unrecogninzed_name alert has to be sent, this by 1746238405Sjkim default is a warning; it becomes fatal with the '-servername_fatal' 1747238405Sjkim option. 1748238405Sjkim 1749238405Sjkim [Peter Sylvester, Remy Allais, Christophe Renou] 1750238405Sjkim 1751238405Sjkim *) Whirlpool hash implementation is added. 1752238405Sjkim [Andy Polyakov] 1753238405Sjkim 1754238405Sjkim *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to 1755238405Sjkim bn(64,32). Because of instruction set limitations it doesn't have 1756238405Sjkim any negative impact on performance. This was done mostly in order 1757238405Sjkim to make it possible to share assembler modules, such as bn_mul_mont 1758238405Sjkim implementations, between 32- and 64-bit builds without hassle. 1759238405Sjkim [Andy Polyakov] 1760238405Sjkim 1761238405Sjkim *) Move code previously exiled into file crypto/ec/ec2_smpt.c 1762238405Sjkim to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP 1763238405Sjkim macro. 1764238405Sjkim [Bodo Moeller] 1765238405Sjkim 1766238405Sjkim *) New candidate for BIGNUM assembler implementation, bn_mul_mont, 1767238405Sjkim dedicated Montgomery multiplication procedure, is introduced. 1768238405Sjkim BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher 1769238405Sjkim "64-bit" performance on certain 32-bit targets. 1770238405Sjkim [Andy Polyakov] 1771238405Sjkim 1772238405Sjkim *) New option SSL_OP_NO_COMP to disable use of compression selectively 1773238405Sjkim in SSL structures. New SSL ctrl to set maximum send fragment size. 1774238405Sjkim Save memory by seeting the I/O buffer sizes dynamically instead of 1775238405Sjkim using the maximum available value. 1776238405Sjkim [Steve Henson] 1777238405Sjkim 1778238405Sjkim *) New option -V for 'openssl ciphers'. This prints the ciphersuite code 1779238405Sjkim in addition to the text details. 1780238405Sjkim [Bodo Moeller] 1781238405Sjkim 1782238405Sjkim *) Very, very preliminary EXPERIMENTAL support for printing of general 1783238405Sjkim ASN1 structures. This currently produces rather ugly output and doesn't 1784238405Sjkim handle several customised structures at all. 1785238405Sjkim [Steve Henson] 1786238405Sjkim 1787238405Sjkim *) Integrated support for PVK file format and some related formats such 1788238405Sjkim as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support 1789238405Sjkim these in the 'rsa' and 'dsa' utilities. 1790238405Sjkim [Steve Henson] 1791238405Sjkim 1792238405Sjkim *) Support for PKCS#1 RSAPublicKey format on rsa utility command line. 1793238405Sjkim [Steve Henson] 1794238405Sjkim 1795238405Sjkim *) Remove the ancient ASN1_METHOD code. This was only ever used in one 1796238405Sjkim place for the (very old) "NETSCAPE" format certificates which are now 1797238405Sjkim handled using new ASN1 code equivalents. 1798238405Sjkim [Steve Henson] 1799238405Sjkim 1800238405Sjkim *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD 1801238405Sjkim pointer and make the SSL_METHOD parameter in SSL_CTX_new, 1802238405Sjkim SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'. 1803238405Sjkim [Nils Larsch] 1804238405Sjkim 1805238405Sjkim *) Modify CRL distribution points extension code to print out previously 1806238405Sjkim unsupported fields. Enhance extension setting code to allow setting of 1807238405Sjkim all fields. 1808238405Sjkim [Steve Henson] 1809238405Sjkim 1810238405Sjkim *) Add print and set support for Issuing Distribution Point CRL extension. 1811238405Sjkim [Steve Henson] 1812238405Sjkim 1813238405Sjkim *) Change 'Configure' script to enable Camellia by default. 1814238405Sjkim [NTT] 1815238405Sjkim 1816264331Sjkim Changes between 0.9.8x and 0.9.8y [5 Feb 2013] 1817264331Sjkim 1818264331Sjkim *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 1819264331Sjkim 1820264331Sjkim This addresses the flaw in CBC record processing discovered by 1821264331Sjkim Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 1822264331Sjkim at: http://www.isg.rhul.ac.uk/tls/ 1823264331Sjkim 1824264331Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 1825264331Sjkim Security Group at Royal Holloway, University of London 1826264331Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 1827264331Sjkim Emilia K�sper for the initial patch. 1828264331Sjkim (CVE-2013-0169) 1829264331Sjkim [Emilia K�sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 1830264331Sjkim 1831264331Sjkim *) Return an error when checking OCSP signatures when key is NULL. 1832264331Sjkim This fixes a DoS attack. (CVE-2013-0166) 1833264331Sjkim [Steve Henson] 1834264331Sjkim 1835264331Sjkim *) Call OCSP Stapling callback after ciphersuite has been chosen, so 1836264331Sjkim the right response is stapled. Also change SSL_get_certificate() 1837264331Sjkim so it returns the certificate actually sent. 1838264331Sjkim See http://rt.openssl.org/Ticket/Display.html?id=2836. 1839264331Sjkim (This is a backport) 1840264331Sjkim [Rob Stradling <rob.stradling@comodo.com>] 1841264331Sjkim 1842264331Sjkim *) Fix possible deadlock when decoding public keys. 1843264331Sjkim [Steve Henson] 1844264331Sjkim 1845264331Sjkim Changes between 0.9.8w and 0.9.8x [10 May 2012] 1846264331Sjkim 1847264331Sjkim *) Sanity check record length before skipping explicit IV in DTLS 1848264331Sjkim to fix DoS attack. 1849264331Sjkim 1850264331Sjkim Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 1851264331Sjkim fuzzing as a service testing platform. 1852264331Sjkim (CVE-2012-2333) 1853264331Sjkim [Steve Henson] 1854264331Sjkim 1855264331Sjkim *) Initialise tkeylen properly when encrypting CMS messages. 1856264331Sjkim Thanks to Solar Designer of Openwall for reporting this issue. 1857264331Sjkim [Steve Henson] 1858264331Sjkim 1859264331Sjkim Changes between 0.9.8v and 0.9.8w [23 Apr 2012] 1860264331Sjkim 1861264331Sjkim *) The fix for CVE-2012-2110 did not take into account that the 1862264331Sjkim 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an 1863264331Sjkim int in OpenSSL 0.9.8, making it still vulnerable. Fix by 1864264331Sjkim rejecting negative len parameter. (CVE-2012-2131) 1865264331Sjkim [Tomas Hoger <thoger@redhat.com>] 1866264331Sjkim 1867264331Sjkim Changes between 0.9.8u and 0.9.8v [19 Apr 2012] 1868264331Sjkim 1869264331Sjkim *) Check for potentially exploitable overflows in asn1_d2i_read_bio 1870264331Sjkim BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 1871264331Sjkim in CRYPTO_realloc_clean. 1872264331Sjkim 1873264331Sjkim Thanks to Tavis Ormandy, Google Security Team, for discovering this 1874264331Sjkim issue and to Adam Langley <agl@chromium.org> for fixing it. 1875264331Sjkim (CVE-2012-2110) 1876264331Sjkim [Adam Langley (Google), Tavis Ormandy, Google Security Team] 1877264331Sjkim 1878264331Sjkim Changes between 0.9.8t and 0.9.8u [12 Mar 2012] 1879264331Sjkim 1880264331Sjkim *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 1881264331Sjkim in CMS and PKCS7 code. When RSA decryption fails use a random key for 1882264331Sjkim content decryption and always return the same error. Note: this attack 1883264331Sjkim needs on average 2^20 messages so it only affects automated senders. The 1884264331Sjkim old behaviour can be reenabled in the CMS code by setting the 1885264331Sjkim CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 1886264331Sjkim an MMA defence is not necessary. 1887264331Sjkim Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 1888264331Sjkim this issue. (CVE-2012-0884) 1889264331Sjkim [Steve Henson] 1890264331Sjkim 1891264331Sjkim *) Fix CVE-2011-4619: make sure we really are receiving a 1892264331Sjkim client hello before rejecting multiple SGC restarts. Thanks to 1893264331Sjkim Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 1894264331Sjkim [Steve Henson] 1895264331Sjkim 1896237657Sjkim Changes between 0.9.8s and 0.9.8t [18 Jan 2012] 1897237657Sjkim 1898237657Sjkim *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 1899237657Sjkim Thanks to Antonio Martin, Enterprise Secure Access Research and 1900237657Sjkim Development, Cisco Systems, Inc. for discovering this bug and 1901237657Sjkim preparing a fix. (CVE-2012-0050) 1902237657Sjkim [Antonio Martin] 1903264331Sjkim 1904237657Sjkim Changes between 0.9.8r and 0.9.8s [4 Jan 2012] 1905237657Sjkim 1906237657Sjkim *) Nadhem Alfardan and Kenny Paterson have discovered an extension 1907237657Sjkim of the Vaudenay padding oracle attack on CBC mode encryption 1908237657Sjkim which enables an efficient plaintext recovery attack against 1909237657Sjkim the OpenSSL implementation of DTLS. Their attack exploits timing 1910237657Sjkim differences arising during decryption processing. A research 1911237657Sjkim paper describing this attack can be found at: 1912237657Sjkim http://www.isg.rhul.ac.uk/~kp/dtls.pdf 1913237657Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 1914237657Sjkim Security Group at Royal Holloway, University of London 1915237657Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 1916237657Sjkim <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 1917237657Sjkim for preparing the fix. (CVE-2011-4108) 1918237657Sjkim [Robin Seggelmann, Michael Tuexen] 1919237657Sjkim 1920237657Sjkim *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109) 1921237657Sjkim [Ben Laurie, Kasper <ekasper@google.com>] 1922237657Sjkim 1923237657Sjkim *) Clear bytes used for block padding of SSL 3.0 records. 1924237657Sjkim (CVE-2011-4576) 1925237657Sjkim [Adam Langley (Google)] 1926237657Sjkim 1927237657Sjkim *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 1928237657Sjkim Kadianakis <desnacked@gmail.com> for discovering this issue and 1929237657Sjkim Adam Langley for preparing the fix. (CVE-2011-4619) 1930237657Sjkim [Adam Langley (Google)] 1931237657Sjkim 1932237657Sjkim *) Prevent malformed RFC3779 data triggering an assertion failure. 1933237657Sjkim Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 1934237657Sjkim and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 1935237657Sjkim [Rob Austein <sra@hactrn.net>] 1936237657Sjkim 1937237657Sjkim *) Fix ssl_ciph.c set-up race. 1938237657Sjkim [Adam Langley (Google)] 1939237657Sjkim 1940237657Sjkim *) Fix spurious failures in ecdsatest.c. 1941237657Sjkim [Emilia K�sper (Google)] 1942237657Sjkim 1943237657Sjkim *) Fix the BIO_f_buffer() implementation (which was mixing different 1944237657Sjkim interpretations of the '..._len' fields). 1945237657Sjkim [Adam Langley (Google)] 1946237657Sjkim 1947237657Sjkim *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 1948237657Sjkim BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 1949237657Sjkim threads won't reuse the same blinding coefficients. 1950237657Sjkim 1951237657Sjkim This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 1952237657Sjkim lock to call BN_BLINDING_invert_ex, and avoids one use of 1953237657Sjkim BN_BLINDING_update for each BN_BLINDING structure (previously, 1954237657Sjkim the last update always remained unused). 1955237657Sjkim [Emilia K�sper (Google)] 1956237657Sjkim 1957237657Sjkim *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 1958237657Sjkim for multi-threaded use of ECDH. 1959237657Sjkim [Adam Langley (Google)] 1960237657Sjkim 1961237657Sjkim *) Fix x509_name_ex_d2i memory leak on bad inputs. 1962237657Sjkim [Bodo Moeller] 1963237657Sjkim 1964237657Sjkim *) Add protection against ECDSA timing attacks as mentioned in the paper 1965237657Sjkim by Billy Bob Brumley and Nicola Tuveri, see: 1966237657Sjkim 1967237657Sjkim http://eprint.iacr.org/2011/232.pdf 1968237657Sjkim 1969237657Sjkim [Billy Bob Brumley and Nicola Tuveri] 1970237657Sjkim 1971237657Sjkim Changes between 0.9.8q and 0.9.8r [8 Feb 2011] 1972237657Sjkim 1973237657Sjkim *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 1974237657Sjkim [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 1975237657Sjkim 1976237657Sjkim *) Fix bug in string printing code: if *any* escaping is enabled we must 1977237657Sjkim escape the escape character (backslash) or the resulting string is 1978237657Sjkim ambiguous. 1979237657Sjkim [Steve Henson] 1980237657Sjkim 1981216166Ssimon Changes between 0.9.8p and 0.9.8q [2 Dec 2010] 1982216166Ssimon 1983216166Ssimon *) Disable code workaround for ancient and obsolete Netscape browsers 1984216166Ssimon and servers: an attacker can use it in a ciphersuite downgrade attack. 1985216166Ssimon Thanks to Martin Rex for discovering this bug. CVE-2010-4180 1986216166Ssimon [Steve Henson] 1987216166Ssimon 1988216166Ssimon *) Fixed J-PAKE implementation error, originally discovered by 1989216166Ssimon Sebastien Martini, further info and confirmation from Stefan 1990216166Ssimon Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 1991216166Ssimon [Ben Laurie] 1992216166Ssimon 1993215697Ssimon Changes between 0.9.8o and 0.9.8p [16 Nov 2010] 1994215697Ssimon 1995215697Ssimon *) Fix extension code to avoid race conditions which can result in a buffer 1996215697Ssimon overrun vulnerability: resumed sessions must not be modified as they can 1997215697Ssimon be shared by multiple threads. CVE-2010-3864 1998215697Ssimon [Steve Henson] 1999215697Ssimon 2000215697Ssimon *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939 2001215697Ssimon [Steve Henson] 2002215697Ssimon 2003215697Ssimon *) Don't reencode certificate when calculating signature: cache and use 2004215697Ssimon the original encoding instead. This makes signature verification of 2005215697Ssimon some broken encodings work correctly. 2006215697Ssimon [Steve Henson] 2007215697Ssimon 2008215697Ssimon *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT 2009215697Ssimon is also one of the inputs. 2010215697Ssimon [Emilia K�sper <emilia.kasper@esat.kuleuven.be> (Google)] 2011215697Ssimon 2012215697Ssimon *) Don't repeatedly append PBE algorithms to table if they already exist. 2013215697Ssimon Sort table on each new add. This effectively makes the table read only 2014215697Ssimon after all algorithms are added and subsequent calls to PKCS12_pbe_add 2015215697Ssimon etc are non-op. 2016215697Ssimon [Steve Henson] 2017215697Ssimon 2018215697Ssimon Changes between 0.9.8n and 0.9.8o [01 Jun 2010] 2019215697Ssimon 2020215697Ssimon [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after 2021215697Ssimon OpenSSL 1.0.0.] 2022215697Ssimon 2023215697Ssimon *) Correct a typo in the CMS ASN1 module which can result in invalid memory 2024215697Ssimon access or freeing data twice (CVE-2010-0742) 2025215697Ssimon [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>] 2026215697Ssimon 2027215697Ssimon *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more 2028215697Ssimon common in certificates and some applications which only call 2029215697Ssimon SSL_library_init and not OpenSSL_add_all_algorithms() will fail. 2030215697Ssimon [Steve Henson] 2031215697Ssimon 2032215697Ssimon *) VMS fixes: 2033215697Ssimon Reduce copying into .apps and .test in makevms.com 2034215697Ssimon Don't try to use blank CA certificate in CA.com 2035215697Ssimon Allow use of C files from original directories in maketests.com 2036215697Ssimon [Steven M. Schweda" <sms@antinode.info>] 2037215697Ssimon 2038206046Ssimon Changes between 0.9.8m and 0.9.8n [24 Mar 2010] 2039206046Ssimon 2040206046Ssimon *) When rejecting SSL/TLS records due to an incorrect version number, never 2041206046Ssimon update s->server with a new major version number. As of 2042206046Ssimon - OpenSSL 0.9.8m if 'short' is a 16-bit type, 2043206046Ssimon - OpenSSL 0.9.8f if 'short' is longer than 16 bits, 2044206046Ssimon the previous behavior could result in a read attempt at NULL when 2045206046Ssimon receiving specific incorrect SSL/TLS records once record payload 2046206046Ssimon protection is active. (CVE-2010-0740) 2047206046Ssimon [Bodo Moeller, Adam Langley <agl@chromium.org>] 2048206046Ssimon 2049206046Ssimon *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL 2050206046Ssimon could be crashed if the relevant tables were not present (e.g. chrooted). 2051206046Ssimon [Tomas Hoger <thoger@redhat.com>] 2052206046Ssimon 2053205128Ssimon Changes between 0.9.8l and 0.9.8m [25 Feb 2010] 2054205128Ssimon 2055205128Ssimon *) Always check bn_wexpend() return values for failure. (CVE-2009-3245) 2056205128Ssimon [Martin Olsson, Neel Mehta] 2057205128Ssimon 2058205128Ssimon *) Fix X509_STORE locking: Every 'objs' access requires a lock (to 2059205128Ssimon accommodate for stack sorting, always a write lock!). 2060205128Ssimon [Bodo Moeller] 2061205128Ssimon 2062205128Ssimon *) On some versions of WIN32 Heap32Next is very slow. This can cause 2063205128Ssimon excessive delays in the RAND_poll(): over a minute. As a workaround 2064205128Ssimon include a time check in the inner Heap32Next loop too. 2065205128Ssimon [Steve Henson] 2066205128Ssimon 2067205128Ssimon *) The code that handled flushing of data in SSL/TLS originally used the 2068205128Ssimon BIO_CTRL_INFO ctrl to see if any data was pending first. This caused 2069205128Ssimon the problem outlined in PR#1949. The fix suggested there however can 2070205128Ssimon trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions 2071205128Ssimon of Apache). So instead simplify the code to flush unconditionally. 2072205128Ssimon This should be fine since flushing with no data to flush is a no op. 2073205128Ssimon [Steve Henson] 2074205128Ssimon 2075205128Ssimon *) Handle TLS versions 2.0 and later properly and correctly use the 2076205128Ssimon highest version of TLS/SSL supported. Although TLS >= 2.0 is some way 2077205128Ssimon off ancient servers have a habit of sticking around for a while... 2078205128Ssimon [Steve Henson] 2079205128Ssimon 2080205128Ssimon *) Modify compression code so it frees up structures without using the 2081205128Ssimon ex_data callbacks. This works around a problem where some applications 2082205128Ssimon call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when 2083205128Ssimon restarting) then use compression (e.g. SSL with compression) later. 2084205128Ssimon This results in significant per-connection memory leaks and 2085205128Ssimon has caused some security issues including CVE-2008-1678 and 2086205128Ssimon CVE-2009-4355. 2087205128Ssimon [Steve Henson] 2088205128Ssimon 2089205128Ssimon *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't 2090205128Ssimon change when encrypting or decrypting. 2091205128Ssimon [Bodo Moeller] 2092205128Ssimon 2093205128Ssimon *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to 2094205128Ssimon connect and renegotiate with servers which do not support RI. 2095205128Ssimon Until RI is more widely deployed this option is enabled by default. 2096205128Ssimon [Steve Henson] 2097205128Ssimon 2098205128Ssimon *) Add "missing" ssl ctrls to clear options and mode. 2099205128Ssimon [Steve Henson] 2100205128Ssimon 2101205128Ssimon *) If client attempts to renegotiate and doesn't support RI respond with 2102205128Ssimon a no_renegotiation alert as required by RFC5746. Some renegotiating 2103205128Ssimon TLS clients will continue a connection gracefully when they receive 2104205128Ssimon the alert. Unfortunately OpenSSL mishandled this alert and would hang 2105205128Ssimon waiting for a server hello which it will never receive. Now we treat a 2106205128Ssimon received no_renegotiation alert as a fatal error. This is because 2107205128Ssimon applications requesting a renegotiation might well expect it to succeed 2108205128Ssimon and would have no code in place to handle the server denying it so the 2109205128Ssimon only safe thing to do is to terminate the connection. 2110205128Ssimon [Steve Henson] 2111205128Ssimon 2112205128Ssimon *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if 2113205128Ssimon peer supports secure renegotiation and 0 otherwise. Print out peer 2114205128Ssimon renegotiation support in s_client/s_server. 2115205128Ssimon [Steve Henson] 2116205128Ssimon 2117205128Ssimon *) Replace the highly broken and deprecated SPKAC certification method with 2118205128Ssimon the updated NID creation version. This should correctly handle UTF8. 2119205128Ssimon [Steve Henson] 2120205128Ssimon 2121205128Ssimon *) Implement RFC5746. Re-enable renegotiation but require the extension 2122205128Ssimon as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION 2123205128Ssimon turns out to be a bad idea. It has been replaced by 2124205128Ssimon SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with 2125205128Ssimon SSL_CTX_set_options(). This is really not recommended unless you 2126205128Ssimon know what you are doing. 2127205128Ssimon [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson] 2128205128Ssimon 2129205128Ssimon *) Fixes to stateless session resumption handling. Use initial_ctx when 2130205128Ssimon issuing and attempting to decrypt tickets in case it has changed during 2131205128Ssimon servername handling. Use a non-zero length session ID when attempting 2132205128Ssimon stateless session resumption: this makes it possible to determine if 2133205128Ssimon a resumption has occurred immediately after receiving server hello 2134205128Ssimon (several places in OpenSSL subtly assume this) instead of later in 2135205128Ssimon the handshake. 2136205128Ssimon [Steve Henson] 2137205128Ssimon 2138205128Ssimon *) The functions ENGINE_ctrl(), OPENSSL_isservice(), 2139205128Ssimon CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error 2140205128Ssimon fixes for a few places where the return code is not checked 2141205128Ssimon correctly. 2142205128Ssimon [Julia Lawall <julia@diku.dk>] 2143205128Ssimon 2144205128Ssimon *) Add --strict-warnings option to Configure script to include devteam 2145205128Ssimon warnings in other configurations. 2146205128Ssimon [Steve Henson] 2147205128Ssimon 2148205128Ssimon *) Add support for --libdir option and LIBDIR variable in makefiles. This 2149205128Ssimon makes it possible to install openssl libraries in locations which 2150205128Ssimon have names other than "lib", for example "/usr/lib64" which some 2151205128Ssimon systems need. 2152205128Ssimon [Steve Henson, based on patch from Jeremy Utley] 2153205128Ssimon 2154205128Ssimon *) Don't allow the use of leading 0x80 in OIDs. This is a violation of 2155205128Ssimon X690 8.9.12 and can produce some misleading textual output of OIDs. 2156205128Ssimon [Steve Henson, reported by Dan Kaminsky] 2157205128Ssimon 2158205128Ssimon *) Delete MD2 from algorithm tables. This follows the recommendation in 2159205128Ssimon several standards that it is not used in new applications due to 2160205128Ssimon several cryptographic weaknesses. For binary compatibility reasons 2161205128Ssimon the MD2 API is still compiled in by default. 2162205128Ssimon [Steve Henson] 2163205128Ssimon 2164205128Ssimon *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved 2165205128Ssimon and restored. 2166205128Ssimon [Steve Henson] 2167205128Ssimon 2168205128Ssimon *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and 2169205128Ssimon OPENSSL_asc2uni conditionally on Netware platforms to avoid a name 2170205128Ssimon clash. 2171205128Ssimon [Guenter <lists@gknw.net>] 2172205128Ssimon 2173205128Ssimon *) Fix the server certificate chain building code to use X509_verify_cert(), 2174205128Ssimon it used to have an ad-hoc builder which was unable to cope with anything 2175205128Ssimon other than a simple chain. 2176205128Ssimon [David Woodhouse <dwmw2@infradead.org>, Steve Henson] 2177205128Ssimon 2178205128Ssimon *) Don't check self signed certificate signatures in X509_verify_cert() 2179205128Ssimon by default (a flag can override this): it just wastes time without 2180205128Ssimon adding any security. As a useful side effect self signed root CAs 2181205128Ssimon with non-FIPS digests are now usable in FIPS mode. 2182205128Ssimon [Steve Henson] 2183205128Ssimon 2184205128Ssimon *) In dtls1_process_out_of_seq_message() the check if the current message 2185205128Ssimon is already buffered was missing. For every new message was memory 2186205128Ssimon allocated, allowing an attacker to perform an denial of service attack 2187205128Ssimon with sending out of seq handshake messages until there is no memory 2188205128Ssimon left. Additionally every future messege was buffered, even if the 2189205128Ssimon sequence number made no sense and would be part of another handshake. 2190205128Ssimon So only messages with sequence numbers less than 10 in advance will be 2191205128Ssimon buffered. (CVE-2009-1378) 2192205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 2193205128Ssimon 2194205128Ssimon *) Records are buffered if they arrive with a future epoch to be 2195205128Ssimon processed after finishing the corresponding handshake. There is 2196205128Ssimon currently no limitation to this buffer allowing an attacker to perform 2197205128Ssimon a DOS attack with sending records with future epochs until there is no 2198205128Ssimon memory left. This patch adds the pqueue_size() function to detemine 2199205128Ssimon the size of a buffer and limits the record buffer to 100 entries. 2200205128Ssimon (CVE-2009-1377) 2201205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 2202205128Ssimon 2203205128Ssimon *) Keep a copy of frag->msg_header.frag_len so it can be used after the 2204205128Ssimon parent structure is freed. (CVE-2009-1379) 2205205128Ssimon [Daniel Mentz] 2206205128Ssimon 2207205128Ssimon *) Handle non-blocking I/O properly in SSL_shutdown() call. 2208205128Ssimon [Darryl Miles <darryl-mailinglists@netbauds.net>] 2209205128Ssimon 2210205128Ssimon *) Add 2.5.4.* OIDs 2211205128Ssimon [Ilya O. <vrghost@gmail.com>] 2212205128Ssimon 2213205128Ssimon Changes between 0.9.8k and 0.9.8l [5 Nov 2009] 2214205128Ssimon 2215205128Ssimon *) Disable renegotiation completely - this fixes a severe security 2216205128Ssimon problem (CVE-2009-3555) at the cost of breaking all 2217205128Ssimon renegotiation. Renegotiation can be re-enabled by setting 2218205128Ssimon SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at 2219205128Ssimon run-time. This is really not recommended unless you know what 2220205128Ssimon you're doing. 2221205128Ssimon [Ben Laurie] 2222205128Ssimon 2223194206Ssimon Changes between 0.9.8j and 0.9.8k [25 Mar 2009] 2224167612Ssimon 2225194206Ssimon *) Don't set val to NULL when freeing up structures, it is freed up by 2226194206Ssimon underlying code. If sizeof(void *) > sizeof(long) this can result in 2227194206Ssimon zeroing past the valid field. (CVE-2009-0789) 2228194206Ssimon [Paolo Ganci <Paolo.Ganci@AdNovum.CH>] 2229194206Ssimon 2230194206Ssimon *) Fix bug where return value of CMS_SignerInfo_verify_content() was not 2231194206Ssimon checked correctly. This would allow some invalid signed attributes to 2232194206Ssimon appear to verify correctly. (CVE-2009-0591) 2233194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 2234194206Ssimon 2235194206Ssimon *) Reject UniversalString and BMPString types with invalid lengths. This 2236194206Ssimon prevents a crash in ASN1_STRING_print_ex() which assumes the strings have 2237194206Ssimon a legal length. (CVE-2009-0590) 2238194206Ssimon [Steve Henson] 2239194206Ssimon 2240194206Ssimon *) Set S/MIME signing as the default purpose rather than setting it 2241194206Ssimon unconditionally. This allows applications to override it at the store 2242194206Ssimon level. 2243194206Ssimon [Steve Henson] 2244194206Ssimon 2245194206Ssimon *) Permit restricted recursion of ASN1 strings. This is needed in practice 2246194206Ssimon to handle some structures. 2247194206Ssimon [Steve Henson] 2248194206Ssimon 2249194206Ssimon *) Improve efficiency of mem_gets: don't search whole buffer each time 2250194206Ssimon for a '\n' 2251194206Ssimon [Jeremy Shapiro <jnshapir@us.ibm.com>] 2252194206Ssimon 2253194206Ssimon *) New -hex option for openssl rand. 2254194206Ssimon [Matthieu Herrb] 2255194206Ssimon 2256194206Ssimon *) Print out UTF8String and NumericString when parsing ASN1. 2257194206Ssimon [Steve Henson] 2258194206Ssimon 2259194206Ssimon *) Support NumericString type for name components. 2260194206Ssimon [Steve Henson] 2261194206Ssimon 2262194206Ssimon *) Allow CC in the environment to override the automatically chosen 2263194206Ssimon compiler. Note that nothing is done to ensure flags work with the 2264194206Ssimon chosen compiler. 2265194206Ssimon [Ben Laurie] 2266194206Ssimon 2267194206Ssimon Changes between 0.9.8i and 0.9.8j [07 Jan 2009] 2268194206Ssimon 2269194206Ssimon *) Properly check EVP_VerifyFinal() and similar return values 2270194206Ssimon (CVE-2008-5077). 2271194206Ssimon [Ben Laurie, Bodo Moeller, Google Security Team] 2272194206Ssimon 2273194206Ssimon *) Enable TLS extensions by default. 2274194206Ssimon [Ben Laurie] 2275194206Ssimon 2276194206Ssimon *) Allow the CHIL engine to be loaded, whether the application is 2277194206Ssimon multithreaded or not. (This does not release the developer from the 2278194206Ssimon obligation to set up the dynamic locking callbacks.) 2279194206Ssimon [Sander Temme <sander@temme.net>] 2280194206Ssimon 2281194206Ssimon *) Use correct exit code if there is an error in dgst command. 2282194206Ssimon [Steve Henson; problem pointed out by Roland Dirlewanger] 2283194206Ssimon 2284194206Ssimon *) Tweak Configure so that you need to say "experimental-jpake" to enable 2285194206Ssimon JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. 2286194206Ssimon [Bodo Moeller] 2287194206Ssimon 2288194206Ssimon *) Add experimental JPAKE support, including demo authentication in 2289194206Ssimon s_client and s_server. 2290194206Ssimon [Ben Laurie] 2291194206Ssimon 2292194206Ssimon *) Set the comparison function in v3_addr_canonize(). 2293194206Ssimon [Rob Austein <sra@hactrn.net>] 2294194206Ssimon 2295194206Ssimon *) Add support for XMPP STARTTLS in s_client. 2296194206Ssimon [Philip Paeps <philip@freebsd.org>] 2297194206Ssimon 2298194206Ssimon *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior 2299194206Ssimon to ensure that even with this option, only ciphersuites in the 2300194206Ssimon server's preference list will be accepted. (Note that the option 2301194206Ssimon applies only when resuming a session, so the earlier behavior was 2302194206Ssimon just about the algorithm choice for symmetric cryptography.) 2303194206Ssimon [Bodo Moeller] 2304194206Ssimon 2305194206Ssimon Changes between 0.9.8h and 0.9.8i [15 Sep 2008] 2306194206Ssimon 2307205128Ssimon *) Fix NULL pointer dereference if a DTLS server received 2308205128Ssimon ChangeCipherSpec as first record (CVE-2009-1386). 2309205128Ssimon [PR #1679] 2310205128Ssimon 2311194206Ssimon *) Fix a state transitition in s3_srvr.c and d1_srvr.c 2312194206Ssimon (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...). 2313194206Ssimon [Nagendra Modadugu] 2314194206Ssimon 2315194206Ssimon *) The fix in 0.9.8c that supposedly got rid of unsafe 2316194206Ssimon double-checked locking was incomplete for RSA blinding, 2317194206Ssimon addressing just one layer of what turns out to have been 2318194206Ssimon doubly unsafe triple-checked locking. 2319194206Ssimon 2320194206Ssimon So now fix this for real by retiring the MONT_HELPER macro 2321194206Ssimon in crypto/rsa/rsa_eay.c. 2322194206Ssimon 2323194206Ssimon [Bodo Moeller; problem pointed out by Marius Schilder] 2324194206Ssimon 2325194206Ssimon *) Various precautionary measures: 2326194206Ssimon 2327194206Ssimon - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). 2328194206Ssimon 2329194206Ssimon - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). 2330194206Ssimon (NB: This would require knowledge of the secret session ticket key 2331194206Ssimon to exploit, in which case you'd be SOL either way.) 2332194206Ssimon 2333194206Ssimon - Change bn_nist.c so that it will properly handle input BIGNUMs 2334194206Ssimon outside the expected range. 2335194206Ssimon 2336194206Ssimon - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG 2337194206Ssimon builds. 2338194206Ssimon 2339194206Ssimon [Neel Mehta, Bodo Moeller] 2340194206Ssimon 2341194206Ssimon *) Allow engines to be "soft loaded" - i.e. optionally don't die if 2342194206Ssimon the load fails. Useful for distros. 2343194206Ssimon [Ben Laurie and the FreeBSD team] 2344194206Ssimon 2345194206Ssimon *) Add support for Local Machine Keyset attribute in PKCS#12 files. 2346194206Ssimon [Steve Henson] 2347194206Ssimon 2348194206Ssimon *) Fix BN_GF2m_mod_arr() top-bit cleanup code. 2349194206Ssimon [Huang Ying] 2350194206Ssimon 2351194206Ssimon *) Expand ENGINE to support engine supplied SSL client certificate functions. 2352194206Ssimon 2353194206Ssimon This work was sponsored by Logica. 2354194206Ssimon [Steve Henson] 2355194206Ssimon 2356194206Ssimon *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows 2357194206Ssimon keystores. Support for SSL/TLS client authentication too. 2358194206Ssimon Not compiled unless enable-capieng specified to Configure. 2359194206Ssimon 2360194206Ssimon This work was sponsored by Logica. 2361194206Ssimon [Steve Henson] 2362194206Ssimon 2363194206Ssimon *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using 2364194206Ssimon ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain 2365194206Ssimon attribute creation routines such as certifcate requests and PKCS#12 2366194206Ssimon files. 2367194206Ssimon [Steve Henson] 2368194206Ssimon 2369194206Ssimon Changes between 0.9.8g and 0.9.8h [28 May 2008] 2370194206Ssimon 2371194206Ssimon *) Fix flaw if 'Server Key exchange message' is omitted from a TLS 2372194206Ssimon handshake which could lead to a cilent crash as found using the 2373194206Ssimon Codenomicon TLS test suite (CVE-2008-1672) 2374194206Ssimon [Steve Henson, Mark Cox] 2375194206Ssimon 2376194206Ssimon *) Fix double free in TLS server name extensions which could lead to 2377194206Ssimon a remote crash found by Codenomicon TLS test suite (CVE-2008-0891) 2378194206Ssimon [Joe Orton] 2379194206Ssimon 2380194206Ssimon *) Clear error queue in SSL_CTX_use_certificate_chain_file() 2381194206Ssimon 2382194206Ssimon Clear the error queue to ensure that error entries left from 2383194206Ssimon older function calls do not interfere with the correct operation. 2384194206Ssimon [Lutz Jaenicke, Erik de Castro Lopo] 2385194206Ssimon 2386194206Ssimon *) Remove root CA certificates of commercial CAs: 2387194206Ssimon 2388194206Ssimon The OpenSSL project does not recommend any specific CA and does not 2389194206Ssimon have any policy with respect to including or excluding any CA. 2390194206Ssimon Therefore it does not make any sense to ship an arbitrary selection 2391194206Ssimon of root CA certificates with the OpenSSL software. 2392194206Ssimon [Lutz Jaenicke] 2393194206Ssimon 2394194206Ssimon *) RSA OAEP patches to fix two separate invalid memory reads. 2395194206Ssimon The first one involves inputs when 'lzero' is greater than 2396194206Ssimon 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes 2397194206Ssimon before the beginning of from). The second one involves inputs where 2398194206Ssimon the 'db' section contains nothing but zeroes (there is a one-byte 2399194206Ssimon invalid read after the end of 'db'). 2400194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 2401194206Ssimon 2402194206Ssimon *) Partial backport from 0.9.9-dev: 2403194206Ssimon 2404194206Ssimon Introduce bn_mul_mont (dedicated Montgomery multiplication 2405194206Ssimon procedure) as a candidate for BIGNUM assembler implementation. 2406194206Ssimon While 0.9.9-dev uses assembler for various architectures, only 2407194206Ssimon x86_64 is available by default here in the 0.9.8 branch, and 2408194206Ssimon 32-bit x86 is available through a compile-time setting. 2409194206Ssimon 2410194206Ssimon To try the 32-bit x86 assembler implementation, use Configure 2411194206Ssimon option "enable-montasm" (which exists only for this backport). 2412194206Ssimon 2413194206Ssimon As "enable-montasm" for 32-bit x86 disclaims code stability 2414194206Ssimon anyway, in this constellation we activate additional code 2415194206Ssimon backported from 0.9.9-dev for further performance improvements, 2416194206Ssimon namely BN_from_montgomery_word. (To enable this otherwise, 2417194206Ssimon e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".) 2418194206Ssimon 2419194206Ssimon [Andy Polyakov (backport partially by Bodo Moeller)] 2420194206Ssimon 2421194206Ssimon *) Add TLS session ticket callback. This allows an application to set 2422194206Ssimon TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed 2423194206Ssimon values. This is useful for key rollover for example where several key 2424194206Ssimon sets may exist with different names. 2425194206Ssimon [Steve Henson] 2426194206Ssimon 2427194206Ssimon *) Reverse ENGINE-internal logic for caching default ENGINE handles. 2428194206Ssimon This was broken until now in 0.9.8 releases, such that the only way 2429194206Ssimon a registered ENGINE could be used (assuming it initialises 2430194206Ssimon successfully on the host) was to explicitly set it as the default 2431194206Ssimon for the relevant algorithms. This is in contradiction with 0.9.7 2432194206Ssimon behaviour and the documentation. With this fix, when an ENGINE is 2433194206Ssimon registered into a given algorithm's table of implementations, the 2434194206Ssimon 'uptodate' flag is reset so that auto-discovery will be used next 2435194206Ssimon time a new context for that algorithm attempts to select an 2436194206Ssimon implementation. 2437194206Ssimon [Ian Lister (tweaked by Geoff Thorpe)] 2438194206Ssimon 2439194206Ssimon *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 2440194206Ssimon implemention in the following ways: 2441194206Ssimon 2442194206Ssimon Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be 2443194206Ssimon hard coded. 2444194206Ssimon 2445194206Ssimon Lack of BER streaming support means one pass streaming processing is 2446194206Ssimon only supported if data is detached: setting the streaming flag is 2447194206Ssimon ignored for embedded content. 2448194206Ssimon 2449194206Ssimon CMS support is disabled by default and must be explicitly enabled 2450194206Ssimon with the enable-cms configuration option. 2451194206Ssimon [Steve Henson] 2452194206Ssimon 2453194206Ssimon *) Update the GMP engine glue to do direct copies between BIGNUM and 2454194206Ssimon mpz_t when openssl and GMP use the same limb size. Otherwise the 2455194206Ssimon existing "conversion via a text string export" trick is still used. 2456194206Ssimon [Paul Sheer <paulsheer@gmail.com>] 2457194206Ssimon 2458194206Ssimon *) Zlib compression BIO. This is a filter BIO which compressed and 2459194206Ssimon uncompresses any data passed through it. 2460194206Ssimon [Steve Henson] 2461194206Ssimon 2462194206Ssimon *) Add AES_wrap_key() and AES_unwrap_key() functions to implement 2463194206Ssimon RFC3394 compatible AES key wrapping. 2464194206Ssimon [Steve Henson] 2465194206Ssimon 2466194206Ssimon *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): 2467194206Ssimon sets string data without copying. X509_ALGOR_set0() and 2468194206Ssimon X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) 2469194206Ssimon data. Attribute function X509at_get0_data_by_OBJ(): retrieves data 2470194206Ssimon from an X509_ATTRIBUTE structure optionally checking it occurs only 2471194206Ssimon once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied 2472194206Ssimon data. 2473194206Ssimon [Steve Henson] 2474194206Ssimon 2475194206Ssimon *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() 2476194206Ssimon to get the expected BN_FLG_CONSTTIME behavior. 2477194206Ssimon [Bodo Moeller (Google)] 2478194206Ssimon 2479194206Ssimon *) Netware support: 2480194206Ssimon 2481194206Ssimon - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets 2482194206Ssimon - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) 2483194206Ssimon - added some more tests to do_tests.pl 2484194206Ssimon - fixed RunningProcess usage so that it works with newer LIBC NDKs too 2485194206Ssimon - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency 2486194206Ssimon - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, 2487194206Ssimon netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc 2488194206Ssimon - various changes to netware.pl to enable gcc-cross builds on Win32 2489194206Ssimon platform 2490194206Ssimon - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) 2491194206Ssimon - various changes to fix missing prototype warnings 2492194206Ssimon - fixed x86nasm.pl to create correct asm files for NASM COFF output 2493194206Ssimon - added AES, WHIRLPOOL and CPUID assembler code to build files 2494194206Ssimon - added missing AES assembler make rules to mk1mf.pl 2495194206Ssimon - fixed order of includes in apps/ocsp.c so that e_os.h settings apply 2496194206Ssimon [Guenter Knauf <eflash@gmx.net>] 2497194206Ssimon 2498194206Ssimon *) Implement certificate status request TLS extension defined in RFC3546. 2499194206Ssimon A client can set the appropriate parameters and receive the encoded 2500194206Ssimon OCSP response via a callback. A server can query the supplied parameters 2501194206Ssimon and set the encoded OCSP response in the callback. Add simplified examples 2502194206Ssimon to s_client and s_server. 2503194206Ssimon [Steve Henson] 2504194206Ssimon 2505194206Ssimon Changes between 0.9.8f and 0.9.8g [19 Oct 2007] 2506194206Ssimon 2507194206Ssimon *) Fix various bugs: 2508194206Ssimon + Binary incompatibility of ssl_ctx_st structure 2509194206Ssimon + DTLS interoperation with non-compliant servers 2510194206Ssimon + Don't call get_session_cb() without proposed session 2511194206Ssimon + Fix ia64 assembler code 2512194206Ssimon [Andy Polyakov, Steve Henson] 2513194206Ssimon 2514194206Ssimon Changes between 0.9.8e and 0.9.8f [11 Oct 2007] 2515194206Ssimon 2516194206Ssimon *) DTLS Handshake overhaul. There were longstanding issues with 2517194206Ssimon OpenSSL DTLS implementation, which were making it impossible for 2518194206Ssimon RFC 4347 compliant client to communicate with OpenSSL server. 2519194206Ssimon Unfortunately just fixing these incompatibilities would "cut off" 2520194206Ssimon pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e 2521194206Ssimon server keeps tolerating non RFC compliant syntax. The opposite is 2522194206Ssimon not true, 0.9.8f client can not communicate with earlier server. 2523194206Ssimon This update even addresses CVE-2007-4995. 2524194206Ssimon [Andy Polyakov] 2525194206Ssimon 2526194206Ssimon *) Changes to avoid need for function casts in OpenSSL: some compilers 2527194206Ssimon (gcc 4.2 and later) reject their use. 2528194206Ssimon [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, 2529194206Ssimon Steve Henson] 2530194206Ssimon 2531194206Ssimon *) Add RFC4507 support to OpenSSL. This includes the corrections in 2532194206Ssimon RFC4507bis. The encrypted ticket format is an encrypted encoded 2533194206Ssimon SSL_SESSION structure, that way new session features are automatically 2534194206Ssimon supported. 2535194206Ssimon 2536194206Ssimon If a client application caches session in an SSL_SESSION structure 2537194206Ssimon support is transparent because tickets are now stored in the encoded 2538194206Ssimon SSL_SESSION. 2539194206Ssimon 2540194206Ssimon The SSL_CTX structure automatically generates keys for ticket 2541194206Ssimon protection in servers so again support should be possible 2542194206Ssimon with no application modification. 2543194206Ssimon 2544194206Ssimon If a client or server wishes to disable RFC4507 support then the option 2545194206Ssimon SSL_OP_NO_TICKET can be set. 2546194206Ssimon 2547194206Ssimon Add a TLS extension debugging callback to allow the contents of any client 2548194206Ssimon or server extensions to be examined. 2549194206Ssimon 2550194206Ssimon This work was sponsored by Google. 2551194206Ssimon [Steve Henson] 2552194206Ssimon 2553194206Ssimon *) Add initial support for TLS extensions, specifically for the server_name 2554194206Ssimon extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 2555194206Ssimon have new members for a host name. The SSL data structure has an 2556194206Ssimon additional member SSL_CTX *initial_ctx so that new sessions can be 2557194206Ssimon stored in that context to allow for session resumption, even after the 2558194206Ssimon SSL has been switched to a new SSL_CTX in reaction to a client's 2559194206Ssimon server_name extension. 2560194206Ssimon 2561194206Ssimon New functions (subject to change): 2562194206Ssimon 2563194206Ssimon SSL_get_servername() 2564194206Ssimon SSL_get_servername_type() 2565194206Ssimon SSL_set_SSL_CTX() 2566194206Ssimon 2567194206Ssimon New CTRL codes and macros (subject to change): 2568194206Ssimon 2569194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 2570194206Ssimon - SSL_CTX_set_tlsext_servername_callback() 2571194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 2572194206Ssimon - SSL_CTX_set_tlsext_servername_arg() 2573194206Ssimon SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 2574194206Ssimon 2575194206Ssimon openssl s_client has a new '-servername ...' option. 2576194206Ssimon 2577194206Ssimon openssl s_server has new options '-servername_host ...', '-cert2 ...', 2578194206Ssimon '-key2 ...', '-servername_fatal' (subject to change). This allows 2579194206Ssimon testing the HostName extension for a specific single host name ('-cert' 2580194206Ssimon and '-key' remain fallbacks for handshakes without HostName 2581194206Ssimon negotiation). If the unrecogninzed_name alert has to be sent, this by 2582194206Ssimon default is a warning; it becomes fatal with the '-servername_fatal' 2583194206Ssimon option. 2584194206Ssimon 2585194206Ssimon [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson] 2586194206Ssimon 2587194206Ssimon *) Add AES and SSE2 assembly language support to VC++ build. 2588194206Ssimon [Steve Henson] 2589194206Ssimon 2590194206Ssimon *) Mitigate attack on final subtraction in Montgomery reduction. 2591194206Ssimon [Andy Polyakov] 2592194206Ssimon 2593194206Ssimon *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 2594194206Ssimon (which previously caused an internal error). 2595194206Ssimon [Bodo Moeller] 2596194206Ssimon 2597194206Ssimon *) Squeeze another 10% out of IGE mode when in != out. 2598194206Ssimon [Ben Laurie] 2599194206Ssimon 2600194206Ssimon *) AES IGE mode speedup. 2601194206Ssimon [Dean Gaudet (Google)] 2602194206Ssimon 2603194206Ssimon *) Add the Korean symmetric 128-bit cipher SEED (see 2604194206Ssimon http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and 2605194206Ssimon add SEED ciphersuites from RFC 4162: 2606194206Ssimon 2607194206Ssimon TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" 2608194206Ssimon TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" 2609194206Ssimon TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" 2610194206Ssimon TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" 2611194206Ssimon 2612194206Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 2613194206Ssimon series, SEED remains excluded from compilation unless OpenSSL 2614194206Ssimon is configured with 'enable-seed'. 2615194206Ssimon [KISA, Bodo Moeller] 2616194206Ssimon 2617194206Ssimon *) Mitigate branch prediction attacks, which can be practical if a 2618194206Ssimon single processor is shared, allowing a spy process to extract 2619194206Ssimon information. For detailed background information, see 2620194206Ssimon http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron, 2621194206Ssimon J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL 2622194206Ssimon and Necessary Software Countermeasures"). The core of the change 2623194206Ssimon are new versions BN_div_no_branch() and 2624194206Ssimon BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), 2625194206Ssimon respectively, which are slower, but avoid the security-relevant 2626194206Ssimon conditional branches. These are automatically called by BN_div() 2627194206Ssimon and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one 2628194206Ssimon of the input BIGNUMs. Also, BN_is_bit_set() has been changed to 2629194206Ssimon remove a conditional branch. 2630194206Ssimon 2631194206Ssimon BN_FLG_CONSTTIME is the new name for the previous 2632194206Ssimon BN_FLG_EXP_CONSTTIME flag, since it now affects more than just 2633194206Ssimon modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag 2634194206Ssimon in the exponent causes BN_mod_exp_mont() to use the alternative 2635194206Ssimon implementation in BN_mod_exp_mont_consttime().) The old name 2636194206Ssimon remains as a deprecated alias. 2637194206Ssimon 2638194206Ssimon Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general 2639194206Ssimon RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses 2640194206Ssimon constant-time implementations for more than just exponentiation. 2641194206Ssimon Here too the old name is kept as a deprecated alias. 2642194206Ssimon 2643194206Ssimon BN_BLINDING_new() will now use BN_dup() for the modulus so that 2644194206Ssimon the BN_BLINDING structure gets an independent copy of the 2645194206Ssimon modulus. This means that the previous "BIGNUM *m" argument to 2646194206Ssimon BN_BLINDING_new() and to BN_BLINDING_create_param() now 2647194206Ssimon essentially becomes "const BIGNUM *m", although we can't actually 2648194206Ssimon change this in the header file before 0.9.9. It allows 2649194206Ssimon RSA_setup_blinding() to use BN_with_flags() on the modulus to 2650194206Ssimon enable BN_FLG_CONSTTIME. 2651194206Ssimon 2652194206Ssimon [Matthew D Wood (Intel Corp)] 2653194206Ssimon 2654194206Ssimon *) In the SSL/TLS server implementation, be strict about session ID 2655194206Ssimon context matching (which matters if an application uses a single 2656194206Ssimon external cache for different purposes). Previously, 2657194206Ssimon out-of-context reuse was forbidden only if SSL_VERIFY_PEER was 2658194206Ssimon set. This did ensure strict client verification, but meant that, 2659194206Ssimon with applications using a single external cache for quite 2660194206Ssimon different requirements, clients could circumvent ciphersuite 2661194206Ssimon restrictions for a given session ID context by starting a session 2662194206Ssimon in a different context. 2663194206Ssimon [Bodo Moeller] 2664194206Ssimon 2665167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 2666167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 2667167612Ssimon authentication-only ciphersuites. 2668167612Ssimon [Bodo Moeller] 2669167612Ssimon 2670194206Ssimon *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was 2671194206Ssimon not complete and could lead to a possible single byte overflow 2672194206Ssimon (CVE-2007-5135) [Ben Laurie] 2673194206Ssimon 2674194206Ssimon Changes between 0.9.8d and 0.9.8e [23 Feb 2007] 2675194206Ssimon 2676167612Ssimon *) Since AES128 and AES256 (and similarly Camellia128 and 2677167612Ssimon Camellia256) share a single mask bit in the logic of 2678167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 2679167612Ssimon kludge to work properly if AES128 is available and AES256 isn't 2680167612Ssimon (or if Camellia128 is available and Camellia256 isn't). 2681167612Ssimon [Victor Duchovni] 2682167612Ssimon 2683167612Ssimon *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c 2684167612Ssimon (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): 2685167612Ssimon When a point or a seed is encoded in a BIT STRING, we need to 2686167612Ssimon prevent the removal of trailing zero bits to get the proper DER 2687167612Ssimon encoding. (By default, crypto/asn1/a_bitstr.c assumes the case 2688167612Ssimon of a NamedBitList, for which trailing 0 bits need to be removed.) 2689167612Ssimon [Bodo Moeller] 2690167612Ssimon 2691167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 2692167612Ssimon protocol version while receiving ClientHello even if the 2693167612Ssimon ClientHello is fragmented. (The server can't insist on the 2694167612Ssimon particular protocol version it has chosen before the ServerHello 2695167612Ssimon message has informed the client about his choice.) 2696167612Ssimon [Bodo Moeller] 2697167612Ssimon 2698167612Ssimon *) Add RFC 3779 support. 2699167612Ssimon [Rob Austein for ARIN, Ben Laurie] 2700167612Ssimon 2701167612Ssimon *) Load error codes if they are not already present instead of using a 2702167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 2703167612Ssimon Improve header file function name parsing. 2704167612Ssimon [Steve Henson] 2705167612Ssimon 2706167612Ssimon *) extend SMTP and IMAP protocol emulation in s_client to use EHLO 2707167612Ssimon or CAPABILITY handshake as required by RFCs. 2708167612Ssimon [Goetz Babin-Ebell] 2709167612Ssimon 2710162911Ssimon Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 2711162911Ssimon 2712162911Ssimon *) Introduce limits to prevent malicious keys being able to 2713162911Ssimon cause a denial of service. (CVE-2006-2940) 2714162911Ssimon [Steve Henson, Bodo Moeller] 2715162911Ssimon 2716162911Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 2717162911Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 2718162911Ssimon 2719162911Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 2720162911Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 2721162911Ssimon 2722162911Ssimon *) Fix SSL client code which could crash if connecting to a 2723162911Ssimon malicious SSLv2 server. (CVE-2006-4343) 2724162911Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 2725162911Ssimon 2726162911Ssimon *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 2727162911Ssimon match only those. Before that, "AES256-SHA" would be interpreted 2728162911Ssimon as a pattern and match "AES128-SHA" too (since AES128-SHA got 2729162911Ssimon the same strength classification in 0.9.7h) as we currently only 2730162911Ssimon have a single AES bit in the ciphersuite description bitmap. 2731162911Ssimon That change, however, also applied to ciphersuite strings such as 2732162911Ssimon "RC4-MD5" that intentionally matched multiple ciphersuites -- 2733162911Ssimon namely, SSL 2.0 ciphersuites in addition to the more common ones 2734162911Ssimon from SSL 3.0/TLS 1.0. 2735162911Ssimon 2736162911Ssimon So we change the selection algorithm again: Naming an explicit 2737162911Ssimon ciphersuite selects this one ciphersuite, and any other similar 2738162911Ssimon ciphersuite (same bitmap) from *other* protocol versions. 2739162911Ssimon Thus, "RC4-MD5" again will properly select both the SSL 2.0 2740162911Ssimon ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 2741162911Ssimon 2742162911Ssimon Since SSL 2.0 does not have any ciphersuites for which the 2743162911Ssimon 128/256 bit distinction would be relevant, this works for now. 2744162911Ssimon The proper fix will be to use different bits for AES128 and 2745162911Ssimon AES256, which would have avoided the problems from the beginning; 2746162911Ssimon however, bits are scarce, so we can only do this in a new release 2747162911Ssimon (not just a patchlevel) when we can change the SSL_CIPHER 2748162911Ssimon definition to split the single 'unsigned long mask' bitmap into 2749162911Ssimon multiple values to extend the available space. 2750162911Ssimon 2751162911Ssimon [Bodo Moeller] 2752162911Ssimon 2753162911Ssimon Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 2754162911Ssimon 2755162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 2756162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 2757162911Ssimon 2758162911Ssimon *) Add AES IGE and biIGE modes. 2759162911Ssimon [Ben Laurie] 2760162911Ssimon 2761162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 2762162911Ssimon possible instead of select(), since the latter has some 2763162911Ssimon undesirable limitations. 2764162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 2765162911Ssimon 2766162911Ssimon *) Disable "ECCdraft" ciphersuites more thoroughly. Now special 2767162911Ssimon treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 2768162911Ssimon cannot be implicitly activated as part of, e.g., the "AES" alias. 2769162911Ssimon However, please upgrade to OpenSSL 0.9.9[-dev] for 2770162911Ssimon non-experimental use of the ECC ciphersuites to get TLS extension 2771162911Ssimon support, which is required for curve and point format negotiation 2772162911Ssimon to avoid potential handshake problems. 2773162911Ssimon [Bodo Moeller] 2774162911Ssimon 2775162911Ssimon *) Disable rogue ciphersuites: 2776162911Ssimon 2777162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 2778162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 2779162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 2780162911Ssimon 2781162911Ssimon The latter two were purportedly from 2782162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 2783162911Ssimon appear there. 2784162911Ssimon 2785167612Ssimon Also deactivate the remaining ciphersuites from 2786162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 2787162911Ssimon unofficial, and the ID has long expired. 2788162911Ssimon [Bodo Moeller] 2789162911Ssimon 2790162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 2791162911Ssimon dual-core machines) and other potential thread-safety issues. 2792162911Ssimon [Bodo Moeller] 2793162911Ssimon 2794162911Ssimon *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 2795162911Ssimon versions), which is now available for royalty-free use 2796162911Ssimon (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). 2797162911Ssimon Also, add Camellia TLS ciphersuites from RFC 4132. 2798162911Ssimon 2799162911Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 2800162911Ssimon series, Camellia remains excluded from compilation unless OpenSSL 2801162911Ssimon is configured with 'enable-camellia'. 2802162911Ssimon [NTT] 2803162911Ssimon 2804162911Ssimon *) Disable the padding bug check when compression is in use. The padding 2805162911Ssimon bug check assumes the first packet is of even length, this is not 2806162911Ssimon necessarily true if compresssion is enabled and can result in false 2807162911Ssimon positives causing handshake failure. The actual bug test is ancient 2808162911Ssimon code so it is hoped that implementations will either have fixed it by 2809162911Ssimon now or any which still have the bug do not support compression. 2810162911Ssimon [Steve Henson] 2811162911Ssimon 2812160814Ssimon Changes between 0.9.8a and 0.9.8b [04 May 2006] 2813160814Ssimon 2814160814Ssimon *) When applying a cipher rule check to see if string match is an explicit 2815160814Ssimon cipher suite and only match that one cipher suite if it is. 2816160814Ssimon [Steve Henson] 2817160814Ssimon 2818160814Ssimon *) Link in manifests for VC++ if needed. 2819160814Ssimon [Austin Ziegler <halostatue@gmail.com>] 2820160814Ssimon 2821160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 2822160814Ssimon draft-ietf-tls-ecc-12.txt with proposed changes (but without 2823160814Ssimon TLS extensions, which are supported starting with the 0.9.9 2824160814Ssimon branch, not in the OpenSSL 0.9.8 branch). 2825160814Ssimon [Douglas Stebila] 2826160814Ssimon 2827160814Ssimon *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 2828160814Ssimon opaque EVP_CIPHER_CTX handling. 2829160814Ssimon [Steve Henson] 2830160814Ssimon 2831160814Ssimon *) Fixes and enhancements to zlib compression code. We now only use 2832160814Ssimon "zlib1.dll" and use the default __cdecl calling convention on Win32 2833160814Ssimon to conform with the standards mentioned here: 2834160814Ssimon http://www.zlib.net/DLL_FAQ.txt 2835160814Ssimon Static zlib linking now works on Windows and the new --with-zlib-include 2836160814Ssimon --with-zlib-lib options to Configure can be used to supply the location 2837160814Ssimon of the headers and library. Gracefully handle case where zlib library 2838160814Ssimon can't be loaded. 2839160814Ssimon [Steve Henson] 2840160814Ssimon 2841160814Ssimon *) Several fixes and enhancements to the OID generation code. The old code 2842160814Ssimon sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 2843160814Ssimon handle numbers larger than ULONG_MAX, truncated printing and had a 2844160814Ssimon non standard OBJ_obj2txt() behaviour. 2845160814Ssimon [Steve Henson] 2846160814Ssimon 2847160814Ssimon *) Add support for building of engines under engine/ as shared libraries 2848160814Ssimon under VC++ build system. 2849160814Ssimon [Steve Henson] 2850160814Ssimon 2851160814Ssimon *) Corrected the numerous bugs in the Win32 path splitter in DSO. 2852160814Ssimon Hopefully, we will not see any false combination of paths any more. 2853160814Ssimon [Richard Levitte] 2854160814Ssimon 2855160814Ssimon Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 2856160814Ssimon 2857160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 2858160814Ssimon (part of SSL_OP_ALL). This option used to disable the 2859160814Ssimon countermeasure against man-in-the-middle protocol-version 2860160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 2861160814Ssimon idea. (CVE-2005-2969) 2862160814Ssimon 2863160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 2864160814Ssimon for Information Security, National Institute of Advanced Industrial 2865160814Ssimon Science and Technology [AIST], Japan)] 2866160814Ssimon 2867160814Ssimon *) Add two function to clear and return the verify parameter flags. 2868160814Ssimon [Steve Henson] 2869160814Ssimon 2870160814Ssimon *) Keep cipherlists sorted in the source instead of sorting them at 2871160814Ssimon runtime, thus removing the need for a lock. 2872160814Ssimon [Nils Larsch] 2873160814Ssimon 2874160814Ssimon *) Avoid some small subgroup attacks in Diffie-Hellman. 2875160814Ssimon [Nick Mathewson and Ben Laurie] 2876160814Ssimon 2877160814Ssimon *) Add functions for well-known primes. 2878160814Ssimon [Nick Mathewson] 2879160814Ssimon 2880160814Ssimon *) Extended Windows CE support. 2881160814Ssimon [Satoshi Nakamura and Andy Polyakov] 2882160814Ssimon 2883160814Ssimon *) Initialize SSL_METHOD structures at compile time instead of during 2884160814Ssimon runtime, thus removing the need for a lock. 2885160814Ssimon [Steve Henson] 2886160814Ssimon 2887160814Ssimon *) Make PKCS7_decrypt() work even if no certificate is supplied by 2888160814Ssimon attempting to decrypt each encrypted key in turn. Add support to 2889160814Ssimon smime utility. 2890160814Ssimon [Steve Henson] 2891160814Ssimon 2892160814Ssimon Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 2893160814Ssimon 2894162911Ssimon [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 2895162911Ssimon OpenSSL 0.9.8.] 2896162911Ssimon 2897160814Ssimon *) Add libcrypto.pc and libssl.pc for those who feel they need them. 2898160814Ssimon [Richard Levitte] 2899160814Ssimon 2900160814Ssimon *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 2901160814Ssimon key into the same file any more. 2902160814Ssimon [Richard Levitte] 2903160814Ssimon 2904160814Ssimon *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 2905160814Ssimon [Andy Polyakov] 2906160814Ssimon 2907160814Ssimon *) Add -utf8 command line and config file option to 'ca'. 2908160814Ssimon [Stefan <stf@udoma.org] 2909160814Ssimon 2910160814Ssimon *) Removed the macro des_crypt(), as it seems to conflict with some 2911160814Ssimon libraries. Use DES_crypt(). 2912160814Ssimon [Richard Levitte] 2913160814Ssimon 2914160814Ssimon *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 2915160814Ssimon involves renaming the source and generated shared-libs for 2916160814Ssimon both. The engines will accept the corrected or legacy ids 2917160814Ssimon ('ncipher' and '4758_cca' respectively) when binding. NB, 2918160814Ssimon this only applies when building 'shared'. 2919160814Ssimon [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 2920160814Ssimon 2921160814Ssimon *) Add attribute functions to EVP_PKEY structure. Modify 2922160814Ssimon PKCS12_create() to recognize a CSP name attribute and 2923160814Ssimon use it. Make -CSP option work again in pkcs12 utility. 2924160814Ssimon [Steve Henson] 2925160814Ssimon 2926160814Ssimon *) Add new functionality to the bn blinding code: 2927160814Ssimon - automatic re-creation of the BN_BLINDING parameters after 2928160814Ssimon a fixed number of uses (currently 32) 2929160814Ssimon - add new function for parameter creation 2930160814Ssimon - introduce flags to control the update behaviour of the 2931160814Ssimon BN_BLINDING parameters 2932160814Ssimon - hide BN_BLINDING structure 2933160814Ssimon Add a second BN_BLINDING slot to the RSA structure to improve 2934160814Ssimon performance when a single RSA object is shared among several 2935160814Ssimon threads. 2936160814Ssimon [Nils Larsch] 2937160814Ssimon 2938160814Ssimon *) Add support for DTLS. 2939160814Ssimon [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 2940160814Ssimon 2941160814Ssimon *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 2942160814Ssimon to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 2943160814Ssimon [Walter Goulet] 2944160814Ssimon 2945160814Ssimon *) Remove buggy and incompletet DH cert support from 2946160814Ssimon ssl/ssl_rsa.c and ssl/s3_both.c 2947160814Ssimon [Nils Larsch] 2948160814Ssimon 2949160814Ssimon *) Use SHA-1 instead of MD5 as the default digest algorithm for 2950160814Ssimon the apps/openssl applications. 2951160814Ssimon [Nils Larsch] 2952160814Ssimon 2953160814Ssimon *) Compile clean with "-Wall -Wmissing-prototypes 2954160814Ssimon -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 2955160814Ssimon DEBUG_SAFESTACK must also be set. 2956160814Ssimon [Ben Laurie] 2957160814Ssimon 2958160814Ssimon *) Change ./Configure so that certain algorithms can be disabled by default. 2959160814Ssimon The new counterpiece to "no-xxx" is "enable-xxx". 2960160814Ssimon 2961160814Ssimon The patented RC5 and MDC2 algorithms will now be disabled unless 2962160814Ssimon "enable-rc5" and "enable-mdc2", respectively, are specified. 2963160814Ssimon 2964160814Ssimon (IDEA remains enabled despite being patented. This is because IDEA 2965160814Ssimon is frequently required for interoperability, and there is no license 2966160814Ssimon fee for non-commercial use. As before, "no-idea" can be used to 2967160814Ssimon avoid this algorithm.) 2968160814Ssimon 2969160814Ssimon [Bodo Moeller] 2970160814Ssimon 2971160814Ssimon *) Add processing of proxy certificates (see RFC 3820). This work was 2972160814Ssimon sponsored by KTH (The Royal Institute of Technology in Stockholm) and 2973160814Ssimon EGEE (Enabling Grids for E-science in Europe). 2974160814Ssimon [Richard Levitte] 2975160814Ssimon 2976160814Ssimon *) RC4 performance overhaul on modern architectures/implementations, such 2977160814Ssimon as Intel P4, IA-64 and AMD64. 2978160814Ssimon [Andy Polyakov] 2979160814Ssimon 2980160814Ssimon *) New utility extract-section.pl. This can be used specify an alternative 2981160814Ssimon section number in a pod file instead of having to treat each file as 2982160814Ssimon a separate case in Makefile. This can be done by adding two lines to the 2983160814Ssimon pod file: 2984160814Ssimon 2985160814Ssimon =for comment openssl_section:XXX 2986160814Ssimon 2987160814Ssimon The blank line is mandatory. 2988160814Ssimon 2989160814Ssimon [Steve Henson] 2990160814Ssimon 2991160814Ssimon *) New arguments -certform, -keyform and -pass for s_client and s_server 2992160814Ssimon to allow alternative format key and certificate files and passphrase 2993160814Ssimon sources. 2994160814Ssimon [Steve Henson] 2995160814Ssimon 2996160814Ssimon *) New structure X509_VERIFY_PARAM which combines current verify parameters, 2997160814Ssimon update associated structures and add various utility functions. 2998160814Ssimon 2999160814Ssimon Add new policy related verify parameters, include policy checking in 3000160814Ssimon standard verify code. Enhance 'smime' application with extra parameters 3001160814Ssimon to support policy checking and print out. 3002160814Ssimon [Steve Henson] 3003160814Ssimon 3004160814Ssimon *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 3005160814Ssimon Nehemiah processors. These extensions support AES encryption in hardware 3006160814Ssimon as well as RNG (though RNG support is currently disabled). 3007160814Ssimon [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 3008160814Ssimon 3009160814Ssimon *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 3010160814Ssimon [Geoff Thorpe] 3011160814Ssimon 3012160814Ssimon *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 3013160814Ssimon [Andy Polyakov and a number of other people] 3014160814Ssimon 3015160814Ssimon *) Improved PowerPC platform support. Most notably BIGNUM assembler 3016160814Ssimon implementation contributed by IBM. 3017160814Ssimon [Suresh Chari, Peter Waltenberg, Andy Polyakov] 3018160814Ssimon 3019160814Ssimon *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 3020160814Ssimon exponent rather than 'unsigned long'. There is a corresponding change to 3021160814Ssimon the new 'rsa_keygen' element of the RSA_METHOD structure. 3022160814Ssimon [Jelte Jansen, Geoff Thorpe] 3023160814Ssimon 3024160814Ssimon *) Functionality for creating the initial serial number file is now 3025160814Ssimon moved from CA.pl to the 'ca' utility with a new option -create_serial. 3026160814Ssimon 3027160814Ssimon (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 3028160814Ssimon number file to 1, which is bound to cause problems. To avoid 3029160814Ssimon the problems while respecting compatibility between different 0.9.7 3030160814Ssimon patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 3031160814Ssimon CA.pl for serial number initialization. With the new release 0.9.8, 3032160814Ssimon we can fix the problem directly in the 'ca' utility.) 3033160814Ssimon [Steve Henson] 3034160814Ssimon 3035160814Ssimon *) Reduced header interdepencies by declaring more opaque objects in 3036160814Ssimon ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 3037160814Ssimon give fewer recursive includes, which could break lazy source code - so 3038160814Ssimon this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 3039160814Ssimon developers should define this symbol when building and using openssl to 3040160814Ssimon ensure they track the recommended behaviour, interfaces, [etc], but 3041160814Ssimon backwards-compatible behaviour prevails when this isn't defined. 3042160814Ssimon [Geoff Thorpe] 3043160814Ssimon 3044160814Ssimon *) New function X509_POLICY_NODE_print() which prints out policy nodes. 3045160814Ssimon [Steve Henson] 3046160814Ssimon 3047160814Ssimon *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 3048160814Ssimon This will generate a random key of the appropriate length based on the 3049160814Ssimon cipher context. The EVP_CIPHER can provide its own random key generation 3050160814Ssimon routine to support keys of a specific form. This is used in the des and 3051160814Ssimon 3des routines to generate a key of the correct parity. Update S/MIME 3052160814Ssimon code to use new functions and hence generate correct parity DES keys. 3053160814Ssimon Add EVP_CHECK_DES_KEY #define to return an error if the key is not 3054160814Ssimon valid (weak or incorrect parity). 3055160814Ssimon [Steve Henson] 3056160814Ssimon 3057160814Ssimon *) Add a local set of CRLs that can be used by X509_verify_cert() as well 3058160814Ssimon as looking them up. This is useful when the verified structure may contain 3059160814Ssimon CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 3060160814Ssimon present unless the new PKCS7_NO_CRL flag is asserted. 3061160814Ssimon [Steve Henson] 3062160814Ssimon 3063160814Ssimon *) Extend ASN1 oid configuration module. It now additionally accepts the 3064160814Ssimon syntax: 3065160814Ssimon 3066160814Ssimon shortName = some long name, 1.2.3.4 3067160814Ssimon [Steve Henson] 3068160814Ssimon 3069160814Ssimon *) Reimplemented the BN_CTX implementation. There is now no more static 3070160814Ssimon limitation on the number of variables it can handle nor the depth of the 3071160814Ssimon "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 3072160814Ssimon information can now expand as required, and rather than having a single 3073160814Ssimon static array of bignums, BN_CTX now uses a linked-list of such arrays 3074160814Ssimon allowing it to expand on demand whilst maintaining the usefulness of 3075160814Ssimon BN_CTX's "bundling". 3076160814Ssimon [Geoff Thorpe] 3077160814Ssimon 3078160814Ssimon *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 3079160814Ssimon to allow all RSA operations to function using a single BN_CTX. 3080160814Ssimon [Geoff Thorpe] 3081160814Ssimon 3082160814Ssimon *) Preliminary support for certificate policy evaluation and checking. This 3083160814Ssimon is initially intended to pass the tests outlined in "Conformance Testing 3084160814Ssimon of Relying Party Client Certificate Path Processing Logic" v1.07. 3085160814Ssimon [Steve Henson] 3086160814Ssimon 3087160814Ssimon *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 3088160814Ssimon remained unused and not that useful. A variety of other little bignum 3089160814Ssimon tweaks and fixes have also been made continuing on from the audit (see 3090160814Ssimon below). 3091160814Ssimon [Geoff Thorpe] 3092160814Ssimon 3093160814Ssimon *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 3094160814Ssimon associated ASN1, EVP and SSL functions and old ASN1 macros. 3095160814Ssimon [Richard Levitte] 3096160814Ssimon 3097160814Ssimon *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 3098160814Ssimon and this should never fail. So the return value from the use of 3099160814Ssimon BN_set_word() (which can fail due to needless expansion) is now deprecated; 3100160814Ssimon if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 3101160814Ssimon [Geoff Thorpe] 3102160814Ssimon 3103160814Ssimon *) BN_CTX_get() should return zero-valued bignums, providing the same 3104160814Ssimon initialised value as BN_new(). 3105160814Ssimon [Geoff Thorpe, suggested by Ulf M�ller] 3106160814Ssimon 3107160814Ssimon *) Support for inhibitAnyPolicy certificate extension. 3108160814Ssimon [Steve Henson] 3109160814Ssimon 3110160814Ssimon *) An audit of the BIGNUM code is underway, for which debugging code is 3111160814Ssimon enabled when BN_DEBUG is defined. This makes stricter enforcements on what 3112160814Ssimon is considered valid when processing BIGNUMs, and causes execution to 3113160814Ssimon assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 3114160814Ssimon further steps are taken to deliberately pollute unused data in BIGNUM 3115160814Ssimon structures to try and expose faulty code further on. For now, openssl will 3116160814Ssimon (in its default mode of operation) continue to tolerate the inconsistent 3117160814Ssimon forms that it has tolerated in the past, but authors and packagers should 3118160814Ssimon consider trying openssl and their own applications when compiled with 3119160814Ssimon these debugging symbols defined. It will help highlight potential bugs in 3120160814Ssimon their own code, and will improve the test coverage for OpenSSL itself. At 3121160814Ssimon some point, these tighter rules will become openssl's default to improve 3122160814Ssimon maintainability, though the assert()s and other overheads will remain only 3123160814Ssimon in debugging configurations. See bn.h for more details. 3124160814Ssimon [Geoff Thorpe, Nils Larsch, Ulf M�ller] 3125160814Ssimon 3126160814Ssimon *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 3127160814Ssimon that can only be obtained through BN_CTX_new() (which implicitly 3128160814Ssimon initialises it). The presence of this function only made it possible 3129160814Ssimon to overwrite an existing structure (and cause memory leaks). 3130160814Ssimon [Geoff Thorpe] 3131160814Ssimon 3132160814Ssimon *) Because of the callback-based approach for implementing LHASH as a 3133160814Ssimon template type, lh_insert() adds opaque objects to hash-tables and 3134160814Ssimon lh_doall() or lh_doall_arg() are typically used with a destructor callback 3135160814Ssimon to clean up those corresponding objects before destroying the hash table 3136160814Ssimon (and losing the object pointers). So some over-zealous constifications in 3137160814Ssimon LHASH have been relaxed so that lh_insert() does not take (nor store) the 3138160814Ssimon objects as "const" and the lh_doall[_arg] callback wrappers are not 3139160814Ssimon prototyped to have "const" restrictions on the object pointers they are 3140160814Ssimon given (and so aren't required to cast them away any more). 3141160814Ssimon [Geoff Thorpe] 3142160814Ssimon 3143160814Ssimon *) The tmdiff.h API was so ugly and minimal that our own timing utility 3144160814Ssimon (speed) prefers to use its own implementation. The two implementations 3145160814Ssimon haven't been consolidated as yet (volunteers?) but the tmdiff API has had 3146160814Ssimon its object type properly exposed (MS_TM) instead of casting to/from "char 3147160814Ssimon *". This may still change yet if someone realises MS_TM and "ms_time_***" 3148160814Ssimon aren't necessarily the greatest nomenclatures - but this is what was used 3149160814Ssimon internally to the implementation so I've used that for now. 3150160814Ssimon [Geoff Thorpe] 3151160814Ssimon 3152160814Ssimon *) Ensure that deprecated functions do not get compiled when 3153160814Ssimon OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 3154160814Ssimon the self-tests were still using deprecated key-generation functions so 3155160814Ssimon these have been updated also. 3156160814Ssimon [Geoff Thorpe] 3157160814Ssimon 3158160814Ssimon *) Reorganise PKCS#7 code to separate the digest location functionality 3159160814Ssimon into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 3160160814Ssimon New function PKCS7_set_digest() to set the digest type for PKCS#7 3161160814Ssimon digestedData type. Add additional code to correctly generate the 3162160814Ssimon digestedData type and add support for this type in PKCS7 initialization 3163160814Ssimon functions. 3164160814Ssimon [Steve Henson] 3165160814Ssimon 3166160814Ssimon *) New function PKCS7_set0_type_other() this initializes a PKCS7 3167160814Ssimon structure of type "other". 3168160814Ssimon [Steve Henson] 3169160814Ssimon 3170160814Ssimon *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 3171160814Ssimon sure the loop does correctly stop and breaking ("division by zero") 3172160814Ssimon modulus operations are not performed. The (pre-generated) prime 3173160814Ssimon table crypto/bn/bn_prime.h was already correct, but it could not be 3174160814Ssimon re-generated on some platforms because of the "division by zero" 3175160814Ssimon situation in the script. 3176160814Ssimon [Ralf S. Engelschall] 3177160814Ssimon 3178160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 3179160814Ssimon draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 3180160814Ssimon SHA-1 now is only used for "small" curves (where the 3181160814Ssimon representation of a field element takes up to 24 bytes); for 3182160814Ssimon larger curves, the field element resulting from ECDH is directly 3183160814Ssimon used as premaster secret. 3184160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 3185160814Ssimon 3186160814Ssimon *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 3187160814Ssimon curve secp160r1 to the tests. 3188160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 3189160814Ssimon 3190160814Ssimon *) Add the possibility to load symbols globally with DSO. 3191160814Ssimon [G�tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 3192160814Ssimon 3193160814Ssimon *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 3194160814Ssimon control of the error stack. 3195160814Ssimon [Richard Levitte] 3196160814Ssimon 3197160814Ssimon *) Add support for STORE in ENGINE. 3198160814Ssimon [Richard Levitte] 3199160814Ssimon 3200160814Ssimon *) Add the STORE type. The intention is to provide a common interface 3201160814Ssimon to certificate and key stores, be they simple file-based stores, or 3202160814Ssimon HSM-type store, or LDAP stores, or... 3203160814Ssimon NOTE: The code is currently UNTESTED and isn't really used anywhere. 3204160814Ssimon [Richard Levitte] 3205160814Ssimon 3206160814Ssimon *) Add a generic structure called OPENSSL_ITEM. This can be used to 3207160814Ssimon pass a list of arguments to any function as well as provide a way 3208160814Ssimon for a function to pass data back to the caller. 3209160814Ssimon [Richard Levitte] 3210160814Ssimon 3211160814Ssimon *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 3212160814Ssimon works like BUF_strdup() but can be used to duplicate a portion of 3213160814Ssimon a string. The copy gets NUL-terminated. BUF_memdup() duplicates 3214160814Ssimon a memory area. 3215160814Ssimon [Richard Levitte] 3216160814Ssimon 3217160814Ssimon *) Add the function sk_find_ex() which works like sk_find(), but will 3218160814Ssimon return an index to an element even if an exact match couldn't be 3219160814Ssimon found. The index is guaranteed to point at the element where the 3220160814Ssimon searched-for key would be inserted to preserve sorting order. 3221160814Ssimon [Richard Levitte] 3222160814Ssimon 3223160814Ssimon *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 3224160814Ssimon takes an extra flags argument for optional functionality. Currently, 3225160814Ssimon the following flags are defined: 3226160814Ssimon 3227160814Ssimon OBJ_BSEARCH_VALUE_ON_NOMATCH 3228160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 3229160814Ssimon element where the comparing function returns a negative or zero 3230160814Ssimon number. 3231160814Ssimon 3232160814Ssimon OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 3233160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 3234160814Ssimon element where the comparing function returns zero. This is useful 3235160814Ssimon if there are more than one element where the comparing function 3236160814Ssimon returns zero. 3237160814Ssimon [Richard Levitte] 3238160814Ssimon 3239160814Ssimon *) Make it possible to create self-signed certificates with 'openssl ca' 3240160814Ssimon in such a way that the self-signed certificate becomes part of the 3241160814Ssimon CA database and uses the same mechanisms for serial number generation 3242160814Ssimon as all other certificate signing. The new flag '-selfsign' enables 3243160814Ssimon this functionality. Adapt CA.sh and CA.pl.in. 3244160814Ssimon [Richard Levitte] 3245160814Ssimon 3246160814Ssimon *) Add functionality to check the public key of a certificate request 3247160814Ssimon against a given private. This is useful to check that a certificate 3248160814Ssimon request can be signed by that key (self-signing). 3249160814Ssimon [Richard Levitte] 3250160814Ssimon 3251160814Ssimon *) Make it possible to have multiple active certificates with the same 3252160814Ssimon subject in the CA index file. This is done only if the keyword 3253160814Ssimon 'unique_subject' is set to 'no' in the main CA section (default 3254160814Ssimon if 'CA_default') of the configuration file. The value is saved 3255160814Ssimon with the database itself in a separate index attribute file, 3256160814Ssimon named like the index file with '.attr' appended to the name. 3257160814Ssimon [Richard Levitte] 3258160814Ssimon 3259160814Ssimon *) Generate muti valued AVAs using '+' notation in config files for 3260160814Ssimon req and dirName. 3261160814Ssimon [Steve Henson] 3262160814Ssimon 3263160814Ssimon *) Support for nameConstraints certificate extension. 3264160814Ssimon [Steve Henson] 3265160814Ssimon 3266160814Ssimon *) Support for policyConstraints certificate extension. 3267160814Ssimon [Steve Henson] 3268160814Ssimon 3269160814Ssimon *) Support for policyMappings certificate extension. 3270160814Ssimon [Steve Henson] 3271160814Ssimon 3272160814Ssimon *) Make sure the default DSA_METHOD implementation only uses its 3273160814Ssimon dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 3274160814Ssimon and change its own handlers to be NULL so as to remove unnecessary 3275160814Ssimon indirection. This lets alternative implementations fallback to the 3276160814Ssimon default implementation more easily. 3277160814Ssimon [Geoff Thorpe] 3278160814Ssimon 3279160814Ssimon *) Support for directoryName in GeneralName related extensions 3280160814Ssimon in config files. 3281160814Ssimon [Steve Henson] 3282160814Ssimon 3283160814Ssimon *) Make it possible to link applications using Makefile.shared. 3284160814Ssimon Make that possible even when linking against static libraries! 3285160814Ssimon [Richard Levitte] 3286160814Ssimon 3287160814Ssimon *) Support for single pass processing for S/MIME signing. This now 3288160814Ssimon means that S/MIME signing can be done from a pipe, in addition 3289160814Ssimon cleartext signing (multipart/signed type) is effectively streaming 3290160814Ssimon and the signed data does not need to be all held in memory. 3291160814Ssimon 3292160814Ssimon This is done with a new flag PKCS7_STREAM. When this flag is set 3293160814Ssimon PKCS7_sign() only initializes the PKCS7 structure and the actual signing 3294160814Ssimon is done after the data is output (and digests calculated) in 3295160814Ssimon SMIME_write_PKCS7(). 3296160814Ssimon [Steve Henson] 3297160814Ssimon 3298160814Ssimon *) Add full support for -rpath/-R, both in shared libraries and 3299160814Ssimon applications, at least on the platforms where it's known how 3300160814Ssimon to do it. 3301160814Ssimon [Richard Levitte] 3302160814Ssimon 3303160814Ssimon *) In crypto/ec/ec_mult.c, implement fast point multiplication with 3304160814Ssimon precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 3305160814Ssimon will now compute a table of multiples of the generator that 3306160814Ssimon makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 3307160814Ssimon faster (notably in the case of a single point multiplication, 3308160814Ssimon scalar * generator). 3309160814Ssimon [Nils Larsch, Bodo Moeller] 3310160814Ssimon 3311160814Ssimon *) IPv6 support for certificate extensions. The various extensions 3312160814Ssimon which use the IP:a.b.c.d can now take IPv6 addresses using the 3313160814Ssimon formats of RFC1884 2.2 . IPv6 addresses are now also displayed 3314160814Ssimon correctly. 3315160814Ssimon [Steve Henson] 3316160814Ssimon 3317160814Ssimon *) Added an ENGINE that implements RSA by performing private key 3318160814Ssimon exponentiations with the GMP library. The conversions to and from 3319160814Ssimon GMP's mpz_t format aren't optimised nor are any montgomery forms 3320160814Ssimon cached, and on x86 it appears OpenSSL's own performance has caught up. 3321160814Ssimon However there are likely to be other architectures where GMP could 3322160814Ssimon provide a boost. This ENGINE is not built in by default, but it can be 3323160814Ssimon specified at Configure time and should be accompanied by the necessary 3324160814Ssimon linker additions, eg; 3325160814Ssimon ./config -DOPENSSL_USE_GMP -lgmp 3326160814Ssimon [Geoff Thorpe] 3327160814Ssimon 3328160814Ssimon *) "openssl engine" will not display ENGINE/DSO load failure errors when 3329160814Ssimon testing availability of engines with "-t" - the old behaviour is 3330160814Ssimon produced by increasing the feature's verbosity with "-tt". 3331160814Ssimon [Geoff Thorpe] 3332160814Ssimon 3333160814Ssimon *) ECDSA routines: under certain error conditions uninitialized BN objects 3334160814Ssimon could be freed. Solution: make sure initialization is performed early 3335160814Ssimon enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 3336160814Ssimon via PR#459) 3337160814Ssimon [Lutz Jaenicke] 3338160814Ssimon 3339160814Ssimon *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 3340160814Ssimon and DH_METHOD (eg. by ENGINE implementations) to override the normal 3341160814Ssimon software implementations. For DSA and DH, parameter generation can 3342160814Ssimon also be overriden by providing the appropriate method callbacks. 3343160814Ssimon [Geoff Thorpe] 3344160814Ssimon 3345160814Ssimon *) Change the "progress" mechanism used in key-generation and 3346160814Ssimon primality testing to functions that take a new BN_GENCB pointer in 3347160814Ssimon place of callback/argument pairs. The new API functions have "_ex" 3348160814Ssimon postfixes and the older functions are reimplemented as wrappers for 3349160814Ssimon the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 3350160814Ssimon declarations of the old functions to help (graceful) attempts to 3351160814Ssimon migrate to the new functions. Also, the new key-generation API 3352160814Ssimon functions operate on a caller-supplied key-structure and return 3353160814Ssimon success/failure rather than returning a key or NULL - this is to 3354160814Ssimon help make "keygen" another member function of RSA_METHOD etc. 3355160814Ssimon 3356160814Ssimon Example for using the new callback interface: 3357160814Ssimon 3358160814Ssimon int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 3359160814Ssimon void *my_arg = ...; 3360160814Ssimon BN_GENCB my_cb; 3361160814Ssimon 3362160814Ssimon BN_GENCB_set(&my_cb, my_callback, my_arg); 3363160814Ssimon 3364160814Ssimon return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 3365160814Ssimon /* For the meaning of a, b in calls to my_callback(), see the 3366160814Ssimon * documentation of the function that calls the callback. 3367160814Ssimon * cb will point to my_cb; my_arg can be retrieved as cb->arg. 3368160814Ssimon * my_callback should return 1 if it wants BN_is_prime_ex() 3369160814Ssimon * to continue, or 0 to stop. 3370160814Ssimon */ 3371160814Ssimon 3372160814Ssimon [Geoff Thorpe] 3373160814Ssimon 3374160814Ssimon *) Change the ZLIB compression method to be stateful, and make it 3375160814Ssimon available to TLS with the number defined in 3376160814Ssimon draft-ietf-tls-compression-04.txt. 3377160814Ssimon [Richard Levitte] 3378160814Ssimon 3379160814Ssimon *) Add the ASN.1 structures and functions for CertificatePair, which 3380160814Ssimon is defined as follows (according to X.509_4thEditionDraftV6.pdf): 3381160814Ssimon 3382160814Ssimon CertificatePair ::= SEQUENCE { 3383160814Ssimon forward [0] Certificate OPTIONAL, 3384160814Ssimon reverse [1] Certificate OPTIONAL, 3385160814Ssimon -- at least one of the pair shall be present -- } 3386160814Ssimon 3387160814Ssimon Also implement the PEM functions to read and write certificate 3388160814Ssimon pairs, and defined the PEM tag as "CERTIFICATE PAIR". 3389160814Ssimon 3390160814Ssimon This needed to be defined, mostly for the sake of the LDAP 3391160814Ssimon attribute crossCertificatePair, but may prove useful elsewhere as 3392160814Ssimon well. 3393160814Ssimon [Richard Levitte] 3394160814Ssimon 3395160814Ssimon *) Make it possible to inhibit symlinking of shared libraries in 3396160814Ssimon Makefile.shared, for Cygwin's sake. 3397160814Ssimon [Richard Levitte] 3398160814Ssimon 3399160814Ssimon *) Extend the BIGNUM API by creating a function 3400160814Ssimon void BN_set_negative(BIGNUM *a, int neg); 3401160814Ssimon and a macro that behave like 3402160814Ssimon int BN_is_negative(const BIGNUM *a); 3403160814Ssimon 3404160814Ssimon to avoid the need to access 'a->neg' directly in applications. 3405160814Ssimon [Nils Larsch] 3406160814Ssimon 3407160814Ssimon *) Implement fast modular reduction for pseudo-Mersenne primes 3408160814Ssimon used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 3409160814Ssimon EC_GROUP_new_curve_GFp() will now automatically use this 3410160814Ssimon if applicable. 3411160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3412160814Ssimon 3413160814Ssimon *) Add new lock type (CRYPTO_LOCK_BN). 3414160814Ssimon [Bodo Moeller] 3415160814Ssimon 3416160814Ssimon *) Change the ENGINE framework to automatically load engines 3417160814Ssimon dynamically from specific directories unless they could be 3418160814Ssimon found to already be built in or loaded. Move all the 3419160814Ssimon current engines except for the cryptodev one to a new 3420160814Ssimon directory engines/. 3421160814Ssimon The engines in engines/ are built as shared libraries if 3422160814Ssimon the "shared" options was given to ./Configure or ./config. 3423160814Ssimon Otherwise, they are inserted in libcrypto.a. 3424160814Ssimon /usr/local/ssl/engines is the default directory for dynamic 3425160814Ssimon engines, but that can be overriden at configure time through 3426160814Ssimon the usual use of --prefix and/or --openssldir, and at run 3427160814Ssimon time with the environment variable OPENSSL_ENGINES. 3428160814Ssimon [Geoff Thorpe and Richard Levitte] 3429160814Ssimon 3430160814Ssimon *) Add Makefile.shared, a helper makefile to build shared 3431160814Ssimon libraries. Addapt Makefile.org. 3432160814Ssimon [Richard Levitte] 3433160814Ssimon 3434160814Ssimon *) Add version info to Win32 DLLs. 3435160814Ssimon [Peter 'Luna' Runestig" <peter@runestig.com>] 3436160814Ssimon 3437160814Ssimon *) Add new 'medium level' PKCS#12 API. Certificates and keys 3438160814Ssimon can be added using this API to created arbitrary PKCS#12 3439160814Ssimon files while avoiding the low level API. 3440160814Ssimon 3441160814Ssimon New options to PKCS12_create(), key or cert can be NULL and 3442160814Ssimon will then be omitted from the output file. The encryption 3443160814Ssimon algorithm NIDs can be set to -1 for no encryption, the mac 3444160814Ssimon iteration count can be set to 0 to omit the mac. 3445160814Ssimon 3446160814Ssimon Enhance pkcs12 utility by making the -nokeys and -nocerts 3447160814Ssimon options work when creating a PKCS#12 file. New option -nomac 3448160814Ssimon to omit the mac, NONE can be set for an encryption algorithm. 3449160814Ssimon New code is modified to use the enhanced PKCS12_create() 3450160814Ssimon instead of the low level API. 3451160814Ssimon [Steve Henson] 3452160814Ssimon 3453160814Ssimon *) Extend ASN1 encoder to support indefinite length constructed 3454160814Ssimon encoding. This can output sequences tags and octet strings in 3455160814Ssimon this form. Modify pk7_asn1.c to support indefinite length 3456160814Ssimon encoding. This is experimental and needs additional code to 3457160814Ssimon be useful, such as an ASN1 bio and some enhanced streaming 3458160814Ssimon PKCS#7 code. 3459160814Ssimon 3460160814Ssimon Extend template encode functionality so that tagging is passed 3461160814Ssimon down to the template encoder. 3462160814Ssimon [Steve Henson] 3463160814Ssimon 3464160814Ssimon *) Let 'openssl req' fail if an argument to '-newkey' is not 3465160814Ssimon recognized instead of using RSA as a default. 3466160814Ssimon [Bodo Moeller] 3467160814Ssimon 3468160814Ssimon *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 3469160814Ssimon As these are not official, they are not included in "ALL"; 3470160814Ssimon the "ECCdraft" ciphersuite group alias can be used to select them. 3471160814Ssimon [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 3472160814Ssimon 3473160814Ssimon *) Add ECDH engine support. 3474160814Ssimon [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 3475160814Ssimon 3476160814Ssimon *) Add ECDH in new directory crypto/ecdh/. 3477160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 3478160814Ssimon 3479160814Ssimon *) Let BN_rand_range() abort with an error after 100 iterations 3480160814Ssimon without success (which indicates a broken PRNG). 3481160814Ssimon [Bodo Moeller] 3482160814Ssimon 3483160814Ssimon *) Change BN_mod_sqrt() so that it verifies that the input value 3484160814Ssimon is really the square of the return value. (Previously, 3485160814Ssimon BN_mod_sqrt would show GIGO behaviour.) 3486160814Ssimon [Bodo Moeller] 3487160814Ssimon 3488160814Ssimon *) Add named elliptic curves over binary fields from X9.62, SECG, 3489160814Ssimon and WAP/WTLS; add OIDs that were still missing. 3490160814Ssimon 3491160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3492160814Ssimon (Sun Microsystems Laboratories)] 3493160814Ssimon 3494160814Ssimon *) Extend the EC library for elliptic curves over binary fields 3495160814Ssimon (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 3496160814Ssimon New EC_METHOD: 3497160814Ssimon 3498160814Ssimon EC_GF2m_simple_method 3499160814Ssimon 3500160814Ssimon New API functions: 3501160814Ssimon 3502160814Ssimon EC_GROUP_new_curve_GF2m 3503160814Ssimon EC_GROUP_set_curve_GF2m 3504160814Ssimon EC_GROUP_get_curve_GF2m 3505160814Ssimon EC_POINT_set_affine_coordinates_GF2m 3506160814Ssimon EC_POINT_get_affine_coordinates_GF2m 3507160814Ssimon EC_POINT_set_compressed_coordinates_GF2m 3508160814Ssimon 3509160814Ssimon Point compression for binary fields is disabled by default for 3510160814Ssimon patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 3511160814Ssimon enable it). 3512160814Ssimon 3513160814Ssimon As binary polynomials are represented as BIGNUMs, various members 3514160814Ssimon of the EC_GROUP and EC_POINT data structures can be shared 3515160814Ssimon between the implementations for prime fields and binary fields; 3516160814Ssimon the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 3517160814Ssimon are essentially identical to their ..._GFp counterparts. 3518160814Ssimon (For simplicity, the '..._GFp' prefix has been dropped from 3519160814Ssimon various internal method names.) 3520160814Ssimon 3521160814Ssimon An internal 'field_div' method (similar to 'field_mul' and 3522160814Ssimon 'field_sqr') has been added; this is used only for binary fields. 3523160814Ssimon 3524160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3525160814Ssimon (Sun Microsystems Laboratories)] 3526160814Ssimon 3527160814Ssimon *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 3528160814Ssimon through methods ('mul', 'precompute_mult'). 3529160814Ssimon 3530160814Ssimon The generic implementations (now internally called 'ec_wNAF_mul' 3531160814Ssimon and 'ec_wNAF_precomputed_mult') remain the default if these 3532160814Ssimon methods are undefined. 3533160814Ssimon 3534160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3535160814Ssimon (Sun Microsystems Laboratories)] 3536160814Ssimon 3537160814Ssimon *) New function EC_GROUP_get_degree, which is defined through 3538160814Ssimon EC_METHOD. For curves over prime fields, this returns the bit 3539160814Ssimon length of the modulus. 3540160814Ssimon 3541160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3542160814Ssimon (Sun Microsystems Laboratories)] 3543160814Ssimon 3544160814Ssimon *) New functions EC_GROUP_dup, EC_POINT_dup. 3545160814Ssimon (These simply call ..._new and ..._copy). 3546160814Ssimon 3547160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3548160814Ssimon (Sun Microsystems Laboratories)] 3549160814Ssimon 3550160814Ssimon *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 3551160814Ssimon Polynomials are represented as BIGNUMs (where the sign bit is not 3552160814Ssimon used) in the following functions [macros]: 3553160814Ssimon 3554160814Ssimon BN_GF2m_add 3555160814Ssimon BN_GF2m_sub [= BN_GF2m_add] 3556160814Ssimon BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 3557160814Ssimon BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 3558160814Ssimon BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 3559160814Ssimon BN_GF2m_mod_inv 3560160814Ssimon BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 3561160814Ssimon BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 3562160814Ssimon BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 3563160814Ssimon BN_GF2m_cmp [= BN_ucmp] 3564160814Ssimon 3565160814Ssimon (Note that only the 'mod' functions are actually for fields GF(2^m). 3566160814Ssimon BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 3567160814Ssimon 3568160814Ssimon For some functions, an the irreducible polynomial defining a 3569160814Ssimon field can be given as an 'unsigned int[]' with strictly 3570160814Ssimon decreasing elements giving the indices of those bits that are set; 3571160814Ssimon i.e., p[] represents the polynomial 3572160814Ssimon f(t) = t^p[0] + t^p[1] + ... + t^p[k] 3573160814Ssimon where 3574160814Ssimon p[0] > p[1] > ... > p[k] = 0. 3575160814Ssimon This applies to the following functions: 3576160814Ssimon 3577160814Ssimon BN_GF2m_mod_arr 3578160814Ssimon BN_GF2m_mod_mul_arr 3579160814Ssimon BN_GF2m_mod_sqr_arr 3580160814Ssimon BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 3581160814Ssimon BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 3582160814Ssimon BN_GF2m_mod_exp_arr 3583160814Ssimon BN_GF2m_mod_sqrt_arr 3584160814Ssimon BN_GF2m_mod_solve_quad_arr 3585160814Ssimon BN_GF2m_poly2arr 3586160814Ssimon BN_GF2m_arr2poly 3587160814Ssimon 3588160814Ssimon Conversion can be performed by the following functions: 3589160814Ssimon 3590160814Ssimon BN_GF2m_poly2arr 3591160814Ssimon BN_GF2m_arr2poly 3592160814Ssimon 3593160814Ssimon bntest.c has additional tests for binary polynomial arithmetic. 3594160814Ssimon 3595160814Ssimon Two implementations for BN_GF2m_mod_div() are available. 3596160814Ssimon The default algorithm simply uses BN_GF2m_mod_inv() and 3597160814Ssimon BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 3598160814Ssimon if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 3599160814Ssimon copyright notice in crypto/bn/bn_gf2m.c before enabling it). 3600160814Ssimon 3601160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3602160814Ssimon (Sun Microsystems Laboratories)] 3603160814Ssimon 3604160814Ssimon *) Add new error code 'ERR_R_DISABLED' that can be used when some 3605160814Ssimon functionality is disabled at compile-time. 3606160814Ssimon [Douglas Stebila <douglas.stebila@sun.com>] 3607160814Ssimon 3608160814Ssimon *) Change default behaviour of 'openssl asn1parse' so that more 3609160814Ssimon information is visible when viewing, e.g., a certificate: 3610160814Ssimon 3611160814Ssimon Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 3612160814Ssimon mode the content of non-printable OCTET STRINGs is output in a 3613160814Ssimon style similar to INTEGERs, but with '[HEX DUMP]' prepended to 3614160814Ssimon avoid the appearance of a printable string. 3615160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3616160814Ssimon 3617160814Ssimon *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 3618160814Ssimon functions 3619160814Ssimon EC_GROUP_set_asn1_flag() 3620160814Ssimon EC_GROUP_get_asn1_flag() 3621160814Ssimon EC_GROUP_set_point_conversion_form() 3622160814Ssimon EC_GROUP_get_point_conversion_form() 3623160814Ssimon These control ASN1 encoding details: 3624160814Ssimon - Curves (i.e., groups) are encoded explicitly unless asn1_flag 3625160814Ssimon has been set to OPENSSL_EC_NAMED_CURVE. 3626160814Ssimon - Points are encoded in uncompressed form by default; options for 3627160814Ssimon asn1_for are as for point2oct, namely 3628160814Ssimon POINT_CONVERSION_COMPRESSED 3629160814Ssimon POINT_CONVERSION_UNCOMPRESSED 3630160814Ssimon POINT_CONVERSION_HYBRID 3631160814Ssimon 3632160814Ssimon Also add 'seed' and 'seed_len' members to EC_GROUP with access 3633160814Ssimon functions 3634160814Ssimon EC_GROUP_set_seed() 3635160814Ssimon EC_GROUP_get0_seed() 3636160814Ssimon EC_GROUP_get_seed_len() 3637160814Ssimon This is used only for ASN1 purposes (so far). 3638160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3639160814Ssimon 3640160814Ssimon *) Add 'field_type' member to EC_METHOD, which holds the NID 3641160814Ssimon of the appropriate field type OID. The new function 3642160814Ssimon EC_METHOD_get_field_type() returns this value. 3643160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3644160814Ssimon 3645160814Ssimon *) Add functions 3646160814Ssimon EC_POINT_point2bn() 3647160814Ssimon EC_POINT_bn2point() 3648160814Ssimon EC_POINT_point2hex() 3649160814Ssimon EC_POINT_hex2point() 3650160814Ssimon providing useful interfaces to EC_POINT_point2oct() and 3651160814Ssimon EC_POINT_oct2point(). 3652160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3653160814Ssimon 3654160814Ssimon *) Change internals of the EC library so that the functions 3655160814Ssimon EC_GROUP_set_generator() 3656160814Ssimon EC_GROUP_get_generator() 3657160814Ssimon EC_GROUP_get_order() 3658160814Ssimon EC_GROUP_get_cofactor() 3659160814Ssimon are implemented directly in crypto/ec/ec_lib.c and not dispatched 3660160814Ssimon to methods, which would lead to unnecessary code duplication when 3661160814Ssimon adding different types of curves. 3662160814Ssimon [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 3663160814Ssimon 3664160814Ssimon *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 3665160814Ssimon arithmetic, and such that modified wNAFs are generated 3666160814Ssimon (which avoid length expansion in many cases). 3667160814Ssimon [Bodo Moeller] 3668160814Ssimon 3669160814Ssimon *) Add a function EC_GROUP_check_discriminant() (defined via 3670160814Ssimon EC_METHOD) that verifies that the curve discriminant is non-zero. 3671160814Ssimon 3672160814Ssimon Add a function EC_GROUP_check() that makes some sanity tests 3673160814Ssimon on a EC_GROUP, its generator and order. This includes 3674160814Ssimon EC_GROUP_check_discriminant(). 3675160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3676160814Ssimon 3677160814Ssimon *) Add ECDSA in new directory crypto/ecdsa/. 3678160814Ssimon 3679160814Ssimon Add applications 'openssl ecparam' and 'openssl ecdsa' 3680160814Ssimon (these are based on 'openssl dsaparam' and 'openssl dsa'). 3681160814Ssimon 3682160814Ssimon ECDSA support is also included in various other files across the 3683160814Ssimon library. Most notably, 3684160814Ssimon - 'openssl req' now has a '-newkey ecdsa:file' option; 3685160814Ssimon - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 3686160814Ssimon - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 3687160814Ssimon d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 3688160814Ssimon them suitable for ECDSA where domain parameters must be 3689160814Ssimon extracted before the specific public key; 3690160814Ssimon - ECDSA engine support has been added. 3691160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3692160814Ssimon 3693160814Ssimon *) Include some named elliptic curves, and add OIDs from X9.62, 3694160814Ssimon SECG, and WAP/WTLS. Each curve can be obtained from the new 3695160814Ssimon function 3696160814Ssimon EC_GROUP_new_by_curve_name(), 3697160814Ssimon and the list of available named curves can be obtained with 3698160814Ssimon EC_get_builtin_curves(). 3699160814Ssimon Also add a 'curve_name' member to EC_GROUP objects, which can be 3700160814Ssimon accessed via 3701160814Ssimon EC_GROUP_set_curve_name() 3702160814Ssimon EC_GROUP_get_curve_name() 3703160814Ssimon [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 3704160814Ssimon 3705160814Ssimon *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 3706160814Ssimon was actually never needed) and in BN_mul(). The removal in BN_mul() 3707160814Ssimon required a small change in bn_mul_part_recursive() and the addition 3708160814Ssimon of the functions bn_cmp_part_words(), bn_sub_part_words() and 3709160814Ssimon bn_add_part_words(), which do the same thing as bn_cmp_words(), 3710160814Ssimon bn_sub_words() and bn_add_words() except they take arrays with 3711160814Ssimon differing sizes. 3712160814Ssimon [Richard Levitte] 3713160814Ssimon 3714194206Ssimon Changes between 0.9.7l and 0.9.7m [23 Feb 2007] 3715194206Ssimon 3716167612Ssimon *) Cleanse PEM buffers before freeing them since they may contain 3717167612Ssimon sensitive data. 3718167612Ssimon [Benjamin Bennett <ben@psc.edu>] 3719167612Ssimon 3720167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 3721167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 3722167612Ssimon authentication-only ciphersuites. 3723167612Ssimon [Bodo Moeller] 3724167612Ssimon 3725167612Ssimon *) Since AES128 and AES256 share a single mask bit in the logic of 3726167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 3727167612Ssimon kludge to work properly if AES128 is available and AES256 isn't. 3728167612Ssimon [Victor Duchovni] 3729167612Ssimon 3730194206Ssimon *) Expand security boundary to match 1.1.1 module. 3731194206Ssimon [Steve Henson] 3732194206Ssimon 3733194206Ssimon *) Remove redundant features: hash file source, editing of test vectors 3734194206Ssimon modify fipsld to use external fips_premain.c signature. 3735194206Ssimon [Steve Henson] 3736194206Ssimon 3737194206Ssimon *) New perl script mkfipsscr.pl to create shell scripts or batch files to 3738194206Ssimon run algorithm test programs. 3739194206Ssimon [Steve Henson] 3740194206Ssimon 3741194206Ssimon *) Make algorithm test programs more tolerant of whitespace. 3742194206Ssimon [Steve Henson] 3743194206Ssimon 3744167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 3745167612Ssimon protocol version while receiving ClientHello even if the 3746167612Ssimon ClientHello is fragmented. (The server can't insist on the 3747167612Ssimon particular protocol version it has chosen before the ServerHello 3748167612Ssimon message has informed the client about his choice.) 3749167612Ssimon [Bodo Moeller] 3750167612Ssimon 3751167612Ssimon *) Load error codes if they are not already present instead of using a 3752167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 3753167612Ssimon [Steve Henson] 3754167612Ssimon 3755167612Ssimon Changes between 0.9.7k and 0.9.7l [28 Sep 2006] 3756167612Ssimon 3757167612Ssimon *) Introduce limits to prevent malicious keys being able to 3758167612Ssimon cause a denial of service. (CVE-2006-2940) 3759167612Ssimon [Steve Henson, Bodo Moeller] 3760167612Ssimon 3761167612Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 3762167612Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 3763167612Ssimon 3764167612Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 3765167612Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 3766167612Ssimon 3767167612Ssimon *) Fix SSL client code which could crash if connecting to a 3768167612Ssimon malicious SSLv2 server. (CVE-2006-4343) 3769167612Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 3770167612Ssimon 3771162911Ssimon *) Change ciphersuite string processing so that an explicit 3772162911Ssimon ciphersuite selects this one ciphersuite (so that "AES256-SHA" 3773162911Ssimon will no longer include "AES128-SHA"), and any other similar 3774162911Ssimon ciphersuite (same bitmap) from *other* protocol versions (so that 3775162911Ssimon "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 3776162911Ssimon SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 3777162911Ssimon changes from 0.9.8b and 0.9.8d. 3778162911Ssimon [Bodo Moeller] 3779162911Ssimon 3780162911Ssimon Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 3781162911Ssimon 3782162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 3783162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 3784162911Ssimon 3785162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 3786162911Ssimon possible instead of select(), since the latter has some 3787162911Ssimon undesirable limitations. 3788162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 3789162911Ssimon 3790162911Ssimon *) Disable rogue ciphersuites: 3791162911Ssimon 3792162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 3793162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 3794162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 3795162911Ssimon 3796162911Ssimon The latter two were purportedly from 3797162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 3798162911Ssimon appear there. 3799162911Ssimon 3800162911Ssimon Also deactive the remaining ciphersuites from 3801162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 3802162911Ssimon unofficial, and the ID has long expired. 3803162911Ssimon [Bodo Moeller] 3804162911Ssimon 3805162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 3806162911Ssimon dual-core machines) and other potential thread-safety issues. 3807162911Ssimon [Bodo Moeller] 3808162911Ssimon 3809162911Ssimon Changes between 0.9.7i and 0.9.7j [04 May 2006] 3810162911Ssimon 3811162911Ssimon *) Adapt fipsld and the build system to link against the validated FIPS 3812162911Ssimon module in FIPS mode. 3813162911Ssimon [Steve Henson] 3814162911Ssimon 3815162911Ssimon *) Fixes for VC++ 2005 build under Windows. 3816162911Ssimon [Steve Henson] 3817162911Ssimon 3818162911Ssimon *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 3819162911Ssimon from a Windows bash shell such as MSYS. It is autodetected from the 3820162911Ssimon "config" script when run from a VC++ environment. Modify standard VC++ 3821162911Ssimon build to use fipscanister.o from the GNU make build. 3822162911Ssimon [Steve Henson] 3823162911Ssimon 3824160814Ssimon Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 3825160814Ssimon 3826160814Ssimon *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 3827160814Ssimon The value now differs depending on if you build for FIPS or not. 3828160814Ssimon BEWARE! A program linked with a shared FIPSed libcrypto can't be 3829160814Ssimon safely run with a non-FIPSed libcrypto, as it may crash because of 3830160814Ssimon the difference induced by this change. 3831160814Ssimon [Andy Polyakov] 3832160814Ssimon 3833160814Ssimon Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 3834160814Ssimon 3835160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 3836160814Ssimon (part of SSL_OP_ALL). This option used to disable the 3837160814Ssimon countermeasure against man-in-the-middle protocol-version 3838160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 3839160814Ssimon idea. (CVE-2005-2969) 3840160814Ssimon 3841160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 3842160814Ssimon for Information Security, National Institute of Advanced Industrial 3843160814Ssimon Science and Technology [AIST], Japan)] 3844160814Ssimon 3845160814Ssimon *) Minimal support for X9.31 signatures and PSS padding modes. This is 3846160814Ssimon mainly for FIPS compliance and not fully integrated at this stage. 3847160814Ssimon [Steve Henson] 3848160814Ssimon 3849160814Ssimon *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 3850160814Ssimon the exponentiation using a fixed-length exponent. (Otherwise, 3851160814Ssimon the information leaked through timing could expose the secret key 3852160814Ssimon after many signatures; cf. Bleichenbacher's attack on DSA with 3853160814Ssimon biased k.) 3854160814Ssimon [Bodo Moeller] 3855160814Ssimon 3856160814Ssimon *) Make a new fixed-window mod_exp implementation the default for 3857160814Ssimon RSA, DSA, and DH private-key operations so that the sequence of 3858160814Ssimon squares and multiplies and the memory access pattern are 3859160814Ssimon independent of the particular secret key. This will mitigate 3860160814Ssimon cache-timing and potential related attacks. 3861160814Ssimon 3862160814Ssimon BN_mod_exp_mont_consttime() is the new exponentiation implementation, 3863160814Ssimon and this is automatically used by BN_mod_exp_mont() if the new flag 3864160814Ssimon BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 3865160814Ssimon will use this BN flag for private exponents unless the flag 3866160814Ssimon RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 3867160814Ssimon DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 3868160814Ssimon 3869160814Ssimon [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 3870160814Ssimon 3871160814Ssimon *) Change the client implementation for SSLv23_method() and 3872160814Ssimon SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 3873160814Ssimon Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 3874160814Ssimon (Previously, the SSL 2.0 backwards compatible Client Hello 3875160814Ssimon message format would be used even with SSL_OP_NO_SSLv2.) 3876160814Ssimon [Bodo Moeller] 3877160814Ssimon 3878160814Ssimon *) Add support for smime-type MIME parameter in S/MIME messages which some 3879160814Ssimon clients need. 3880160814Ssimon [Steve Henson] 3881160814Ssimon 3882160814Ssimon *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 3883160814Ssimon a threadsafe manner. Modify rsa code to use new function and add calls 3884160814Ssimon to dsa and dh code (which had race conditions before). 3885160814Ssimon [Steve Henson] 3886160814Ssimon 3887160814Ssimon *) Include the fixed error library code in the C error file definitions 3888160814Ssimon instead of fixing them up at runtime. This keeps the error code 3889160814Ssimon structures constant. 3890160814Ssimon [Steve Henson] 3891160814Ssimon 3892160814Ssimon Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 3893160814Ssimon 3894160814Ssimon [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 3895160814Ssimon OpenSSL 0.9.8.] 3896160814Ssimon 3897160814Ssimon *) Fixes for newer kerberos headers. NB: the casts are needed because 3898160814Ssimon the 'length' field is signed on one version and unsigned on another 3899160814Ssimon with no (?) obvious way to tell the difference, without these VC++ 3900160814Ssimon complains. Also the "definition" of FAR (blank) is no longer included 3901160814Ssimon nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 3902160814Ssimon some needed definitions. 3903160814Ssimon [Steve Henson] 3904160814Ssimon 3905160814Ssimon *) Undo Cygwin change. 3906160814Ssimon [Ulf M�ller] 3907160814Ssimon 3908160814Ssimon *) Added support for proxy certificates according to RFC 3820. 3909160814Ssimon Because they may be a security thread to unaware applications, 3910160814Ssimon they must be explicitely allowed in run-time. See 3911160814Ssimon docs/HOWTO/proxy_certificates.txt for further information. 3912160814Ssimon [Richard Levitte] 3913160814Ssimon 3914160814Ssimon Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 3915160814Ssimon 3916160814Ssimon *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 3917160814Ssimon server and client random values. Previously 3918160814Ssimon (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 3919160814Ssimon less random data when sizeof(time_t) > 4 (some 64 bit platforms). 3920160814Ssimon 3921160814Ssimon This change has negligible security impact because: 3922160814Ssimon 3923160814Ssimon 1. Server and client random values still have 24 bytes of pseudo random 3924160814Ssimon data. 3925160814Ssimon 3926160814Ssimon 2. Server and client random values are sent in the clear in the initial 3927160814Ssimon handshake. 3928160814Ssimon 3929160814Ssimon 3. The master secret is derived using the premaster secret (48 bytes in 3930160814Ssimon size for static RSA ciphersuites) as well as client server and random 3931160814Ssimon values. 3932160814Ssimon 3933160814Ssimon The OpenSSL team would like to thank the UK NISCC for bringing this issue 3934160814Ssimon to our attention. 3935160814Ssimon 3936160814Ssimon [Stephen Henson, reported by UK NISCC] 3937160814Ssimon 3938160814Ssimon *) Use Windows randomness collection on Cygwin. 3939160814Ssimon [Ulf M�ller] 3940160814Ssimon 3941160814Ssimon *) Fix hang in EGD/PRNGD query when communication socket is closed 3942160814Ssimon prematurely by EGD/PRNGD. 3943160814Ssimon [Darren Tucker <dtucker@zip.com.au> via Lutz J�nicke, resolves #1014] 3944160814Ssimon 3945160814Ssimon *) Prompt for pass phrases when appropriate for PKCS12 input format. 3946160814Ssimon [Steve Henson] 3947160814Ssimon 3948160814Ssimon *) Back-port of selected performance improvements from development 3949160814Ssimon branch, as well as improved support for PowerPC platforms. 3950160814Ssimon [Andy Polyakov] 3951160814Ssimon 3952160814Ssimon *) Add lots of checks for memory allocation failure, error codes to indicate 3953160814Ssimon failure and freeing up memory if a failure occurs. 3954160814Ssimon [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 3955160814Ssimon 3956160814Ssimon *) Add new -passin argument to dgst. 3957160814Ssimon [Steve Henson] 3958160814Ssimon 3959160814Ssimon *) Perform some character comparisons of different types in X509_NAME_cmp: 3960160814Ssimon this is needed for some certificates that reencode DNs into UTF8Strings 3961160814Ssimon (in violation of RFC3280) and can't or wont issue name rollover 3962160814Ssimon certificates. 3963160814Ssimon [Steve Henson] 3964160814Ssimon 3965160814Ssimon *) Make an explicit check during certificate validation to see that 3966160814Ssimon the CA setting in each certificate on the chain is correct. As a 3967160814Ssimon side effect always do the following basic checks on extensions, 3968160814Ssimon not just when there's an associated purpose to the check: 3969160814Ssimon 3970160814Ssimon - if there is an unhandled critical extension (unless the user 3971160814Ssimon has chosen to ignore this fault) 3972160814Ssimon - if the path length has been exceeded (if one is set at all) 3973160814Ssimon - that certain extensions fit the associated purpose (if one has 3974160814Ssimon been given) 3975160814Ssimon [Richard Levitte] 3976160814Ssimon 3977142425Snectar Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 3978142425Snectar 3979142425Snectar *) Avoid a race condition when CRLs are checked in a multi threaded 3980142425Snectar environment. This would happen due to the reordering of the revoked 3981142425Snectar entries during signature checking and serial number lookup. Now the 3982142425Snectar encoding is cached and the serial number sort performed under a lock. 3983142425Snectar Add new STACK function sk_is_sorted(). 3984142425Snectar [Steve Henson] 3985142425Snectar 3986142425Snectar *) Add Delta CRL to the extension code. 3987142425Snectar [Steve Henson] 3988142425Snectar 3989142425Snectar *) Various fixes to s3_pkt.c so alerts are sent properly. 3990142425Snectar [David Holmes <d.holmes@f5.com>] 3991142425Snectar 3992142425Snectar *) Reduce the chances of duplicate issuer name and serial numbers (in 3993142425Snectar violation of RFC3280) using the OpenSSL certificate creation utilities. 3994142425Snectar This is done by creating a random 64 bit value for the initial serial 3995142425Snectar number when a serial number file is created or when a self signed 3996142425Snectar certificate is created using 'openssl req -x509'. The initial serial 3997142425Snectar number file is created using 'openssl x509 -next_serial' in CA.pl 3998142425Snectar rather than being initialized to 1. 3999142425Snectar [Steve Henson] 4000142425Snectar 4001127128Snectar Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 4002127128Snectar 4003127128Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 4004160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 4005127128Snectar [Joe Orton, Steve Henson] 4006127128Snectar 4007127128Snectar *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 4008160814Ssimon (CVE-2004-0112) 4009127128Snectar [Joe Orton, Steve Henson] 4010127128Snectar 4011127128Snectar *) Make it possible to have multiple active certificates with the same 4012127128Snectar subject in the CA index file. This is done only if the keyword 4013127128Snectar 'unique_subject' is set to 'no' in the main CA section (default 4014127128Snectar if 'CA_default') of the configuration file. The value is saved 4015127128Snectar with the database itself in a separate index attribute file, 4016127128Snectar named like the index file with '.attr' appended to the name. 4017127128Snectar [Richard Levitte] 4018127128Snectar 4019127128Snectar *) X509 verify fixes. Disable broken certificate workarounds when 4020127128Snectar X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 4021127128Snectar keyUsage extension present. Don't accept CRLs with unhandled critical 4022127128Snectar extensions: since verify currently doesn't process CRL extensions this 4023127128Snectar rejects a CRL with *any* critical extensions. Add new verify error codes 4024127128Snectar for these cases. 4025127128Snectar [Steve Henson] 4026127128Snectar 4027127128Snectar *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 4028127128Snectar A clarification of RFC2560 will require the use of OCTET STRINGs and 4029127128Snectar some implementations cannot handle the current raw format. Since OpenSSL 4030127128Snectar copies and compares OCSP nonces as opaque blobs without any attempt at 4031127128Snectar parsing them this should not create any compatibility issues. 4032127128Snectar [Steve Henson] 4033127128Snectar 4034127128Snectar *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 4035127128Snectar calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 4036127128Snectar this HMAC (and other) operations are several times slower than OpenSSL 4037127128Snectar < 0.9.7. 4038127128Snectar [Steve Henson] 4039127128Snectar 4040127128Snectar *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 4041127128Snectar [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 4042127128Snectar 4043127128Snectar *) Use the correct content when signing type "other". 4044127128Snectar [Steve Henson] 4045127128Snectar 4046120631Snectar Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 4047120631Snectar 4048120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 4049120631Snectar 4050120631Snectar Stop out of bounds reads in the ASN1 code when presented with 4051160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 4052120631Snectar 4053160814Ssimon Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 4054120631Snectar 4055120631Snectar If verify callback ignores invalid public key errors don't try to check 4056120631Snectar certificate signature with the NULL public key. 4057120631Snectar 4058120631Snectar [Steve Henson] 4059120631Snectar 4060120631Snectar *) New -ignore_err option in ocsp application to stop the server 4061120631Snectar exiting on the first error in a request. 4062120631Snectar [Steve Henson] 4063120631Snectar 4064120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 4065120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 4066120631Snectar specifications. 4067120631Snectar [Steve Henson] 4068120631Snectar 4069120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 4070120631Snectar extra data after the compression methods not only for TLS 1.0 4071120631Snectar but also for SSL 3.0 (as required by the specification). 4072120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 4073120631Snectar 4074120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 4075120631Snectar when it's 512 *bits* long, not 512 bytes. 4076120631Snectar [Richard Levitte] 4077120631Snectar 4078120631Snectar *) Change AES_cbc_encrypt() so it outputs exact multiple of 4079120631Snectar blocks during encryption. 4080120631Snectar [Richard Levitte] 4081120631Snectar 4082120631Snectar *) Various fixes to base64 BIO and non blocking I/O. On write 4083120631Snectar flushes were not handled properly if the BIO retried. On read 4084120631Snectar data was not being buffered properly and had various logic bugs. 4085120631Snectar This also affects blocking I/O when the data being decoded is a 4086120631Snectar certain size. 4087120631Snectar [Steve Henson] 4088120631Snectar 4089120631Snectar *) Various S/MIME bugfixes and compatibility changes: 4090120631Snectar output correct application/pkcs7 MIME type if 4091120631Snectar PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 4092120631Snectar Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 4093120631Snectar of files as .eml work). Correctly handle very long lines in MIME 4094120631Snectar parser. 4095120631Snectar [Steve Henson] 4096120631Snectar 4097120631Snectar Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 4098120631Snectar 4099120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 4100120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 4101120631Snectar a protocol version number mismatch like a decryption error 4102120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 4103120631Snectar [Bodo Moeller] 4104120631Snectar 4105120631Snectar *) Turn on RSA blinding by default in the default implementation 4106120631Snectar to avoid a timing attack. Applications that don't want it can call 4107120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 4108120631Snectar They would be ill-advised to do so in most cases. 4109120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 4110120631Snectar 4111120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 4112120631Snectar seeded (in this case, the secret RSA exponent is abused as 4113120631Snectar an unpredictable seed -- if it is not unpredictable, there 4114120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 4115120631Snectar by remembering the creator's thread ID in rsa->blinding and 4116120631Snectar having all other threads use local one-time blinding factors 4117120631Snectar (this requires more computation than sharing rsa->blinding, but 4118120631Snectar avoids excessive locking; and if an RSA object is not shared 4119120631Snectar between threads, blinding will still be very fast). 4120120631Snectar [Bodo Moeller] 4121120631Snectar 4122120631Snectar *) Fixed a typo bug that would cause ENGINE_set_default() to set an 4123120631Snectar ENGINE as defaults for all supported algorithms irrespective of 4124120631Snectar the 'flags' parameter. 'flags' is now honoured, so applications 4125120631Snectar should make sure they are passing it correctly. 4126120631Snectar [Geoff Thorpe] 4127120631Snectar 4128120631Snectar *) Target "mingw" now allows native Windows code to be generated in 4129120631Snectar the Cygwin environment as well as with the MinGW compiler. 4130120631Snectar [Ulf Moeller] 4131120631Snectar 4132111147Snectar Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 4133111147Snectar 4134111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 4135111147Snectar via timing by performing a MAC computation even if incorrrect 4136111147Snectar block cipher padding has been found. This is a countermeasure 4137111147Snectar against active attacks where the attacker has to distinguish 4138160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 4139111147Snectar 4140111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 4141111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 4142111147Snectar Martin Vuagnoux (EPFL, Ilion)] 4143111147Snectar 4144111147Snectar *) Make the no-err option work as intended. The intention with no-err 4145111147Snectar is not to have the whole error stack handling routines removed from 4146111147Snectar libcrypto, it's only intended to remove all the function name and 4147111147Snectar reason texts, thereby removing some of the footprint that may not 4148111147Snectar be interesting if those errors aren't displayed anyway. 4149111147Snectar 4150111147Snectar NOTE: it's still possible for any application or module to have it's 4151111147Snectar own set of error texts inserted. The routines are there, just not 4152111147Snectar used by default when no-err is given. 4153111147Snectar [Richard Levitte] 4154111147Snectar 4155111147Snectar *) Add support for FreeBSD on IA64. 4156111147Snectar [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 4157111147Snectar 4158111147Snectar *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 4159111147Snectar Kerberos function mit_des_cbc_cksum(). Before this change, 4160111147Snectar the value returned by DES_cbc_cksum() was like the one from 4161111147Snectar mit_des_cbc_cksum(), except the bytes were swapped. 4162111147Snectar [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 4163111147Snectar 4164111147Snectar *) Allow an application to disable the automatic SSL chain building. 4165111147Snectar Before this a rather primitive chain build was always performed in 4166111147Snectar ssl3_output_cert_chain(): an application had no way to send the 4167111147Snectar correct chain if the automatic operation produced an incorrect result. 4168111147Snectar 4169111147Snectar Now the chain builder is disabled if either: 4170111147Snectar 4171111147Snectar 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 4172111147Snectar 4173111147Snectar 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 4174111147Snectar 4175111147Snectar The reasoning behind this is that an application would not want the 4176111147Snectar auto chain building to take place if extra chain certificates are 4177111147Snectar present and it might also want a means of sending no additional 4178111147Snectar certificates (for example the chain has two certificates and the 4179111147Snectar root is omitted). 4180111147Snectar [Steve Henson] 4181111147Snectar 4182111147Snectar *) Add the possibility to build without the ENGINE framework. 4183111147Snectar [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 4184111147Snectar 4185111147Snectar *) Under Win32 gmtime() can return NULL: check return value in 4186111147Snectar OPENSSL_gmtime(). Add error code for case where gmtime() fails. 4187111147Snectar [Steve Henson] 4188111147Snectar 4189111147Snectar *) DSA routines: under certain error conditions uninitialized BN objects 4190111147Snectar could be freed. Solution: make sure initialization is performed early 4191111147Snectar enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 4192111147Snectar Nils Larsch <nla@trustcenter.de> via PR#459) 4193111147Snectar [Lutz Jaenicke] 4194111147Snectar 4195111147Snectar *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 4196111147Snectar checked on reconnect on the client side, therefore session resumption 4197111147Snectar could still fail with a "ssl session id is different" error. This 4198111147Snectar behaviour is masked when SSL_OP_ALL is used due to 4199111147Snectar SSL_OP_MICROSOFT_SESS_ID_BUG being set. 4200111147Snectar Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4201111147Snectar followup to PR #377. 4202111147Snectar [Lutz Jaenicke] 4203111147Snectar 4204111147Snectar *) IA-32 assembler support enhancements: unified ELF targets, support 4205111147Snectar for SCO/Caldera platforms, fix for Cygwin shared build. 4206111147Snectar [Andy Polyakov] 4207111147Snectar 4208111147Snectar *) Add support for FreeBSD on sparc64. As a consequence, support for 4209111147Snectar FreeBSD on non-x86 processors is separate from x86 processors on 4210111147Snectar the config script, much like the NetBSD support. 4211111147Snectar [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 4212111147Snectar 4213109998Smarkm Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 4214109998Smarkm 4215120631Snectar [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 4216120631Snectar OpenSSL 0.9.7.] 4217120631Snectar 4218109998Smarkm *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 4219109998Smarkm code (06) was taken as the first octet of the session ID and the last 4220109998Smarkm octet was ignored consequently. As a result SSLv2 client side session 4221109998Smarkm caching could not have worked due to the session ID mismatch between 4222109998Smarkm client and server. 4223109998Smarkm Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4224109998Smarkm PR #377. 4225109998Smarkm [Lutz Jaenicke] 4226109998Smarkm 4227109998Smarkm *) Change the declaration of needed Kerberos libraries to use EX_LIBS 4228109998Smarkm instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 4229109998Smarkm removed entirely. 4230109998Smarkm [Richard Levitte] 4231109998Smarkm 4232109998Smarkm *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 4233109998Smarkm seems that in spite of existing for more than a year, many application 4234109998Smarkm author have done nothing to provide the necessary callbacks, which 4235109998Smarkm means that this particular engine will not work properly anywhere. 4236109998Smarkm This is a very unfortunate situation which forces us, in the name 4237109998Smarkm of usability, to give the hw_ncipher.c a static lock, which is part 4238109998Smarkm of libcrypto. 4239109998Smarkm NOTE: This is for the 0.9.7 series ONLY. This hack will never 4240109998Smarkm appear in 0.9.8 or later. We EXPECT application authors to have 4241109998Smarkm dealt properly with this when 0.9.8 is released (unless we actually 4242109998Smarkm make such changes in the libcrypto locking code that changes will 4243109998Smarkm have to be made anyway). 4244109998Smarkm [Richard Levitte] 4245109998Smarkm 4246109998Smarkm *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 4247109998Smarkm octets have been read, EOF or an error occurs. Without this change 4248109998Smarkm some truncated ASN1 structures will not produce an error. 4249109998Smarkm [Steve Henson] 4250109998Smarkm 4251109998Smarkm *) Disable Heimdal support, since it hasn't been fully implemented. 4252109998Smarkm Still give the possibility to force the use of Heimdal, but with 4253109998Smarkm warnings and a request that patches get sent to openssl-dev. 4254109998Smarkm [Richard Levitte] 4255109998Smarkm 4256109998Smarkm *) Add the VC-CE target, introduce the WINCE sysname, and add 4257109998Smarkm INSTALL.WCE and appropriate conditionals to make it build. 4258109998Smarkm [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 4259109998Smarkm 4260109998Smarkm *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 4261109998Smarkm cygssl-x.y.z.dll, where x, y and z are the major, minor and 4262109998Smarkm edit numbers of the version. 4263109998Smarkm [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 4264109998Smarkm 4265109998Smarkm *) Introduce safe string copy and catenation functions 4266109998Smarkm (BUF_strlcpy() and BUF_strlcat()). 4267109998Smarkm [Ben Laurie (CHATS) and Richard Levitte] 4268109998Smarkm 4269109998Smarkm *) Avoid using fixed-size buffers for one-line DNs. 4270109998Smarkm [Ben Laurie (CHATS)] 4271109998Smarkm 4272109998Smarkm *) Add BUF_MEM_grow_clean() to avoid information leakage when 4273109998Smarkm resizing buffers containing secrets, and use where appropriate. 4274109998Smarkm [Ben Laurie (CHATS)] 4275109998Smarkm 4276109998Smarkm *) Avoid using fixed size buffers for configuration file location. 4277109998Smarkm [Ben Laurie (CHATS)] 4278109998Smarkm 4279109998Smarkm *) Avoid filename truncation for various CA files. 4280109998Smarkm [Ben Laurie (CHATS)] 4281109998Smarkm 4282109998Smarkm *) Use sizeof in preference to magic numbers. 4283109998Smarkm [Ben Laurie (CHATS)] 4284109998Smarkm 4285109998Smarkm *) Avoid filename truncation in cert requests. 4286109998Smarkm [Ben Laurie (CHATS)] 4287109998Smarkm 4288109998Smarkm *) Add assertions to check for (supposedly impossible) buffer 4289109998Smarkm overflows. 4290109998Smarkm [Ben Laurie (CHATS)] 4291109998Smarkm 4292109998Smarkm *) Don't cache truncated DNS entries in the local cache (this could 4293109998Smarkm potentially lead to a spoofing attack). 4294109998Smarkm [Ben Laurie (CHATS)] 4295109998Smarkm 4296109998Smarkm *) Fix various buffers to be large enough for hex/decimal 4297109998Smarkm representations in a platform independent manner. 4298109998Smarkm [Ben Laurie (CHATS)] 4299109998Smarkm 4300109998Smarkm *) Add CRYPTO_realloc_clean() to avoid information leakage when 4301109998Smarkm resizing buffers containing secrets, and use where appropriate. 4302109998Smarkm [Ben Laurie (CHATS)] 4303109998Smarkm 4304109998Smarkm *) Add BIO_indent() to avoid much slightly worrying code to do 4305109998Smarkm indents. 4306109998Smarkm [Ben Laurie (CHATS)] 4307109998Smarkm 4308109998Smarkm *) Convert sprintf()/BIO_puts() to BIO_printf(). 4309109998Smarkm [Ben Laurie (CHATS)] 4310109998Smarkm 4311109998Smarkm *) buffer_gets() could terminate with the buffer only half 4312109998Smarkm full. Fixed. 4313109998Smarkm [Ben Laurie (CHATS)] 4314109998Smarkm 4315109998Smarkm *) Add assertions to prevent user-supplied crypto functions from 4316109998Smarkm overflowing internal buffers by having large block sizes, etc. 4317109998Smarkm [Ben Laurie (CHATS)] 4318109998Smarkm 4319109998Smarkm *) New OPENSSL_assert() macro (similar to assert(), but enabled 4320109998Smarkm unconditionally). 4321109998Smarkm [Ben Laurie (CHATS)] 4322109998Smarkm 4323109998Smarkm *) Eliminate unused copy of key in RC4. 4324109998Smarkm [Ben Laurie (CHATS)] 4325109998Smarkm 4326109998Smarkm *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 4327109998Smarkm [Ben Laurie (CHATS)] 4328109998Smarkm 4329109998Smarkm *) Fix off-by-one error in EGD path. 4330109998Smarkm [Ben Laurie (CHATS)] 4331109998Smarkm 4332109998Smarkm *) If RANDFILE path is too long, ignore instead of truncating. 4333109998Smarkm [Ben Laurie (CHATS)] 4334109998Smarkm 4335109998Smarkm *) Eliminate unused and incorrectly sized X.509 structure 4336109998Smarkm CBCParameter. 4337109998Smarkm [Ben Laurie (CHATS)] 4338109998Smarkm 4339109998Smarkm *) Eliminate unused and dangerous function knumber(). 4340109998Smarkm [Ben Laurie (CHATS)] 4341109998Smarkm 4342109998Smarkm *) Eliminate unused and dangerous structure, KSSL_ERR. 4343109998Smarkm [Ben Laurie (CHATS)] 4344109998Smarkm 4345109998Smarkm *) Protect against overlong session ID context length in an encoded 4346109998Smarkm session object. Since these are local, this does not appear to be 4347109998Smarkm exploitable. 4348109998Smarkm [Ben Laurie (CHATS)] 4349109998Smarkm 4350109998Smarkm *) Change from security patch (see 0.9.6e below) that did not affect 4351109998Smarkm the 0.9.6 release series: 4352109998Smarkm 4353109998Smarkm Remote buffer overflow in SSL3 protocol - an attacker could 4354109998Smarkm supply an oversized master key in Kerberos-enabled versions. 4355160814Ssimon (CVE-2002-0657) 4356109998Smarkm [Ben Laurie (CHATS)] 4357109998Smarkm 4358109998Smarkm *) Change the SSL kerb5 codes to match RFC 2712. 4359109998Smarkm [Richard Levitte] 4360109998Smarkm 4361109998Smarkm *) Make -nameopt work fully for req and add -reqopt switch. 4362109998Smarkm [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 4363109998Smarkm 4364109998Smarkm *) The "block size" for block ciphers in CFB and OFB mode should be 1. 4365109998Smarkm [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 4366109998Smarkm 4367109998Smarkm *) Make sure tests can be performed even if the corresponding algorithms 4368109998Smarkm have been removed entirely. This was also the last step to make 4369109998Smarkm OpenSSL compilable with DJGPP under all reasonable conditions. 4370109998Smarkm [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 4371109998Smarkm 4372109998Smarkm *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 4373109998Smarkm to allow version independent disabling of normally unselected ciphers, 4374109998Smarkm which may be activated as a side-effect of selecting a single cipher. 4375109998Smarkm 4376109998Smarkm (E.g., cipher list string "RSA" enables ciphersuites that are left 4377109998Smarkm out of "ALL" because they do not provide symmetric encryption. 4378109998Smarkm "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 4379109998Smarkm [Lutz Jaenicke, Bodo Moeller] 4380109998Smarkm 4381109998Smarkm *) Add appropriate support for separate platform-dependent build 4382109998Smarkm directories. The recommended way to make a platform-dependent 4383109998Smarkm build directory is the following (tested on Linux), maybe with 4384109998Smarkm some local tweaks: 4385109998Smarkm 4386109998Smarkm # Place yourself outside of the OpenSSL source tree. In 4387109998Smarkm # this example, the environment variable OPENSSL_SOURCE 4388109998Smarkm # is assumed to contain the absolute OpenSSL source directory. 4389109998Smarkm mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 4390109998Smarkm cd objtree/"`uname -s`-`uname -r`-`uname -m`" 4391111147Snectar (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 4392109998Smarkm mkdir -p `dirname $F` 4393109998Smarkm ln -s $OPENSSL_SOURCE/$F $F 4394109998Smarkm done 4395109998Smarkm 4396109998Smarkm To be absolutely sure not to disturb the source tree, a "make clean" 4397109998Smarkm is a good thing. If it isn't successfull, don't worry about it, 4398109998Smarkm it probably means the source directory is very clean. 4399109998Smarkm [Richard Levitte] 4400109998Smarkm 4401109998Smarkm *) Make sure any ENGINE control commands make local copies of string 4402109998Smarkm pointers passed to them whenever necessary. Otherwise it is possible 4403109998Smarkm the caller may have overwritten (or deallocated) the original string 4404109998Smarkm data when a later ENGINE operation tries to use the stored values. 4405109998Smarkm [G�tz Babin-Ebell <babinebell@trustcenter.de>] 4406109998Smarkm 4407109998Smarkm *) Improve diagnostics in file reading and command-line digests. 4408109998Smarkm [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 4409109998Smarkm 4410109998Smarkm *) Add AES modes CFB and OFB to the object database. Correct an 4411109998Smarkm error in AES-CFB decryption. 4412109998Smarkm [Richard Levitte] 4413109998Smarkm 4414109998Smarkm *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 4415109998Smarkm allows existing EVP_CIPHER_CTX structures to be reused after 4416109998Smarkm calling EVP_*Final(). This behaviour is used by encryption 4417109998Smarkm BIOs and some applications. This has the side effect that 4418109998Smarkm applications must explicitly clean up cipher contexts with 4419109998Smarkm EVP_CIPHER_CTX_cleanup() or they will leak memory. 4420109998Smarkm [Steve Henson] 4421109998Smarkm 4422109998Smarkm *) Check the values of dna and dnb in bn_mul_recursive before calling 4423109998Smarkm bn_mul_comba (a non zero value means the a or b arrays do not contain 4424109998Smarkm n2 elements) and fallback to bn_mul_normal if either is not zero. 4425109998Smarkm [Steve Henson] 4426109998Smarkm 4427109998Smarkm *) Fix escaping of non-ASCII characters when using the -subj option 4428109998Smarkm of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 4429109998Smarkm [Lutz Jaenicke] 4430109998Smarkm 4431109998Smarkm *) Make object definitions compliant to LDAP (RFC2256): SN is the short 4432109998Smarkm form for "surname", serialNumber has no short form. 4433109998Smarkm Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 4434109998Smarkm therefore remove "mail" short name for "internet 7". 4435109998Smarkm The OID for unique identifiers in X509 certificates is 4436109998Smarkm x500UniqueIdentifier, not uniqueIdentifier. 4437109998Smarkm Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 4438109998Smarkm [Lutz Jaenicke] 4439109998Smarkm 4440109998Smarkm *) Add an "init" command to the ENGINE config module and auto initialize 4441109998Smarkm ENGINEs. Without any "init" command the ENGINE will be initialized 4442109998Smarkm after all ctrl commands have been executed on it. If init=1 the 4443109998Smarkm ENGINE is initailized at that point (ctrls before that point are run 4444109998Smarkm on the uninitialized ENGINE and after on the initialized one). If 4445109998Smarkm init=0 then the ENGINE will not be iniatialized at all. 4446109998Smarkm [Steve Henson] 4447109998Smarkm 4448109998Smarkm *) Fix the 'app_verify_callback' interface so that the user-defined 4449109998Smarkm argument is actually passed to the callback: In the 4450109998Smarkm SSL_CTX_set_cert_verify_callback() prototype, the callback 4451109998Smarkm declaration has been changed from 4452109998Smarkm int (*cb)() 4453109998Smarkm into 4454109998Smarkm int (*cb)(X509_STORE_CTX *,void *); 4455109998Smarkm in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 4456109998Smarkm i=s->ctx->app_verify_callback(&ctx) 4457109998Smarkm has been changed into 4458109998Smarkm i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 4459109998Smarkm 4460109998Smarkm To update applications using SSL_CTX_set_cert_verify_callback(), 4461109998Smarkm a dummy argument can be added to their callback functions. 4462109998Smarkm [D. K. Smetters <smetters@parc.xerox.com>] 4463109998Smarkm 4464109998Smarkm *) Added the '4758cca' ENGINE to support IBM 4758 cards. 4465109998Smarkm [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 4466109998Smarkm 4467109998Smarkm *) Add and OPENSSL_LOAD_CONF define which will cause 4468109998Smarkm OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 4469109998Smarkm This allows older applications to transparently support certain 4470109998Smarkm OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 4471109998Smarkm Two new functions OPENSSL_add_all_algorithms_noconf() which will never 4472109998Smarkm load the config file and OPENSSL_add_all_algorithms_conf() which will 4473109998Smarkm always load it have also been added. 4474109998Smarkm [Steve Henson] 4475109998Smarkm 4476109998Smarkm *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 4477109998Smarkm Adjust NIDs and EVP layer. 4478109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4479109998Smarkm 4480109998Smarkm *) Config modules support in openssl utility. 4481109998Smarkm 4482109998Smarkm Most commands now load modules from the config file, 4483109998Smarkm though in a few (such as version) this isn't done 4484109998Smarkm because it couldn't be used for anything. 4485109998Smarkm 4486109998Smarkm In the case of ca and req the config file used is 4487109998Smarkm the same as the utility itself: that is the -config 4488109998Smarkm command line option can be used to specify an 4489109998Smarkm alternative file. 4490109998Smarkm [Steve Henson] 4491109998Smarkm 4492109998Smarkm *) Move default behaviour from OPENSSL_config(). If appname is NULL 4493109998Smarkm use "openssl_conf" if filename is NULL use default openssl config file. 4494109998Smarkm [Steve Henson] 4495109998Smarkm 4496109998Smarkm *) Add an argument to OPENSSL_config() to allow the use of an alternative 4497109998Smarkm config section name. Add a new flag to tolerate a missing config file 4498109998Smarkm and move code to CONF_modules_load_file(). 4499109998Smarkm [Steve Henson] 4500109998Smarkm 4501109998Smarkm *) Support for crypto accelerator cards from Accelerated Encryption 4502109998Smarkm Processing, www.aep.ie. (Use engine 'aep') 4503109998Smarkm The support was copied from 0.9.6c [engine] and adapted/corrected 4504109998Smarkm to work with the new engine framework. 4505109998Smarkm [AEP Inc. and Richard Levitte] 4506109998Smarkm 4507109998Smarkm *) Support for SureWare crypto accelerator cards from Baltimore 4508109998Smarkm Technologies. (Use engine 'sureware') 4509109998Smarkm The support was copied from 0.9.6c [engine] and adapted 4510109998Smarkm to work with the new engine framework. 4511109998Smarkm [Richard Levitte] 4512109998Smarkm 4513109998Smarkm *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 4514109998Smarkm make the newer ENGINE framework commands for the CHIL engine work. 4515109998Smarkm [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 4516109998Smarkm 4517109998Smarkm *) Make it possible to produce shared libraries on ReliantUNIX. 4518109998Smarkm [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 4519109998Smarkm 4520109998Smarkm *) Add the configuration target debug-linux-ppro. 4521109998Smarkm Make 'openssl rsa' use the general key loading routines 4522109998Smarkm implemented in apps.c, and make those routines able to 4523109998Smarkm handle the key format FORMAT_NETSCAPE and the variant 4524109998Smarkm FORMAT_IISSGC. 4525109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4526109998Smarkm 4527109998Smarkm *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 4528109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4529109998Smarkm 4530109998Smarkm *) Add -keyform to rsautl, and document -engine. 4531109998Smarkm [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 4532109998Smarkm 4533109998Smarkm *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 4534109998Smarkm BIO_R_NO_SUCH_FILE error code rather than the generic 4535109998Smarkm ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 4536109998Smarkm [Ben Laurie] 4537109998Smarkm 4538109998Smarkm *) Add new functions 4539109998Smarkm ERR_peek_last_error 4540109998Smarkm ERR_peek_last_error_line 4541109998Smarkm ERR_peek_last_error_line_data. 4542109998Smarkm These are similar to 4543109998Smarkm ERR_peek_error 4544109998Smarkm ERR_peek_error_line 4545109998Smarkm ERR_peek_error_line_data, 4546109998Smarkm but report on the latest error recorded rather than the first one 4547109998Smarkm still in the error queue. 4548109998Smarkm [Ben Laurie, Bodo Moeller] 4549109998Smarkm 4550109998Smarkm *) default_algorithms option in ENGINE config module. This allows things 4551109998Smarkm like: 4552109998Smarkm default_algorithms = ALL 4553109998Smarkm default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 4554109998Smarkm [Steve Henson] 4555109998Smarkm 4556109998Smarkm *) Prelminary ENGINE config module. 4557109998Smarkm [Steve Henson] 4558109998Smarkm 4559109998Smarkm *) New experimental application configuration code. 4560109998Smarkm [Steve Henson] 4561109998Smarkm 4562109998Smarkm *) Change the AES code to follow the same name structure as all other 4563109998Smarkm symmetric ciphers, and behave the same way. Move everything to 4564109998Smarkm the directory crypto/aes, thereby obsoleting crypto/rijndael. 4565109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4566109998Smarkm 4567109998Smarkm *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 4568109998Smarkm [Ben Laurie and Theo de Raadt] 4569109998Smarkm 4570109998Smarkm *) Add option to output public keys in req command. 4571109998Smarkm [Massimiliano Pala madwolf@openca.org] 4572109998Smarkm 4573109998Smarkm *) Use wNAFs in EC_POINTs_mul() for improved efficiency 4574109998Smarkm (up to about 10% better than before for P-192 and P-224). 4575109998Smarkm [Bodo Moeller] 4576109998Smarkm 4577109998Smarkm *) New functions/macros 4578109998Smarkm 4579109998Smarkm SSL_CTX_set_msg_callback(ctx, cb) 4580109998Smarkm SSL_CTX_set_msg_callback_arg(ctx, arg) 4581109998Smarkm SSL_set_msg_callback(ssl, cb) 4582109998Smarkm SSL_set_msg_callback_arg(ssl, arg) 4583109998Smarkm 4584109998Smarkm to request calling a callback function 4585109998Smarkm 4586109998Smarkm void cb(int write_p, int version, int content_type, 4587109998Smarkm const void *buf, size_t len, SSL *ssl, void *arg) 4588109998Smarkm 4589109998Smarkm whenever a protocol message has been completely received 4590109998Smarkm (write_p == 0) or sent (write_p == 1). Here 'version' is the 4591109998Smarkm protocol version according to which the SSL library interprets 4592109998Smarkm the current protocol message (SSL2_VERSION, SSL3_VERSION, or 4593109998Smarkm TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 4594109998Smarkm the content type as defined in the SSL 3.0/TLS 1.0 protocol 4595109998Smarkm specification (change_cipher_spec(20), alert(21), handshake(22)). 4596109998Smarkm 'buf' and 'len' point to the actual message, 'ssl' to the 4597109998Smarkm SSL object, and 'arg' is the application-defined value set by 4598109998Smarkm SSL[_CTX]_set_msg_callback_arg(). 4599109998Smarkm 4600109998Smarkm 'openssl s_client' and 'openssl s_server' have new '-msg' options 4601109998Smarkm to enable a callback that displays all protocol messages. 4602109998Smarkm [Bodo Moeller] 4603109998Smarkm 4604109998Smarkm *) Change the shared library support so shared libraries are built as 4605109998Smarkm soon as the corresponding static library is finished, and thereby get 4606109998Smarkm openssl and the test programs linked against the shared library. 4607109998Smarkm This still only happens when the keyword "shard" has been given to 4608109998Smarkm the configuration scripts. 4609109998Smarkm 4610109998Smarkm NOTE: shared library support is still an experimental thing, and 4611109998Smarkm backward binary compatibility is still not guaranteed. 4612109998Smarkm ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 4613109998Smarkm 4614109998Smarkm *) Add support for Subject Information Access extension. 4615109998Smarkm [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 4616109998Smarkm 4617109998Smarkm *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 4618109998Smarkm additional bytes when new memory had to be allocated, not just 4619109998Smarkm when reusing an existing buffer. 4620109998Smarkm [Bodo Moeller] 4621109998Smarkm 4622109998Smarkm *) New command line and configuration option 'utf8' for the req command. 4623109998Smarkm This allows field values to be specified as UTF8 strings. 4624109998Smarkm [Steve Henson] 4625109998Smarkm 4626109998Smarkm *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 4627109998Smarkm runs for the former and machine-readable output for the latter. 4628109998Smarkm [Ben Laurie] 4629109998Smarkm 4630109998Smarkm *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 4631109998Smarkm of the e-mail address in the DN (i.e., it will go into a certificate 4632109998Smarkm extension only). The new configuration file option 'email_in_dn = no' 4633109998Smarkm has the same effect. 4634109998Smarkm [Massimiliano Pala madwolf@openca.org] 4635109998Smarkm 4636109998Smarkm *) Change all functions with names starting with des_ to be starting 4637109998Smarkm with DES_ instead. Add wrappers that are compatible with libdes, 4638109998Smarkm but are named _ossl_old_des_*. Finally, add macros that map the 4639109998Smarkm des_* symbols to the corresponding _ossl_old_des_* if libdes 4640109998Smarkm compatibility is desired. If OpenSSL 0.9.6c compatibility is 4641109998Smarkm desired, the des_* symbols will be mapped to DES_*, with one 4642109998Smarkm exception. 4643109998Smarkm 4644109998Smarkm Since we provide two compatibility mappings, the user needs to 4645109998Smarkm define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 4646109998Smarkm compatibility is desired. The default (i.e., when that macro 4647109998Smarkm isn't defined) is OpenSSL 0.9.6c compatibility. 4648109998Smarkm 4649109998Smarkm There are also macros that enable and disable the support of old 4650109998Smarkm des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 4651109998Smarkm and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 4652109998Smarkm are defined, the default will apply: to support the old des routines. 4653109998Smarkm 4654109998Smarkm In either case, one must include openssl/des.h to get the correct 4655109998Smarkm definitions. Do not try to just include openssl/des_old.h, that 4656109998Smarkm won't work. 4657109998Smarkm 4658109998Smarkm NOTE: This is a major break of an old API into a new one. Software 4659109998Smarkm authors are encouraged to switch to the DES_ style functions. Some 4660109998Smarkm time in the future, des_old.h and the libdes compatibility functions 4661109998Smarkm will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 4662109998Smarkm default), and then completely removed. 4663109998Smarkm [Richard Levitte] 4664109998Smarkm 4665109998Smarkm *) Test for certificates which contain unsupported critical extensions. 4666109998Smarkm If such a certificate is found during a verify operation it is 4667109998Smarkm rejected by default: this behaviour can be overridden by either 4668109998Smarkm handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 4669109998Smarkm by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 4670109998Smarkm X509_supported_extension() has also been added which returns 1 if a 4671109998Smarkm particular extension is supported. 4672109998Smarkm [Steve Henson] 4673109998Smarkm 4674109998Smarkm *) Modify the behaviour of EVP cipher functions in similar way to digests 4675109998Smarkm to retain compatibility with existing code. 4676109998Smarkm [Steve Henson] 4677109998Smarkm 4678109998Smarkm *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 4679109998Smarkm compatibility with existing code. In particular the 'ctx' parameter does 4680109998Smarkm not have to be to be initialized before the call to EVP_DigestInit() and 4681109998Smarkm it is tidied up after a call to EVP_DigestFinal(). New function 4682109998Smarkm EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 4683109998Smarkm EVP_MD_CTX_copy() changed to not require the destination to be 4684109998Smarkm initialized valid and new function EVP_MD_CTX_copy_ex() added which 4685109998Smarkm requires the destination to be valid. 4686109998Smarkm 4687109998Smarkm Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 4688109998Smarkm EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 4689109998Smarkm [Steve Henson] 4690109998Smarkm 4691109998Smarkm *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 4692109998Smarkm so that complete 'Handshake' protocol structures are kept in memory 4693109998Smarkm instead of overwriting 'msg_type' and 'length' with 'body' data. 4694109998Smarkm [Bodo Moeller] 4695109998Smarkm 4696109998Smarkm *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 4697109998Smarkm [Massimo Santin via Richard Levitte] 4698109998Smarkm 4699109998Smarkm *) Major restructuring to the underlying ENGINE code. This includes 4700109998Smarkm reduction of linker bloat, separation of pure "ENGINE" manipulation 4701109998Smarkm (initialisation, etc) from functionality dealing with implementations 4702109998Smarkm of specific crypto iterfaces. This change also introduces integrated 4703109998Smarkm support for symmetric ciphers and digest implementations - so ENGINEs 4704109998Smarkm can now accelerate these by providing EVP_CIPHER and EVP_MD 4705109998Smarkm implementations of their own. This is detailed in crypto/engine/README 4706109998Smarkm as it couldn't be adequately described here. However, there are a few 4707109998Smarkm API changes worth noting - some RSA, DSA, DH, and RAND functions that 4708109998Smarkm were changed in the original introduction of ENGINE code have now 4709109998Smarkm reverted back - the hooking from this code to ENGINE is now a good 4710109998Smarkm deal more passive and at run-time, operations deal directly with 4711109998Smarkm RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 4712109998Smarkm dereferencing through an ENGINE pointer any more. Also, the ENGINE 4713109998Smarkm functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 4714109998Smarkm they were not being used by the framework as there is no concept of a 4715109998Smarkm BIGNUM_METHOD and they could not be generalised to the new 4716109998Smarkm 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 4717109998Smarkm ENGINE_cpy() has been removed as it cannot be consistently defined in 4718109998Smarkm the new code. 4719109998Smarkm [Geoff Thorpe] 4720109998Smarkm 4721109998Smarkm *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 4722109998Smarkm [Steve Henson] 4723109998Smarkm 4724109998Smarkm *) Change mkdef.pl to sort symbols that get the same entry number, 4725109998Smarkm and make sure the automatically generated functions ERR_load_* 4726109998Smarkm become part of libeay.num as well. 4727109998Smarkm [Richard Levitte] 4728109998Smarkm 4729109998Smarkm *) New function SSL_renegotiate_pending(). This returns true once 4730109998Smarkm renegotiation has been requested (either SSL_renegotiate() call 4731109998Smarkm or HelloRequest/ClientHello receveived from the peer) and becomes 4732109998Smarkm false once a handshake has been completed. 4733109998Smarkm (For servers, SSL_renegotiate() followed by SSL_do_handshake() 4734109998Smarkm sends a HelloRequest, but does not ensure that a handshake takes 4735109998Smarkm place. SSL_renegotiate_pending() is useful for checking if the 4736109998Smarkm client has followed the request.) 4737109998Smarkm [Bodo Moeller] 4738109998Smarkm 4739109998Smarkm *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 4740109998Smarkm By default, clients may request session resumption even during 4741109998Smarkm renegotiation (if session ID contexts permit); with this option, 4742109998Smarkm session resumption is possible only in the first handshake. 4743109998Smarkm 4744109998Smarkm SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 4745109998Smarkm more bits available for options that should not be part of 4746109998Smarkm SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 4747109998Smarkm [Bodo Moeller] 4748109998Smarkm 4749109998Smarkm *) Add some demos for certificate and certificate request creation. 4750109998Smarkm [Steve Henson] 4751109998Smarkm 4752109998Smarkm *) Make maximum certificate chain size accepted from the peer application 4753109998Smarkm settable (SSL*_get/set_max_cert_list()), as proposed by 4754109998Smarkm "Douglas E. Engert" <deengert@anl.gov>. 4755109998Smarkm [Lutz Jaenicke] 4756109998Smarkm 4757109998Smarkm *) Add support for shared libraries for Unixware-7 4758109998Smarkm (Boyd Lynn Gerber <gerberb@zenez.com>). 4759109998Smarkm [Lutz Jaenicke] 4760109998Smarkm 4761109998Smarkm *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 4762109998Smarkm be done prior to destruction. Use this to unload error strings from 4763109998Smarkm ENGINEs that load their own error strings. NB: This adds two new API 4764109998Smarkm functions to "get" and "set" this destroy handler in an ENGINE. 4765109998Smarkm [Geoff Thorpe] 4766109998Smarkm 4767109998Smarkm *) Alter all existing ENGINE implementations (except "openssl" and 4768109998Smarkm "openbsd") to dynamically instantiate their own error strings. This 4769109998Smarkm makes them more flexible to be built both as statically-linked ENGINEs 4770109998Smarkm and self-contained shared-libraries loadable via the "dynamic" ENGINE. 4771109998Smarkm Also, add stub code to each that makes building them as self-contained 4772109998Smarkm shared-libraries easier (see README.ENGINE). 4773109998Smarkm [Geoff Thorpe] 4774109998Smarkm 4775109998Smarkm *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 4776109998Smarkm implementations into applications that are completely implemented in 4777109998Smarkm self-contained shared-libraries. The "dynamic" ENGINE exposes control 4778109998Smarkm commands that can be used to configure what shared-library to load and 4779109998Smarkm to control aspects of the way it is handled. Also, made an update to 4780109998Smarkm the README.ENGINE file that brings its information up-to-date and 4781109998Smarkm provides some information and instructions on the "dynamic" ENGINE 4782109998Smarkm (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 4783109998Smarkm [Geoff Thorpe] 4784109998Smarkm 4785109998Smarkm *) Make it possible to unload ranges of ERR strings with a new 4786109998Smarkm "ERR_unload_strings" function. 4787109998Smarkm [Geoff Thorpe] 4788109998Smarkm 4789109998Smarkm *) Add a copy() function to EVP_MD. 4790109998Smarkm [Ben Laurie] 4791109998Smarkm 4792109998Smarkm *) Make EVP_MD routines take a context pointer instead of just the 4793109998Smarkm md_data void pointer. 4794109998Smarkm [Ben Laurie] 4795109998Smarkm 4796109998Smarkm *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 4797109998Smarkm that the digest can only process a single chunk of data 4798109998Smarkm (typically because it is provided by a piece of 4799109998Smarkm hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 4800109998Smarkm is only going to provide a single chunk of data, and hence the 4801109998Smarkm framework needn't accumulate the data for oneshot drivers. 4802109998Smarkm [Ben Laurie] 4803109998Smarkm 4804109998Smarkm *) As with "ERR", make it possible to replace the underlying "ex_data" 4805109998Smarkm functions. This change also alters the storage and management of global 4806109998Smarkm ex_data state - it's now all inside ex_data.c and all "class" code (eg. 4807109998Smarkm RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 4808109998Smarkm index counters. The API functions that use this state have been changed 4809109998Smarkm to take a "class_index" rather than pointers to the class's local STACK 4810109998Smarkm and counter, and there is now an API function to dynamically create new 4811109998Smarkm classes. This centralisation allows us to (a) plug a lot of the 4812109998Smarkm thread-safety problems that existed, and (b) makes it possible to clean 4813109998Smarkm up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 4814109998Smarkm such data would previously have always leaked in application code and 4815109998Smarkm workarounds were in place to make the memory debugging turn a blind eye 4816109998Smarkm to it. Application code that doesn't use this new function will still 4817109998Smarkm leak as before, but their memory debugging output will announce it now 4818109998Smarkm rather than letting it slide. 4819109998Smarkm 4820109998Smarkm Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 4821109998Smarkm induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 4822109998Smarkm has a return value to indicate success or failure. 4823109998Smarkm [Geoff Thorpe] 4824109998Smarkm 4825109998Smarkm *) Make it possible to replace the underlying "ERR" functions such that the 4826109998Smarkm global state (2 LHASH tables and 2 locks) is only used by the "default" 4827109998Smarkm implementation. This change also adds two functions to "get" and "set" 4828109998Smarkm the implementation prior to it being automatically set the first time 4829109998Smarkm any other ERR function takes place. Ie. an application can call "get", 4830109998Smarkm pass the return value to a module it has just loaded, and that module 4831109998Smarkm can call its own "set" function using that value. This means the 4832109998Smarkm module's "ERR" operations will use (and modify) the error state in the 4833109998Smarkm application and not in its own statically linked copy of OpenSSL code. 4834109998Smarkm [Geoff Thorpe] 4835109998Smarkm 4836109998Smarkm *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 4837109998Smarkm reference counts. This performs normal REF_PRINT/REF_CHECK macros on 4838109998Smarkm the operation, and provides a more encapsulated way for external code 4839109998Smarkm (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 4840109998Smarkm to use these functions rather than manually incrementing the counts. 4841109998Smarkm 4842109998Smarkm Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 4843109998Smarkm [Geoff Thorpe] 4844109998Smarkm 4845109998Smarkm *) Add EVP test program. 4846109998Smarkm [Ben Laurie] 4847109998Smarkm 4848109998Smarkm *) Add symmetric cipher support to ENGINE. Expect the API to change! 4849109998Smarkm [Ben Laurie] 4850109998Smarkm 4851109998Smarkm *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 4852109998Smarkm X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 4853109998Smarkm X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 4854109998Smarkm These allow a CRL to be built without having to access X509_CRL fields 4855109998Smarkm directly. Modify 'ca' application to use new functions. 4856109998Smarkm [Steve Henson] 4857109998Smarkm 4858109998Smarkm *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 4859109998Smarkm bug workarounds. Rollback attack detection is a security feature. 4860109998Smarkm The problem will only arise on OpenSSL servers when TLSv1 is not 4861109998Smarkm available (sslv3_server_method() or SSL_OP_NO_TLSv1). 4862109998Smarkm Software authors not wanting to support TLSv1 will have special reasons 4863109998Smarkm for their choice and can explicitly enable this option. 4864109998Smarkm [Bodo Moeller, Lutz Jaenicke] 4865109998Smarkm 4866109998Smarkm *) Rationalise EVP so it can be extended: don't include a union of 4867109998Smarkm cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 4868109998Smarkm (similar to those existing for EVP_CIPHER_CTX). 4869109998Smarkm Usage example: 4870109998Smarkm 4871109998Smarkm EVP_MD_CTX md; 4872109998Smarkm 4873109998Smarkm EVP_MD_CTX_init(&md); /* new function call */ 4874109998Smarkm EVP_DigestInit(&md, EVP_sha1()); 4875109998Smarkm EVP_DigestUpdate(&md, in, len); 4876109998Smarkm EVP_DigestFinal(&md, out, NULL); 4877109998Smarkm EVP_MD_CTX_cleanup(&md); /* new function call */ 4878109998Smarkm 4879109998Smarkm [Ben Laurie] 4880109998Smarkm 4881109998Smarkm *) Make DES key schedule conform to the usual scheme, as well as 4882109998Smarkm correcting its structure. This means that calls to DES functions 4883109998Smarkm now have to pass a pointer to a des_key_schedule instead of a 4884109998Smarkm plain des_key_schedule (which was actually always a pointer 4885109998Smarkm anyway): E.g., 4886109998Smarkm 4887109998Smarkm des_key_schedule ks; 4888109998Smarkm 4889109998Smarkm des_set_key_checked(..., &ks); 4890109998Smarkm des_ncbc_encrypt(..., &ks, ...); 4891109998Smarkm 4892109998Smarkm (Note that a later change renames 'des_...' into 'DES_...'.) 4893109998Smarkm [Ben Laurie] 4894109998Smarkm 4895109998Smarkm *) Initial reduction of linker bloat: the use of some functions, such as 4896109998Smarkm PEM causes large amounts of unused functions to be linked in due to 4897109998Smarkm poor organisation. For example pem_all.c contains every PEM function 4898109998Smarkm which has a knock on effect of linking in large amounts of (unused) 4899109998Smarkm ASN1 code. Grouping together similar functions and splitting unrelated 4900109998Smarkm functions prevents this. 4901109998Smarkm [Steve Henson] 4902109998Smarkm 4903109998Smarkm *) Cleanup of EVP macros. 4904109998Smarkm [Ben Laurie] 4905109998Smarkm 4906109998Smarkm *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 4907109998Smarkm correct _ecb suffix. 4908109998Smarkm [Ben Laurie] 4909109998Smarkm 4910109998Smarkm *) Add initial OCSP responder support to ocsp application. The 4911109998Smarkm revocation information is handled using the text based index 4912109998Smarkm use by the ca application. The responder can either handle 4913109998Smarkm requests generated internally, supplied in files (for example 4914109998Smarkm via a CGI script) or using an internal minimal server. 4915109998Smarkm [Steve Henson] 4916109998Smarkm 4917109998Smarkm *) Add configuration choices to get zlib compression for TLS. 4918109998Smarkm [Richard Levitte] 4919109998Smarkm 4920109998Smarkm *) Changes to Kerberos SSL for RFC 2712 compliance: 4921109998Smarkm 1. Implemented real KerberosWrapper, instead of just using 4922109998Smarkm KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 4923109998Smarkm 2. Implemented optional authenticator field of KerberosWrapper. 4924109998Smarkm 4925109998Smarkm Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 4926109998Smarkm and authenticator structs; see crypto/krb5/. 4927109998Smarkm 4928109998Smarkm Generalized Kerberos calls to support multiple Kerberos libraries. 4929109998Smarkm [Vern Staats <staatsvr@asc.hpc.mil>, 4930109998Smarkm Jeffrey Altman <jaltman@columbia.edu> 4931109998Smarkm via Richard Levitte] 4932109998Smarkm 4933109998Smarkm *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 4934109998Smarkm already does with RSA. testdsa.h now has 'priv_key/pub_key' 4935109998Smarkm values for each of the key sizes rather than having just 4936109998Smarkm parameters (and 'speed' generating keys each time). 4937109998Smarkm [Geoff Thorpe] 4938109998Smarkm 4939109998Smarkm *) Speed up EVP routines. 4940109998Smarkm Before: 4941109998Smarkmencrypt 4942109998Smarkmtype 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 4943109998Smarkmdes-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 4944109998Smarkmdes-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 4945109998Smarkmdes-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 4946109998Smarkmdecrypt 4947109998Smarkmdes-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 4948109998Smarkmdes-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 4949109998Smarkmdes-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 4950109998Smarkm After: 4951109998Smarkmencrypt 4952109998Smarkmdes-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 4953109998Smarkmdecrypt 4954109998Smarkmdes-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 4955109998Smarkm [Ben Laurie] 4956109998Smarkm 4957109998Smarkm *) Added the OS2-EMX target. 4958109998Smarkm ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 4959109998Smarkm 4960109998Smarkm *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 4961109998Smarkm to support NCONF routines in extension code. New function CONF_set_nconf() 4962109998Smarkm to allow functions which take an NCONF to also handle the old LHASH 4963109998Smarkm structure: this means that the old CONF compatible routines can be 4964109998Smarkm retained (in particular wrt extensions) without having to duplicate the 4965109998Smarkm code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 4966109998Smarkm [Steve Henson] 4967109998Smarkm 4968109998Smarkm *) Enhance the general user interface with mechanisms for inner control 4969109998Smarkm and with possibilities to have yes/no kind of prompts. 4970109998Smarkm [Richard Levitte] 4971109998Smarkm 4972109998Smarkm *) Change all calls to low level digest routines in the library and 4973109998Smarkm applications to use EVP. Add missing calls to HMAC_cleanup() and 4974109998Smarkm don't assume HMAC_CTX can be copied using memcpy(). 4975109998Smarkm [Verdon Walker <VWalker@novell.com>, Steve Henson] 4976109998Smarkm 4977109998Smarkm *) Add the possibility to control engines through control names but with 4978109998Smarkm arbitrary arguments instead of just a string. 4979109998Smarkm Change the key loaders to take a UI_METHOD instead of a callback 4980109998Smarkm function pointer. NOTE: this breaks binary compatibility with earlier 4981109998Smarkm versions of OpenSSL [engine]. 4982109998Smarkm Adapt the nCipher code for these new conditions and add a card insertion 4983109998Smarkm callback. 4984109998Smarkm [Richard Levitte] 4985109998Smarkm 4986109998Smarkm *) Enhance the general user interface with mechanisms to better support 4987109998Smarkm dialog box interfaces, application-defined prompts, the possibility 4988109998Smarkm to use defaults (for example default passwords from somewhere else) 4989109998Smarkm and interrupts/cancellations. 4990109998Smarkm [Richard Levitte] 4991109998Smarkm 4992109998Smarkm *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 4993109998Smarkm attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 4994109998Smarkm [Steve Henson] 4995109998Smarkm 4996109998Smarkm *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 4997109998Smarkm tidy up some unnecessarily weird code in 'sk_new()'). 4998109998Smarkm [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 4999109998Smarkm 5000109998Smarkm *) Change the key loading routines for ENGINEs to use the same kind 5001109998Smarkm callback (pem_password_cb) as all other routines that need this 5002109998Smarkm kind of callback. 5003109998Smarkm [Richard Levitte] 5004109998Smarkm 5005109998Smarkm *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 5006109998Smarkm 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 5007109998Smarkm than this minimum value is recommended. 5008109998Smarkm [Lutz Jaenicke] 5009109998Smarkm 5010109998Smarkm *) New random seeder for OpenVMS, using the system process statistics 5011109998Smarkm that are easily reachable. 5012109998Smarkm [Richard Levitte] 5013109998Smarkm 5014109998Smarkm *) Windows apparently can't transparently handle global 5015109998Smarkm variables defined in DLLs. Initialisations such as: 5016109998Smarkm 5017109998Smarkm const ASN1_ITEM *it = &ASN1_INTEGER_it; 5018109998Smarkm 5019109998Smarkm wont compile. This is used by the any applications that need to 5020109998Smarkm declare their own ASN1 modules. This was fixed by adding the option 5021109998Smarkm EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 5022109998Smarkm needed for static libraries under Win32. 5023109998Smarkm [Steve Henson] 5024109998Smarkm 5025109998Smarkm *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 5026109998Smarkm setting of purpose and trust fields. New X509_STORE trust and 5027109998Smarkm purpose functions and tidy up setting in other SSL functions. 5028109998Smarkm [Steve Henson] 5029109998Smarkm 5030109998Smarkm *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 5031109998Smarkm structure. These are inherited by X509_STORE_CTX when it is 5032109998Smarkm initialised. This allows various defaults to be set in the 5033109998Smarkm X509_STORE structure (such as flags for CRL checking and custom 5034109998Smarkm purpose or trust settings) for functions which only use X509_STORE_CTX 5035109998Smarkm internally such as S/MIME. 5036109998Smarkm 5037109998Smarkm Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 5038109998Smarkm trust settings if they are not set in X509_STORE. This allows X509_STORE 5039109998Smarkm purposes and trust (in S/MIME for example) to override any set by default. 5040109998Smarkm 5041109998Smarkm Add command line options for CRL checking to smime, s_client and s_server 5042109998Smarkm applications. 5043109998Smarkm [Steve Henson] 5044109998Smarkm 5045109998Smarkm *) Initial CRL based revocation checking. If the CRL checking flag(s) 5046109998Smarkm are set then the CRL is looked up in the X509_STORE structure and 5047109998Smarkm its validity and signature checked, then if the certificate is found 5048109998Smarkm in the CRL the verify fails with a revoked error. 5049109998Smarkm 5050109998Smarkm Various new CRL related callbacks added to X509_STORE_CTX structure. 5051109998Smarkm 5052109998Smarkm Command line options added to 'verify' application to support this. 5053109998Smarkm 5054109998Smarkm This needs some additional work, such as being able to handle multiple 5055109998Smarkm CRLs with different times, extension based lookup (rather than just 5056109998Smarkm by subject name) and ultimately more complete V2 CRL extension 5057109998Smarkm handling. 5058109998Smarkm [Steve Henson] 5059109998Smarkm 5060109998Smarkm *) Add a general user interface API (crypto/ui/). This is designed 5061109998Smarkm to replace things like des_read_password and friends (backward 5062109998Smarkm compatibility functions using this new API are provided). 5063109998Smarkm The purpose is to remove prompting functions from the DES code 5064109998Smarkm section as well as provide for prompting through dialog boxes in 5065109998Smarkm a window system and the like. 5066109998Smarkm [Richard Levitte] 5067109998Smarkm 5068109998Smarkm *) Add "ex_data" support to ENGINE so implementations can add state at a 5069109998Smarkm per-structure level rather than having to store it globally. 5070109998Smarkm [Geoff] 5071109998Smarkm 5072109998Smarkm *) Make it possible for ENGINE structures to be copied when retrieved by 5073109998Smarkm ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 5074109998Smarkm This causes the "original" ENGINE structure to act like a template, 5075109998Smarkm analogous to the RSA vs. RSA_METHOD type of separation. Because of this 5076109998Smarkm operational state can be localised to each ENGINE structure, despite the 5077109998Smarkm fact they all share the same "methods". New ENGINE structures returned in 5078109998Smarkm this case have no functional references and the return value is the single 5079109998Smarkm structural reference. This matches the single structural reference returned 5080109998Smarkm by ENGINE_by_id() normally, when it is incremented on the pre-existing 5081109998Smarkm ENGINE structure. 5082109998Smarkm [Geoff] 5083109998Smarkm 5084109998Smarkm *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 5085109998Smarkm needs to match any other type at all we need to manually clear the 5086109998Smarkm tag cache. 5087109998Smarkm [Steve Henson] 5088109998Smarkm 5089109998Smarkm *) Changes to the "openssl engine" utility to include; 5090109998Smarkm - verbosity levels ('-v', '-vv', and '-vvv') that provide information 5091109998Smarkm about an ENGINE's available control commands. 5092109998Smarkm - executing control commands from command line arguments using the 5093109998Smarkm '-pre' and '-post' switches. '-post' is only used if '-t' is 5094109998Smarkm specified and the ENGINE is successfully initialised. The syntax for 5095109998Smarkm the individual commands are colon-separated, for example; 5096109998Smarkm openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 5097109998Smarkm [Geoff] 5098109998Smarkm 5099109998Smarkm *) New dynamic control command support for ENGINEs. ENGINEs can now 5100109998Smarkm declare their own commands (numbers), names (strings), descriptions, 5101109998Smarkm and input types for run-time discovery by calling applications. A 5102109998Smarkm subset of these commands are implicitly classed as "executable" 5103109998Smarkm depending on their input type, and only these can be invoked through 5104109998Smarkm the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 5105109998Smarkm can be based on user input, config files, etc). The distinction is 5106109998Smarkm that "executable" commands cannot return anything other than a boolean 5107109998Smarkm result and can only support numeric or string input, whereas some 5108109998Smarkm discoverable commands may only be for direct use through 5109109998Smarkm ENGINE_ctrl(), eg. supporting the exchange of binary data, function 5110109998Smarkm pointers, or other custom uses. The "executable" commands are to 5111109998Smarkm support parameterisations of ENGINE behaviour that can be 5112109998Smarkm unambiguously defined by ENGINEs and used consistently across any 5113109998Smarkm OpenSSL-based application. Commands have been added to all the 5114109998Smarkm existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 5115109998Smarkm control over shared-library paths without source code alterations. 5116109998Smarkm [Geoff] 5117109998Smarkm 5118109998Smarkm *) Changed all ENGINE implementations to dynamically allocate their 5119109998Smarkm ENGINEs rather than declaring them statically. Apart from this being 5120109998Smarkm necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 5121109998Smarkm this also allows the implementations to compile without using the 5122109998Smarkm internal engine_int.h header. 5123109998Smarkm [Geoff] 5124109998Smarkm 5125109998Smarkm *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 5126109998Smarkm 'const' value. Any code that should be able to modify a RAND_METHOD 5127109998Smarkm should already have non-const pointers to it (ie. they should only 5128109998Smarkm modify their own ones). 5129109998Smarkm [Geoff] 5130109998Smarkm 5131109998Smarkm *) Made a variety of little tweaks to the ENGINE code. 5132109998Smarkm - "atalla" and "ubsec" string definitions were moved from header files 5133109998Smarkm to C code. "nuron" string definitions were placed in variables 5134109998Smarkm rather than hard-coded - allowing parameterisation of these values 5135109998Smarkm later on via ctrl() commands. 5136109998Smarkm - Removed unused "#if 0"'d code. 5137109998Smarkm - Fixed engine list iteration code so it uses ENGINE_free() to release 5138109998Smarkm structural references. 5139109998Smarkm - Constified the RAND_METHOD element of ENGINE structures. 5140109998Smarkm - Constified various get/set functions as appropriate and added 5141109998Smarkm missing functions (including a catch-all ENGINE_cpy that duplicates 5142109998Smarkm all ENGINE values onto a new ENGINE except reference counts/state). 5143109998Smarkm - Removed NULL parameter checks in get/set functions. Setting a method 5144109998Smarkm or function to NULL is a way of cancelling out a previously set 5145109998Smarkm value. Passing a NULL ENGINE parameter is just plain stupid anyway 5146109998Smarkm and doesn't justify the extra error symbols and code. 5147109998Smarkm - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 5148109998Smarkm flags from engine_int.h to engine.h. 5149109998Smarkm - Changed prototypes for ENGINE handler functions (init(), finish(), 5150109998Smarkm ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 5151109998Smarkm [Geoff] 5152109998Smarkm 5153109998Smarkm *) Implement binary inversion algorithm for BN_mod_inverse in addition 5154109998Smarkm to the algorithm using long division. The binary algorithm can be 5155109998Smarkm used only if the modulus is odd. On 32-bit systems, it is faster 5156109998Smarkm only for relatively small moduli (roughly 20-30% for 128-bit moduli, 5157109998Smarkm roughly 5-15% for 256-bit moduli), so we use it only for moduli 5158109998Smarkm up to 450 bits. In 64-bit environments, the binary algorithm 5159109998Smarkm appears to be advantageous for much longer moduli; here we use it 5160109998Smarkm for moduli up to 2048 bits. 5161109998Smarkm [Bodo Moeller] 5162109998Smarkm 5163109998Smarkm *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 5164109998Smarkm could not support the combine flag in choice fields. 5165109998Smarkm [Steve Henson] 5166109998Smarkm 5167109998Smarkm *) Add a 'copy_extensions' option to the 'ca' utility. This copies 5168109998Smarkm extensions from a certificate request to the certificate. 5169109998Smarkm [Steve Henson] 5170109998Smarkm 5171109998Smarkm *) Allow multiple 'certopt' and 'nameopt' options to be separated 5172109998Smarkm by commas. Add 'namopt' and 'certopt' options to the 'ca' config 5173109998Smarkm file: this allows the display of the certificate about to be 5174109998Smarkm signed to be customised, to allow certain fields to be included 5175109998Smarkm or excluded and extension details. The old system didn't display 5176109998Smarkm multicharacter strings properly, omitted fields not in the policy 5177109998Smarkm and couldn't display additional details such as extensions. 5178109998Smarkm [Steve Henson] 5179109998Smarkm 5180109998Smarkm *) Function EC_POINTs_mul for multiple scalar multiplication 5181109998Smarkm of an arbitrary number of elliptic curve points 5182109998Smarkm \sum scalars[i]*points[i], 5183109998Smarkm optionally including the generator defined for the EC_GROUP: 5184109998Smarkm scalar*generator + \sum scalars[i]*points[i]. 5185109998Smarkm 5186109998Smarkm EC_POINT_mul is a simple wrapper function for the typical case 5187109998Smarkm that the point list has just one item (besides the optional 5188109998Smarkm generator). 5189109998Smarkm [Bodo Moeller] 5190109998Smarkm 5191109998Smarkm *) First EC_METHODs for curves over GF(p): 5192109998Smarkm 5193109998Smarkm EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 5194109998Smarkm operations and provides various method functions that can also 5195109998Smarkm operate with faster implementations of modular arithmetic. 5196109998Smarkm 5197109998Smarkm EC_GFp_mont_method() reuses most functions that are part of 5198109998Smarkm EC_GFp_simple_method, but uses Montgomery arithmetic. 5199109998Smarkm 5200109998Smarkm [Bodo Moeller; point addition and point doubling 5201109998Smarkm implementation directly derived from source code provided by 5202109998Smarkm Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 5203109998Smarkm 5204109998Smarkm *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 5205109998Smarkm crypto/ec/ec_lib.c): 5206109998Smarkm 5207109998Smarkm Curves are EC_GROUP objects (with an optional group generator) 5208109998Smarkm based on EC_METHODs that are built into the library. 5209109998Smarkm 5210109998Smarkm Points are EC_POINT objects based on EC_GROUP objects. 5211109998Smarkm 5212109998Smarkm Most of the framework would be able to handle curves over arbitrary 5213109998Smarkm finite fields, but as there are no obvious types for fields other 5214109998Smarkm than GF(p), some functions are limited to that for now. 5215109998Smarkm [Bodo Moeller] 5216109998Smarkm 5217109998Smarkm *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 5218109998Smarkm that the file contains a complete HTTP response. 5219109998Smarkm [Richard Levitte] 5220109998Smarkm 5221109998Smarkm *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 5222109998Smarkm change the def and num file printf format specifier from "%-40sXXX" 5223109998Smarkm to "%-39s XXX". The latter will always guarantee a space after the 5224109998Smarkm field while the former will cause them to run together if the field 5225109998Smarkm is 40 of more characters long. 5226109998Smarkm [Steve Henson] 5227109998Smarkm 5228109998Smarkm *) Constify the cipher and digest 'method' functions and structures 5229109998Smarkm and modify related functions to take constant EVP_MD and EVP_CIPHER 5230109998Smarkm pointers. 5231109998Smarkm [Steve Henson] 5232109998Smarkm 5233109998Smarkm *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 5234109998Smarkm in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 5235109998Smarkm [Bodo Moeller] 5236109998Smarkm 5237109998Smarkm *) Modify EVP_Digest*() routines so they now return values. Although the 5238109998Smarkm internal software routines can never fail additional hardware versions 5239109998Smarkm might. 5240109998Smarkm [Steve Henson] 5241109998Smarkm 5242109998Smarkm *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 5243109998Smarkm 5244109998Smarkm Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 5245109998Smarkm (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 5246109998Smarkm 5247109998Smarkm ASN1 error codes 5248109998Smarkm ERR_R_NESTED_ASN1_ERROR 5249109998Smarkm ... 5250109998Smarkm ERR_R_MISSING_ASN1_EOS 5251109998Smarkm were 4 .. 9, conflicting with 5252109998Smarkm ERR_LIB_RSA (= ERR_R_RSA_LIB) 5253109998Smarkm ... 5254109998Smarkm ERR_LIB_PEM (= ERR_R_PEM_LIB). 5255109998Smarkm They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 5256109998Smarkm 5257109998Smarkm Add new error code 'ERR_R_INTERNAL_ERROR'. 5258109998Smarkm [Bodo Moeller] 5259109998Smarkm 5260109998Smarkm *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 5261109998Smarkm suffices. 5262109998Smarkm [Bodo Moeller] 5263109998Smarkm 5264109998Smarkm *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 5265109998Smarkm sets the subject name for a new request or supersedes the 5266109998Smarkm subject name in a given request. Formats that can be parsed are 5267109998Smarkm 'CN=Some Name, OU=myOU, C=IT' 5268109998Smarkm and 5269109998Smarkm 'CN=Some Name/OU=myOU/C=IT'. 5270109998Smarkm 5271109998Smarkm Add options '-batch' and '-verbose' to 'openssl req'. 5272109998Smarkm [Massimiliano Pala <madwolf@hackmasters.net>] 5273109998Smarkm 5274109998Smarkm *) Introduce the possibility to access global variables through 5275109998Smarkm functions on platform were that's the best way to handle exporting 5276109998Smarkm global variables in shared libraries. To enable this functionality, 5277109998Smarkm one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 5278109998Smarkm "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 5279109998Smarkm is normally done by Configure or something similar). 5280109998Smarkm 5281109998Smarkm To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 5282109998Smarkm in the source file (foo.c) like this: 5283109998Smarkm 5284109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 5285109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(double,bar); 5286109998Smarkm 5287109998Smarkm To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 5288109998Smarkm and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 5289109998Smarkm 5290109998Smarkm OPENSSL_DECLARE_GLOBAL(int,foo); 5291109998Smarkm #define foo OPENSSL_GLOBAL_REF(foo) 5292109998Smarkm OPENSSL_DECLARE_GLOBAL(double,bar); 5293109998Smarkm #define bar OPENSSL_GLOBAL_REF(bar) 5294109998Smarkm 5295109998Smarkm The #defines are very important, and therefore so is including the 5296109998Smarkm header file everywhere where the defined globals are used. 5297109998Smarkm 5298109998Smarkm The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 5299109998Smarkm of ASN.1 items, but that structure is a bit different. 5300109998Smarkm 5301109998Smarkm The largest change is in util/mkdef.pl which has been enhanced with 5302109998Smarkm better and easier to understand logic to choose which symbols should 5303109998Smarkm go into the Windows .def files as well as a number of fixes and code 5304109998Smarkm cleanup (among others, algorithm keywords are now sorted 5305109998Smarkm lexicographically to avoid constant rewrites). 5306109998Smarkm [Richard Levitte] 5307109998Smarkm 5308109998Smarkm *) In BN_div() keep a copy of the sign of 'num' before writing the 5309109998Smarkm result to 'rm' because if rm==num the value will be overwritten 5310109998Smarkm and produce the wrong result if 'num' is negative: this caused 5311109998Smarkm problems with BN_mod() and BN_nnmod(). 5312109998Smarkm [Steve Henson] 5313109998Smarkm 5314109998Smarkm *) Function OCSP_request_verify(). This checks the signature on an 5315109998Smarkm OCSP request and verifies the signer certificate. The signer 5316109998Smarkm certificate is just checked for a generic purpose and OCSP request 5317109998Smarkm trust settings. 5318109998Smarkm [Steve Henson] 5319109998Smarkm 5320109998Smarkm *) Add OCSP_check_validity() function to check the validity of OCSP 5321109998Smarkm responses. OCSP responses are prepared in real time and may only 5322109998Smarkm be a few seconds old. Simply checking that the current time lies 5323109998Smarkm between thisUpdate and nextUpdate max reject otherwise valid responses 5324109998Smarkm caused by either OCSP responder or client clock inaccuracy. Instead 5325109998Smarkm we allow thisUpdate and nextUpdate to fall within a certain period of 5326109998Smarkm the current time. The age of the response can also optionally be 5327109998Smarkm checked. Two new options -validity_period and -status_age added to 5328109998Smarkm ocsp utility. 5329109998Smarkm [Steve Henson] 5330109998Smarkm 5331109998Smarkm *) If signature or public key algorithm is unrecognized print out its 5332109998Smarkm OID rather that just UNKNOWN. 5333109998Smarkm [Steve Henson] 5334109998Smarkm 5335109998Smarkm *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 5336109998Smarkm OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 5337109998Smarkm ID to be generated from the issuer certificate alone which can then be 5338109998Smarkm passed to OCSP_id_issuer_cmp(). 5339109998Smarkm [Steve Henson] 5340109998Smarkm 5341109998Smarkm *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 5342109998Smarkm ASN1 modules to export functions returning ASN1_ITEM pointers 5343109998Smarkm instead of the ASN1_ITEM structures themselves. This adds several 5344109998Smarkm new macros which allow the underlying ASN1 function/structure to 5345109998Smarkm be accessed transparently. As a result code should not use ASN1_ITEM 5346109998Smarkm references directly (such as &X509_it) but instead use the relevant 5347109998Smarkm macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 5348109998Smarkm use of the new ASN1 code on platforms where exporting structures 5349109998Smarkm is problematical (for example in shared libraries) but exporting 5350109998Smarkm functions returning pointers to structures is not. 5351109998Smarkm [Steve Henson] 5352109998Smarkm 5353109998Smarkm *) Add support for overriding the generation of SSL/TLS session IDs. 5354109998Smarkm These callbacks can be registered either in an SSL_CTX or per SSL. 5355109998Smarkm The purpose of this is to allow applications to control, if they wish, 5356109998Smarkm the arbitrary values chosen for use as session IDs, particularly as it 5357109998Smarkm can be useful for session caching in multiple-server environments. A 5358109998Smarkm command-line switch for testing this (and any client code that wishes 5359109998Smarkm to use such a feature) has been added to "s_server". 5360109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 5361109998Smarkm 5362109998Smarkm *) Modify mkdef.pl to recognise and parse preprocessor conditionals 5363109998Smarkm of the form '#if defined(...) || defined(...) || ...' and 5364109998Smarkm '#if !defined(...) && !defined(...) && ...'. This also avoids 5365109998Smarkm the growing number of special cases it was previously handling. 5366109998Smarkm [Richard Levitte] 5367109998Smarkm 5368109998Smarkm *) Make all configuration macros available for application by making 5369109998Smarkm sure they are available in opensslconf.h, by giving them names starting 5370109998Smarkm with "OPENSSL_" to avoid conflicts with other packages and by making 5371109998Smarkm sure e_os2.h will cover all platform-specific cases together with 5372109998Smarkm opensslconf.h. 5373109998Smarkm Additionally, it is now possible to define configuration/platform- 5374109998Smarkm specific names (called "system identities"). In the C code, these 5375109998Smarkm are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 5376109998Smarkm macro with the name beginning with "OPENSSL_SYS_", which is determined 5377109998Smarkm from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 5378109998Smarkm what is available. 5379109998Smarkm [Richard Levitte] 5380109998Smarkm 5381109998Smarkm *) New option -set_serial to 'req' and 'x509' this allows the serial 5382109998Smarkm number to use to be specified on the command line. Previously self 5383109998Smarkm signed certificates were hard coded with serial number 0 and the 5384109998Smarkm CA options of 'x509' had to use a serial number in a file which was 5385109998Smarkm auto incremented. 5386109998Smarkm [Steve Henson] 5387109998Smarkm 5388109998Smarkm *) New options to 'ca' utility to support V2 CRL entry extensions. 5389109998Smarkm Currently CRL reason, invalidity date and hold instruction are 5390109998Smarkm supported. Add new CRL extensions to V3 code and some new objects. 5391109998Smarkm [Steve Henson] 5392109998Smarkm 5393109998Smarkm *) New function EVP_CIPHER_CTX_set_padding() this is used to 5394109998Smarkm disable standard block padding (aka PKCS#5 padding) in the EVP 5395109998Smarkm API, which was previously mandatory. This means that the data is 5396109998Smarkm not padded in any way and so the total length much be a multiple 5397109998Smarkm of the block size, otherwise an error occurs. 5398109998Smarkm [Steve Henson] 5399109998Smarkm 5400109998Smarkm *) Initial (incomplete) OCSP SSL support. 5401109998Smarkm [Steve Henson] 5402109998Smarkm 5403109998Smarkm *) New function OCSP_parse_url(). This splits up a URL into its host, 5404109998Smarkm port and path components: primarily to parse OCSP URLs. New -url 5405109998Smarkm option to ocsp utility. 5406109998Smarkm [Steve Henson] 5407109998Smarkm 5408109998Smarkm *) New nonce behavior. The return value of OCSP_check_nonce() now 5409109998Smarkm reflects the various checks performed. Applications can decide 5410109998Smarkm whether to tolerate certain situations such as an absent nonce 5411109998Smarkm in a response when one was present in a request: the ocsp application 5412109998Smarkm just prints out a warning. New function OCSP_add1_basic_nonce() 5413109998Smarkm this is to allow responders to include a nonce in a response even if 5414109998Smarkm the request is nonce-less. 5415109998Smarkm [Steve Henson] 5416109998Smarkm 5417109998Smarkm *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 5418109998Smarkm skipped when using openssl x509 multiple times on a single input file, 5419109998Smarkm e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 5420109998Smarkm [Bodo Moeller] 5421109998Smarkm 5422109998Smarkm *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 5423109998Smarkm set string type: to handle setting ASN1_TIME structures. Fix ca 5424109998Smarkm utility to correctly initialize revocation date of CRLs. 5425109998Smarkm [Steve Henson] 5426109998Smarkm 5427109998Smarkm *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 5428109998Smarkm the clients preferred ciphersuites and rather use its own preferences. 5429109998Smarkm Should help to work around M$ SGC (Server Gated Cryptography) bug in 5430109998Smarkm Internet Explorer by ensuring unchanged hash method during stepup. 5431109998Smarkm (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 5432109998Smarkm [Lutz Jaenicke] 5433109998Smarkm 5434109998Smarkm *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 5435109998Smarkm to aes and add a new 'exist' option to print out symbols that don't 5436109998Smarkm appear to exist. 5437109998Smarkm [Steve Henson] 5438109998Smarkm 5439109998Smarkm *) Additional options to ocsp utility to allow flags to be set and 5440109998Smarkm additional certificates supplied. 5441109998Smarkm [Steve Henson] 5442109998Smarkm 5443109998Smarkm *) Add the option -VAfile to 'openssl ocsp', so the user can give the 5444109998Smarkm OCSP client a number of certificate to only verify the response 5445109998Smarkm signature against. 5446109998Smarkm [Richard Levitte] 5447109998Smarkm 5448109998Smarkm *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 5449109998Smarkm handle the new API. Currently only ECB, CBC modes supported. Add new 5450109998Smarkm AES OIDs. 5451109998Smarkm 5452109998Smarkm Add TLS AES ciphersuites as described in RFC3268, "Advanced 5453109998Smarkm Encryption Standard (AES) Ciphersuites for Transport Layer 5454109998Smarkm Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 5455109998Smarkm not enabled by default and were not part of the "ALL" ciphersuite 5456109998Smarkm alias because they were not yet official; they could be 5457109998Smarkm explicitly requested by specifying the "AESdraft" ciphersuite 5458109998Smarkm group alias. In the final release of OpenSSL 0.9.7, the group 5459109998Smarkm alias is called "AES" and is part of "ALL".) 5460109998Smarkm [Ben Laurie, Steve Henson, Bodo Moeller] 5461109998Smarkm 5462109998Smarkm *) New function OCSP_copy_nonce() to copy nonce value (if present) from 5463109998Smarkm request to response. 5464109998Smarkm [Steve Henson] 5465109998Smarkm 5466109998Smarkm *) Functions for OCSP responders. OCSP_request_onereq_count(), 5467109998Smarkm OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 5468109998Smarkm extract information from a certificate request. OCSP_response_create() 5469109998Smarkm creates a response and optionally adds a basic response structure. 5470109998Smarkm OCSP_basic_add1_status() adds a complete single response to a basic 5471109998Smarkm response and returns the OCSP_SINGLERESP structure just added (to allow 5472109998Smarkm extensions to be included for example). OCSP_basic_add1_cert() adds a 5473109998Smarkm certificate to a basic response and OCSP_basic_sign() signs a basic 5474109998Smarkm response with various flags. New helper functions ASN1_TIME_check() 5475109998Smarkm (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 5476109998Smarkm (converts ASN1_TIME to GeneralizedTime). 5477109998Smarkm [Steve Henson] 5478109998Smarkm 5479109998Smarkm *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 5480109998Smarkm in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 5481109998Smarkm structure from a certificate. X509_pubkey_digest() digests the public_key 5482109998Smarkm contents: this is used in various key identifiers. 5483109998Smarkm [Steve Henson] 5484109998Smarkm 5485109998Smarkm *) Make sk_sort() tolerate a NULL argument. 5486109998Smarkm [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 5487109998Smarkm 5488109998Smarkm *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 5489109998Smarkm passed by the function are trusted implicitly. If any of them signed the 5490109998Smarkm response then it is assumed to be valid and is not verified. 5491109998Smarkm [Steve Henson] 5492109998Smarkm 5493109998Smarkm *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 5494109998Smarkm to data. This was previously part of the PKCS7 ASN1 code. This 5495109998Smarkm was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 5496109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 5497109998Smarkm <support@securenetterm.com>] 5498109998Smarkm 5499109998Smarkm *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 5500109998Smarkm routines: without these tracing memory leaks is very painful. 5501109998Smarkm Fix leaks in PKCS12 and PKCS7 routines. 5502109998Smarkm [Steve Henson] 5503109998Smarkm 5504109998Smarkm *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 5505109998Smarkm Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 5506109998Smarkm effectively meant GeneralizedTime would never be used. Now it 5507109998Smarkm is initialised to -1 but X509_time_adj() now has to check the value 5508109998Smarkm and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 5509109998Smarkm V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 5510109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 5511109998Smarkm <support@securenetterm.com>] 5512109998Smarkm 5513109998Smarkm *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 5514109998Smarkm result in a zero length in the ASN1_INTEGER structure which was 5515109998Smarkm not consistent with the structure when d2i_ASN1_INTEGER() was used 5516109998Smarkm and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 5517109998Smarkm to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 5518109998Smarkm where it did not print out a minus for negative ASN1_INTEGER. 5519109998Smarkm [Steve Henson] 5520109998Smarkm 5521109998Smarkm *) Add summary printout to ocsp utility. The various functions which 5522109998Smarkm convert status values to strings have been renamed to: 5523109998Smarkm OCSP_response_status_str(), OCSP_cert_status_str() and 5524109998Smarkm OCSP_crl_reason_str() and are no longer static. New options 5525109998Smarkm to verify nonce values and to disable verification. OCSP response 5526109998Smarkm printout format cleaned up. 5527109998Smarkm [Steve Henson] 5528109998Smarkm 5529109998Smarkm *) Add additional OCSP certificate checks. These are those specified 5530109998Smarkm in RFC2560. This consists of two separate checks: the CA of the 5531109998Smarkm certificate being checked must either be the OCSP signer certificate 5532109998Smarkm or the issuer of the OCSP signer certificate. In the latter case the 5533109998Smarkm OCSP signer certificate must contain the OCSP signing extended key 5534109998Smarkm usage. This check is performed by attempting to match the OCSP 5535109998Smarkm signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 5536109998Smarkm in the OCSP_CERTID structures of the response. 5537109998Smarkm [Steve Henson] 5538109998Smarkm 5539109998Smarkm *) Initial OCSP certificate verification added to OCSP_basic_verify() 5540109998Smarkm and related routines. This uses the standard OpenSSL certificate 5541109998Smarkm verify routines to perform initial checks (just CA validity) and 5542109998Smarkm to obtain the certificate chain. Then additional checks will be 5543109998Smarkm performed on the chain. Currently the root CA is checked to see 5544109998Smarkm if it is explicitly trusted for OCSP signing. This is used to set 5545109998Smarkm a root CA as a global signing root: that is any certificate that 5546109998Smarkm chains to that CA is an acceptable OCSP signing certificate. 5547109998Smarkm [Steve Henson] 5548109998Smarkm 5549109998Smarkm *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 5550109998Smarkm extensions from a separate configuration file. 5551109998Smarkm As when reading extensions from the main configuration file, 5552109998Smarkm the '-extensions ...' option may be used for specifying the 5553109998Smarkm section to use. 5554109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 5555109998Smarkm 5556109998Smarkm *) New OCSP utility. Allows OCSP requests to be generated or 5557109998Smarkm read. The request can be sent to a responder and the output 5558109998Smarkm parsed, outputed or printed in text form. Not complete yet: 5559109998Smarkm still needs to check the OCSP response validity. 5560109998Smarkm [Steve Henson] 5561109998Smarkm 5562109998Smarkm *) New subcommands for 'openssl ca': 5563109998Smarkm 'openssl ca -status <serial>' prints the status of the cert with 5564109998Smarkm the given serial number (according to the index file). 5565109998Smarkm 'openssl ca -updatedb' updates the expiry status of certificates 5566109998Smarkm in the index file. 5567109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 5568109998Smarkm 5569109998Smarkm *) New '-newreq-nodes' command option to CA.pl. This is like 5570109998Smarkm '-newreq', but calls 'openssl req' with the '-nodes' option 5571109998Smarkm so that the resulting key is not encrypted. 5572109998Smarkm [Damien Miller <djm@mindrot.org>] 5573109998Smarkm 5574109998Smarkm *) New configuration for the GNU Hurd. 5575109998Smarkm [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 5576109998Smarkm 5577109998Smarkm *) Initial code to implement OCSP basic response verify. This 5578109998Smarkm is currently incomplete. Currently just finds the signer's 5579109998Smarkm certificate and verifies the signature on the response. 5580109998Smarkm [Steve Henson] 5581109998Smarkm 5582109998Smarkm *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 5583109998Smarkm value of OPENSSLDIR. This is available via the new '-d' option 5584109998Smarkm to 'openssl version', and is also included in 'openssl version -a'. 5585109998Smarkm [Bodo Moeller] 5586109998Smarkm 5587109998Smarkm *) Allowing defining memory allocation callbacks that will be given 5588109998Smarkm file name and line number information in additional arguments 5589109998Smarkm (a const char* and an int). The basic functionality remains, as 5590109998Smarkm well as the original possibility to just replace malloc(), 5591109998Smarkm realloc() and free() by functions that do not know about these 5592109998Smarkm additional arguments. To register and find out the current 5593109998Smarkm settings for extended allocation functions, the following 5594109998Smarkm functions are provided: 5595109998Smarkm 5596109998Smarkm CRYPTO_set_mem_ex_functions 5597109998Smarkm CRYPTO_set_locked_mem_ex_functions 5598109998Smarkm CRYPTO_get_mem_ex_functions 5599109998Smarkm CRYPTO_get_locked_mem_ex_functions 5600109998Smarkm 5601109998Smarkm These work the same way as CRYPTO_set_mem_functions and friends. 5602109998Smarkm CRYPTO_get_[locked_]mem_functions now writes 0 where such an 5603109998Smarkm extended allocation function is enabled. 5604109998Smarkm Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 5605109998Smarkm a conventional allocation function is enabled. 5606109998Smarkm [Richard Levitte, Bodo Moeller] 5607109998Smarkm 5608109998Smarkm *) Finish off removing the remaining LHASH function pointer casts. 5609109998Smarkm There should no longer be any prototype-casting required when using 5610109998Smarkm the LHASH abstraction, and any casts that remain are "bugs". See 5611109998Smarkm the callback types and macros at the head of lhash.h for details 5612109998Smarkm (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 5613109998Smarkm [Geoff Thorpe] 5614109998Smarkm 5615109998Smarkm *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 5616109998Smarkm If /dev/[u]random devices are not available or do not return enough 5617109998Smarkm entropy, EGD style sockets (served by EGD or PRNGD) will automatically 5618109998Smarkm be queried. 5619109998Smarkm The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 5620109998Smarkm /etc/entropy will be queried once each in this sequence, quering stops 5621109998Smarkm when enough entropy was collected without querying more sockets. 5622109998Smarkm [Lutz Jaenicke] 5623109998Smarkm 5624109998Smarkm *) Change the Unix RAND_poll() variant to be able to poll several 5625109998Smarkm random devices, as specified by DEVRANDOM, until a sufficient amount 5626109998Smarkm of data has been collected. We spend at most 10 ms on each file 5627109998Smarkm (select timeout) and read in non-blocking mode. DEVRANDOM now 5628109998Smarkm defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 5629109998Smarkm (previously it was just the string "/dev/urandom"), so on typical 5630109998Smarkm platforms the 10 ms delay will never occur. 5631109998Smarkm Also separate out the Unix variant to its own file, rand_unix.c. 5632109998Smarkm For VMS, there's a currently-empty rand_vms.c. 5633109998Smarkm [Richard Levitte] 5634109998Smarkm 5635109998Smarkm *) Move OCSP client related routines to ocsp_cl.c. These 5636109998Smarkm provide utility functions which an application needing 5637109998Smarkm to issue a request to an OCSP responder and analyse the 5638109998Smarkm response will typically need: as opposed to those which an 5639109998Smarkm OCSP responder itself would need which will be added later. 5640109998Smarkm 5641109998Smarkm OCSP_request_sign() signs an OCSP request with an API similar 5642109998Smarkm to PKCS7_sign(). OCSP_response_status() returns status of OCSP 5643109998Smarkm response. OCSP_response_get1_basic() extracts basic response 5644109998Smarkm from response. OCSP_resp_find_status(): finds and extracts status 5645109998Smarkm information from an OCSP_CERTID structure (which will be created 5646109998Smarkm when the request structure is built). These are built from lower 5647109998Smarkm level functions which work on OCSP_SINGLERESP structures but 5648109998Smarkm wont normally be used unless the application wishes to examine 5649109998Smarkm extensions in the OCSP response for example. 5650109998Smarkm 5651109998Smarkm Replace nonce routines with a pair of functions. 5652109998Smarkm OCSP_request_add1_nonce() adds a nonce value and optionally 5653109998Smarkm generates a random value. OCSP_check_nonce() checks the 5654109998Smarkm validity of the nonce in an OCSP response. 5655109998Smarkm [Steve Henson] 5656109998Smarkm 5657109998Smarkm *) Change function OCSP_request_add() to OCSP_request_add0_id(). 5658109998Smarkm This doesn't copy the supplied OCSP_CERTID and avoids the 5659109998Smarkm need to free up the newly created id. Change return type 5660109998Smarkm to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 5661109998Smarkm This can then be used to add extensions to the request. 5662109998Smarkm Deleted OCSP_request_new(), since most of its functionality 5663109998Smarkm is now in OCSP_REQUEST_new() (and the case insensitive name 5664109998Smarkm clash) apart from the ability to set the request name which 5665109998Smarkm will be added elsewhere. 5666109998Smarkm [Steve Henson] 5667109998Smarkm 5668109998Smarkm *) Update OCSP API. Remove obsolete extensions argument from 5669109998Smarkm various functions. Extensions are now handled using the new 5670109998Smarkm OCSP extension code. New simple OCSP HTTP function which 5671109998Smarkm can be used to send requests and parse the response. 5672109998Smarkm [Steve Henson] 5673109998Smarkm 5674109998Smarkm *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 5675109998Smarkm ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 5676109998Smarkm uses the special reorder version of SET OF to sort the attributes 5677109998Smarkm and reorder them to match the encoded order. This resolves a long 5678109998Smarkm standing problem: a verify on a PKCS7 structure just after signing 5679109998Smarkm it used to fail because the attribute order did not match the 5680109998Smarkm encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 5681109998Smarkm it uses the received order. This is necessary to tolerate some broken 5682109998Smarkm software that does not order SET OF. This is handled by encoding 5683109998Smarkm as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 5684109998Smarkm to produce the required SET OF. 5685109998Smarkm [Steve Henson] 5686109998Smarkm 5687109998Smarkm *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 5688109998Smarkm OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 5689109998Smarkm files to get correct declarations of the ASN.1 item variables. 5690109998Smarkm [Richard Levitte] 5691109998Smarkm 5692109998Smarkm *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 5693109998Smarkm PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 5694109998Smarkm asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 5695109998Smarkm NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 5696109998Smarkm New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 5697109998Smarkm ASN1_ITEM and no wrapper functions. 5698109998Smarkm [Steve Henson] 5699109998Smarkm 5700109998Smarkm *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 5701109998Smarkm replace the old function pointer based I/O routines. Change most of 5702109998Smarkm the *_d2i_bio() and *_d2i_fp() functions to use these. 5703109998Smarkm [Steve Henson] 5704109998Smarkm 5705109998Smarkm *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 5706109998Smarkm lines, recognice more "algorithms" that can be deselected, and make 5707109998Smarkm it complain about algorithm deselection that isn't recognised. 5708109998Smarkm [Richard Levitte] 5709109998Smarkm 5710109998Smarkm *) New ASN1 functions to handle dup, sign, verify, digest, pack and 5711109998Smarkm unpack operations in terms of ASN1_ITEM. Modify existing wrappers 5712109998Smarkm to use new functions. Add NO_ASN1_OLD which can be set to remove 5713109998Smarkm some old style ASN1 functions: this can be used to determine if old 5714109998Smarkm code will still work when these eventually go away. 5715109998Smarkm [Steve Henson] 5716109998Smarkm 5717109998Smarkm *) New extension functions for OCSP structures, these follow the 5718109998Smarkm same conventions as certificates and CRLs. 5719109998Smarkm [Steve Henson] 5720109998Smarkm 5721109998Smarkm *) New function X509V3_add1_i2d(). This automatically encodes and 5722109998Smarkm adds an extension. Its behaviour can be customised with various 5723109998Smarkm flags to append, replace or delete. Various wrappers added for 5724109998Smarkm certifcates and CRLs. 5725109998Smarkm [Steve Henson] 5726109998Smarkm 5727109998Smarkm *) Fix to avoid calling the underlying ASN1 print routine when 5728109998Smarkm an extension cannot be parsed. Correct a typo in the 5729109998Smarkm OCSP_SERVICELOC extension. Tidy up print OCSP format. 5730109998Smarkm [Steve Henson] 5731109998Smarkm 5732109998Smarkm *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 5733109998Smarkm entries for variables. 5734109998Smarkm [Steve Henson] 5735109998Smarkm 5736109998Smarkm *) Add functionality to apps/openssl.c for detecting locking 5737109998Smarkm problems: As the program is single-threaded, all we have 5738109998Smarkm to do is register a locking callback using an array for 5739109998Smarkm storing which locks are currently held by the program. 5740109998Smarkm [Bodo Moeller] 5741109998Smarkm 5742109998Smarkm *) Use a lock around the call to CRYPTO_get_ex_new_index() in 5743109998Smarkm SSL_get_ex_data_X509_STORE_idx(), which is used in 5744109998Smarkm ssl_verify_cert_chain() and thus can be called at any time 5745109998Smarkm during TLS/SSL handshakes so that thread-safety is essential. 5746109998Smarkm Unfortunately, the ex_data design is not at all suited 5747109998Smarkm for multi-threaded use, so it probably should be abolished. 5748109998Smarkm [Bodo Moeller] 5749109998Smarkm 5750109998Smarkm *) Added Broadcom "ubsec" ENGINE to OpenSSL. 5751109998Smarkm [Broadcom, tweaked and integrated by Geoff Thorpe] 5752109998Smarkm 5753109998Smarkm *) Move common extension printing code to new function 5754109998Smarkm X509V3_print_extensions(). Reorganise OCSP print routines and 5755109998Smarkm implement some needed OCSP ASN1 functions. Add OCSP extensions. 5756109998Smarkm [Steve Henson] 5757109998Smarkm 5758109998Smarkm *) New function X509_signature_print() to remove duplication in some 5759109998Smarkm print routines. 5760109998Smarkm [Steve Henson] 5761109998Smarkm 5762109998Smarkm *) Add a special meaning when SET OF and SEQUENCE OF flags are both 5763109998Smarkm set (this was treated exactly the same as SET OF previously). This 5764109998Smarkm is used to reorder the STACK representing the structure to match the 5765109998Smarkm encoding. This will be used to get round a problem where a PKCS7 5766109998Smarkm structure which was signed could not be verified because the STACK 5767109998Smarkm order did not reflect the encoded order. 5768109998Smarkm [Steve Henson] 5769109998Smarkm 5770109998Smarkm *) Reimplement the OCSP ASN1 module using the new code. 5771109998Smarkm [Steve Henson] 5772109998Smarkm 5773109998Smarkm *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 5774109998Smarkm for its ASN1 operations. The old style function pointers still exist 5775109998Smarkm for now but they will eventually go away. 5776109998Smarkm [Steve Henson] 5777109998Smarkm 5778109998Smarkm *) Merge in replacement ASN1 code from the ASN1 branch. This almost 5779109998Smarkm completely replaces the old ASN1 functionality with a table driven 5780109998Smarkm encoder and decoder which interprets an ASN1_ITEM structure describing 5781109998Smarkm the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 5782109998Smarkm largely maintained. Almost all of the old asn1_mac.h macro based ASN1 5783109998Smarkm has also been converted to the new form. 5784109998Smarkm [Steve Henson] 5785109998Smarkm 5786109998Smarkm *) Change BN_mod_exp_recp so that negative moduli are tolerated 5787109998Smarkm (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 5788109998Smarkm so that BN_mod_exp_mont and BN_mod_exp_mont_word work 5789109998Smarkm for negative moduli. 5790109998Smarkm [Bodo Moeller] 5791109998Smarkm 5792109998Smarkm *) Fix BN_uadd and BN_usub: Always return non-negative results instead 5793109998Smarkm of not touching the result's sign bit. 5794109998Smarkm [Bodo Moeller] 5795109998Smarkm 5796109998Smarkm *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 5797109998Smarkm set. 5798109998Smarkm [Bodo Moeller] 5799109998Smarkm 5800109998Smarkm *) Changed the LHASH code to use prototypes for callbacks, and created 5801109998Smarkm macros to declare and implement thin (optionally static) functions 5802109998Smarkm that provide type-safety and avoid function pointer casting for the 5803109998Smarkm type-specific callbacks. 5804109998Smarkm [Geoff Thorpe] 5805109998Smarkm 5806109998Smarkm *) Added Kerberos Cipher Suites to be used with TLS, as written in 5807109998Smarkm RFC 2712. 5808109998Smarkm [Veers Staats <staatsvr@asc.hpc.mil>, 5809109998Smarkm Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 5810109998Smarkm 5811109998Smarkm *) Reformat the FAQ so the different questions and answers can be divided 5812109998Smarkm in sections depending on the subject. 5813109998Smarkm [Richard Levitte] 5814109998Smarkm 5815109998Smarkm *) Have the zlib compression code load ZLIB.DLL dynamically under 5816109998Smarkm Windows. 5817109998Smarkm [Richard Levitte] 5818109998Smarkm 5819109998Smarkm *) New function BN_mod_sqrt for computing square roots modulo a prime 5820109998Smarkm (using the probabilistic Tonelli-Shanks algorithm unless 5821109998Smarkm p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 5822109998Smarkm be handled deterministically). 5823109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5824109998Smarkm 5825109998Smarkm *) Make BN_mod_inverse faster by explicitly handling small quotients 5826109998Smarkm in the Euclid loop. (Speed gain about 20% for small moduli [256 or 5827109998Smarkm 512 bits], about 30% for larger ones [1024 or 2048 bits].) 5828109998Smarkm [Bodo Moeller] 5829109998Smarkm 5830109998Smarkm *) New function BN_kronecker. 5831109998Smarkm [Bodo Moeller] 5832109998Smarkm 5833109998Smarkm *) Fix BN_gcd so that it works on negative inputs; the result is 5834109998Smarkm positive unless both parameters are zero. 5835109998Smarkm Previously something reasonably close to an infinite loop was 5836109998Smarkm possible because numbers could be growing instead of shrinking 5837109998Smarkm in the implementation of Euclid's algorithm. 5838109998Smarkm [Bodo Moeller] 5839109998Smarkm 5840109998Smarkm *) Fix BN_is_word() and BN_is_one() macros to take into account the 5841109998Smarkm sign of the number in question. 5842109998Smarkm 5843109998Smarkm Fix BN_is_word(a,w) to work correctly for w == 0. 5844109998Smarkm 5845109998Smarkm The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 5846109998Smarkm because its test if the absolute value of 'a' equals 'w'. 5847109998Smarkm Note that BN_abs_is_word does *not* handle w == 0 reliably; 5848109998Smarkm it exists mostly for use in the implementations of BN_is_zero(), 5849109998Smarkm BN_is_one(), and BN_is_word(). 5850109998Smarkm [Bodo Moeller] 5851109998Smarkm 5852109998Smarkm *) New function BN_swap. 5853109998Smarkm [Bodo Moeller] 5854109998Smarkm 5855109998Smarkm *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 5856109998Smarkm the exponentiation functions are more likely to produce reasonable 5857109998Smarkm results on negative inputs. 5858109998Smarkm [Bodo Moeller] 5859109998Smarkm 5860109998Smarkm *) Change BN_mod_mul so that the result is always non-negative. 5861109998Smarkm Previously, it could be negative if one of the factors was negative; 5862109998Smarkm I don't think anyone really wanted that behaviour. 5863109998Smarkm [Bodo Moeller] 5864109998Smarkm 5865109998Smarkm *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 5866109998Smarkm (except for exponentiation, which stays in crypto/bn/bn_exp.c, 5867109998Smarkm and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 5868109998Smarkm and add new functions: 5869109998Smarkm 5870109998Smarkm BN_nnmod 5871109998Smarkm BN_mod_sqr 5872109998Smarkm BN_mod_add 5873109998Smarkm BN_mod_add_quick 5874109998Smarkm BN_mod_sub 5875109998Smarkm BN_mod_sub_quick 5876109998Smarkm BN_mod_lshift1 5877109998Smarkm BN_mod_lshift1_quick 5878109998Smarkm BN_mod_lshift 5879109998Smarkm BN_mod_lshift_quick 5880109998Smarkm 5881109998Smarkm These functions always generate non-negative results. 5882109998Smarkm 5883109998Smarkm BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 5884109998Smarkm such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 5885109998Smarkm 5886109998Smarkm BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 5887109998Smarkm BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 5888109998Smarkm be reduced modulo m. 5889109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5890109998Smarkm 5891111147Snectar#if 0 5892111147Snectar The following entry accidentily appeared in the CHANGES file 5893111147Snectar distributed with OpenSSL 0.9.7. The modifications described in 5894111147Snectar it do *not* apply to OpenSSL 0.9.7. 5895111147Snectar 5896109998Smarkm *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 5897109998Smarkm was actually never needed) and in BN_mul(). The removal in BN_mul() 5898109998Smarkm required a small change in bn_mul_part_recursive() and the addition 5899109998Smarkm of the functions bn_cmp_part_words(), bn_sub_part_words() and 5900109998Smarkm bn_add_part_words(), which do the same thing as bn_cmp_words(), 5901109998Smarkm bn_sub_words() and bn_add_words() except they take arrays with 5902109998Smarkm differing sizes. 5903109998Smarkm [Richard Levitte] 5904111147Snectar#endif 5905109998Smarkm 5906109998Smarkm *) In 'openssl passwd', verify passwords read from the terminal 5907109998Smarkm unless the '-salt' option is used (which usually means that 5908109998Smarkm verification would just waste user's time since the resulting 5909109998Smarkm hash is going to be compared with some given password hash) 5910109998Smarkm or the new '-noverify' option is used. 5911109998Smarkm 5912109998Smarkm This is an incompatible change, but it does not affect 5913109998Smarkm non-interactive use of 'openssl passwd' (passwords on the command 5914109998Smarkm line, '-stdin' option, '-in ...' option) and thus should not 5915109998Smarkm cause any problems. 5916109998Smarkm [Bodo Moeller] 5917109998Smarkm 5918109998Smarkm *) Remove all references to RSAref, since there's no more need for it. 5919109998Smarkm [Richard Levitte] 5920109998Smarkm 5921109998Smarkm *) Make DSO load along a path given through an environment variable 5922109998Smarkm (SHLIB_PATH) with shl_load(). 5923109998Smarkm [Richard Levitte] 5924109998Smarkm 5925109998Smarkm *) Constify the ENGINE code as a result of BIGNUM constification. 5926109998Smarkm Also constify the RSA code and most things related to it. In a 5927109998Smarkm few places, most notable in the depth of the ASN.1 code, ugly 5928109998Smarkm casts back to non-const were required (to be solved at a later 5929109998Smarkm time) 5930109998Smarkm [Richard Levitte] 5931109998Smarkm 5932109998Smarkm *) Make it so the openssl application has all engines loaded by default. 5933109998Smarkm [Richard Levitte] 5934109998Smarkm 5935109998Smarkm *) Constify the BIGNUM routines a little more. 5936109998Smarkm [Richard Levitte] 5937109998Smarkm 5938109998Smarkm *) Add the following functions: 5939109998Smarkm 5940109998Smarkm ENGINE_load_cswift() 5941109998Smarkm ENGINE_load_chil() 5942109998Smarkm ENGINE_load_atalla() 5943109998Smarkm ENGINE_load_nuron() 5944109998Smarkm ENGINE_load_builtin_engines() 5945109998Smarkm 5946109998Smarkm That way, an application can itself choose if external engines that 5947109998Smarkm are built-in in OpenSSL shall ever be used or not. The benefit is 5948109998Smarkm that applications won't have to be linked with libdl or other dso 5949109998Smarkm libraries unless it's really needed. 5950109998Smarkm 5951109998Smarkm Changed 'openssl engine' to load all engines on demand. 5952109998Smarkm Changed the engine header files to avoid the duplication of some 5953109998Smarkm declarations (they differed!). 5954109998Smarkm [Richard Levitte] 5955109998Smarkm 5956109998Smarkm *) 'openssl engine' can now list capabilities. 5957109998Smarkm [Richard Levitte] 5958109998Smarkm 5959109998Smarkm *) Better error reporting in 'openssl engine'. 5960109998Smarkm [Richard Levitte] 5961109998Smarkm 5962109998Smarkm *) Never call load_dh_param(NULL) in s_server. 5963109998Smarkm [Bodo Moeller] 5964109998Smarkm 5965109998Smarkm *) Add engine application. It can currently list engines by name and 5966109998Smarkm identity, and test if they are actually available. 5967109998Smarkm [Richard Levitte] 5968109998Smarkm 5969109998Smarkm *) Improve RPM specification file by forcing symbolic linking and making 5970109998Smarkm sure the installed documentation is also owned by root.root. 5971109998Smarkm [Damien Miller <djm@mindrot.org>] 5972109998Smarkm 5973109998Smarkm *) Give the OpenSSL applications more possibilities to make use of 5974109998Smarkm keys (public as well as private) handled by engines. 5975109998Smarkm [Richard Levitte] 5976109998Smarkm 5977109998Smarkm *) Add OCSP code that comes from CertCo. 5978109998Smarkm [Richard Levitte] 5979109998Smarkm 5980109998Smarkm *) Add VMS support for the Rijndael code. 5981109998Smarkm [Richard Levitte] 5982109998Smarkm 5983109998Smarkm *) Added untested support for Nuron crypto accelerator. 5984109998Smarkm [Ben Laurie] 5985109998Smarkm 5986109998Smarkm *) Add support for external cryptographic devices. This code was 5987109998Smarkm previously distributed separately as the "engine" branch. 5988109998Smarkm [Geoff Thorpe, Richard Levitte] 5989109998Smarkm 5990109998Smarkm *) Rework the filename-translation in the DSO code. It is now possible to 5991109998Smarkm have far greater control over how a "name" is turned into a filename 5992109998Smarkm depending on the operating environment and any oddities about the 5993109998Smarkm different shared library filenames on each system. 5994109998Smarkm [Geoff Thorpe] 5995109998Smarkm 5996109998Smarkm *) Support threads on FreeBSD-elf in Configure. 5997109998Smarkm [Richard Levitte] 5998109998Smarkm 5999109998Smarkm *) Fix for SHA1 assembly problem with MASM: it produces 6000109998Smarkm warnings about corrupt line number information when assembling 6001109998Smarkm with debugging information. This is caused by the overlapping 6002109998Smarkm of two sections. 6003109998Smarkm [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 6004109998Smarkm 6005109998Smarkm *) NCONF changes. 6006109998Smarkm NCONF_get_number() has no error checking at all. As a replacement, 6007109998Smarkm NCONF_get_number_e() is defined (_e for "error checking") and is 6008109998Smarkm promoted strongly. The old NCONF_get_number is kept around for 6009109998Smarkm binary backward compatibility. 6010109998Smarkm Make it possible for methods to load from something other than a BIO, 6011109998Smarkm by providing a function pointer that is given a name instead of a BIO. 6012109998Smarkm For example, this could be used to load configuration data from an 6013109998Smarkm LDAP server. 6014109998Smarkm [Richard Levitte] 6015109998Smarkm 6016109998Smarkm *) Fix for non blocking accept BIOs. Added new I/O special reason 6017109998Smarkm BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 6018109998Smarkm with non blocking I/O was not possible because no retry code was 6019109998Smarkm implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 6020109998Smarkm this case. 6021109998Smarkm [Steve Henson] 6022109998Smarkm 6023109998Smarkm *) Added the beginnings of Rijndael support. 6024109998Smarkm [Ben Laurie] 6025109998Smarkm 6026109998Smarkm *) Fix for bug in DirectoryString mask setting. Add support for 6027109998Smarkm X509_NAME_print_ex() in 'req' and X509_print_ex() function 6028109998Smarkm to allow certificate printing to more controllable, additional 6029109998Smarkm 'certopt' option to 'x509' to allow new printing options to be 6030109998Smarkm set. 6031109998Smarkm [Steve Henson] 6032109998Smarkm 6033109998Smarkm *) Clean old EAY MD5 hack from e_os.h. 6034109998Smarkm [Richard Levitte] 6035109998Smarkm 6036142425Snectar Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 6037142425Snectar 6038142425Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 6039160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 6040142425Snectar [Joe Orton, Steve Henson] 6041142425Snectar 6042142425Snectar Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 6043142425Snectar 6044142425Snectar *) Fix additional bug revealed by the NISCC test suite: 6045142425Snectar 6046142425Snectar Stop bug triggering large recursion when presented with 6047160814Ssimon certain ASN.1 tags (CVE-2003-0851) 6048142425Snectar [Steve Henson] 6049142425Snectar 6050120631Snectar Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 6051120631Snectar 6052120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 6053120631Snectar 6054120631Snectar Stop out of bounds reads in the ASN1 code when presented with 6055160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 6056120631Snectar 6057120631Snectar If verify callback ignores invalid public key errors don't try to check 6058120631Snectar certificate signature with the NULL public key. 6059120631Snectar 6060120631Snectar [Steve Henson] 6061120631Snectar 6062120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 6063120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 6064120631Snectar specifications. 6065120631Snectar [Steve Henson] 6066120631Snectar 6067120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 6068120631Snectar extra data after the compression methods not only for TLS 1.0 6069120631Snectar but also for SSL 3.0 (as required by the specification). 6070120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 6071120631Snectar 6072120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 6073120631Snectar when it's 512 *bits* long, not 512 bytes. 6074120631Snectar [Richard Levitte] 6075120631Snectar 6076120631Snectar Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 6077120631Snectar 6078120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 6079120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 6080120631Snectar a protocol version number mismatch like a decryption error 6081120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 6082120631Snectar [Bodo Moeller] 6083120631Snectar 6084120631Snectar *) Turn on RSA blinding by default in the default implementation 6085120631Snectar to avoid a timing attack. Applications that don't want it can call 6086120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 6087120631Snectar They would be ill-advised to do so in most cases. 6088120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 6089120631Snectar 6090120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 6091120631Snectar seeded (in this case, the secret RSA exponent is abused as 6092120631Snectar an unpredictable seed -- if it is not unpredictable, there 6093120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 6094120631Snectar by remembering the creator's thread ID in rsa->blinding and 6095120631Snectar having all other threads use local one-time blinding factors 6096120631Snectar (this requires more computation than sharing rsa->blinding, but 6097120631Snectar avoids excessive locking; and if an RSA object is not shared 6098120631Snectar between threads, blinding will still be very fast). 6099120631Snectar [Bodo Moeller] 6100120631Snectar 6101111147Snectar Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 6102111147Snectar 6103111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 6104111147Snectar via timing by performing a MAC computation even if incorrrect 6105111147Snectar block cipher padding has been found. This is a countermeasure 6106111147Snectar against active attacks where the attacker has to distinguish 6107160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 6108111147Snectar 6109111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 6110111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 6111111147Snectar Martin Vuagnoux (EPFL, Ilion)] 6112111147Snectar 6113109998Smarkm Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 6114109998Smarkm 6115109998Smarkm *) New function OPENSSL_cleanse(), which is used to cleanse a section of 6116109998Smarkm memory from it's contents. This is done with a counter that will 6117109998Smarkm place alternating values in each byte. This can be used to solve 6118109998Smarkm two issues: 1) the removal of calls to memset() by highly optimizing 6119109998Smarkm compilers, and 2) cleansing with other values than 0, since those can 6120109998Smarkm be read through on certain media, for example a swap space on disk. 6121109998Smarkm [Geoff Thorpe] 6122109998Smarkm 6123109998Smarkm *) Bugfix: client side session caching did not work with external caching, 6124109998Smarkm because the session->cipher setting was not restored when reloading 6125109998Smarkm from the external cache. This problem was masked, when 6126109998Smarkm SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 6127109998Smarkm (Found by Steve Haslam <steve@araqnid.ddts.net>.) 6128109998Smarkm [Lutz Jaenicke] 6129109998Smarkm 6130109998Smarkm *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 6131109998Smarkm length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 6132109998Smarkm [Zeev Lieber <zeev-l@yahoo.com>] 6133109998Smarkm 6134109998Smarkm *) Undo an undocumented change introduced in 0.9.6e which caused 6135109998Smarkm repeated calls to OpenSSL_add_all_ciphers() and 6136109998Smarkm OpenSSL_add_all_digests() to be ignored, even after calling 6137109998Smarkm EVP_cleanup(). 6138109998Smarkm [Richard Levitte] 6139109998Smarkm 6140109998Smarkm *) Change the default configuration reader to deal with last line not 6141109998Smarkm being properly terminated. 6142109998Smarkm [Richard Levitte] 6143109998Smarkm 6144109998Smarkm *) Change X509_NAME_cmp() so it applies the special rules on handling 6145109998Smarkm DN values that are of type PrintableString, as well as RDNs of type 6146109998Smarkm emailAddress where the value has the type ia5String. 6147109998Smarkm [stefank@valicert.com via Richard Levitte] 6148109998Smarkm 6149109998Smarkm *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 6150109998Smarkm the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 6151109998Smarkm doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 6152109998Smarkm the bitwise-OR of the two for use by the majority of applications 6153109998Smarkm wanting this behaviour, and update the docs. The documented 6154109998Smarkm behaviour and actual behaviour were inconsistent and had been 6155109998Smarkm changing anyway, so this is more a bug-fix than a behavioural 6156109998Smarkm change. 6157109998Smarkm [Geoff Thorpe, diagnosed by Nadav Har'El] 6158109998Smarkm 6159109998Smarkm *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 6160109998Smarkm (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 6161109998Smarkm [Bodo Moeller] 6162109998Smarkm 6163109998Smarkm *) Fix initialization code race conditions in 6164109998Smarkm SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 6165109998Smarkm SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 6166109998Smarkm SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 6167109998Smarkm TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 6168109998Smarkm ssl2_get_cipher_by_char(), 6169109998Smarkm ssl3_get_cipher_by_char(). 6170109998Smarkm [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 6171109998Smarkm 6172109998Smarkm *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 6173109998Smarkm the cached sessions are flushed, as the remove_cb() might use ex_data 6174109998Smarkm contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 6175109998Smarkm (see [openssl.org #212]). 6176109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 6177109998Smarkm 6178109998Smarkm *) Fix typo in OBJ_txt2obj which incorrectly passed the content 6179109998Smarkm length, instead of the encoding length to d2i_ASN1_OBJECT. 6180109998Smarkm [Steve Henson] 6181109998Smarkm 6182101618Snectar Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 6183101618Snectar 6184101618Snectar *) [In 0.9.6g-engine release:] 6185101618Snectar Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 6186101618Snectar [Lynn Gazis <lgazis@rainbow.com>] 6187101618Snectar 6188101613Snectar Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 6189101613Snectar 6190101613Snectar *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 6191101613Snectar and get fix the header length calculation. 6192101613Snectar [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 6193101613Snectar Alon Kantor <alonk@checkpoint.com> (and others), 6194101613Snectar Steve Henson] 6195101613Snectar 6196101613Snectar *) Use proper error handling instead of 'assertions' in buffer 6197101613Snectar overflow checks added in 0.9.6e. This prevents DoS (the 6198101613Snectar assertions could call abort()). 6199101613Snectar [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 6200101613Snectar 6201100936Snectar Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 6202100936Snectar 6203109998Smarkm *) Add various sanity checks to asn1_get_length() to reject 6204109998Smarkm the ASN1 length bytes if they exceed sizeof(long), will appear 6205109998Smarkm negative or the content length exceeds the length of the 6206109998Smarkm supplied buffer. 6207109998Smarkm [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6208109998Smarkm 6209100936Snectar *) Fix cipher selection routines: ciphers without encryption had no flags 6210100936Snectar for the cipher strength set and where therefore not handled correctly 6211100936Snectar by the selection routines (PR #130). 6212100936Snectar [Lutz Jaenicke] 6213100936Snectar 6214100936Snectar *) Fix EVP_dsa_sha macro. 6215100936Snectar [Nils Larsch] 6216100936Snectar 6217100936Snectar *) New option 6218100936Snectar SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 6219100936Snectar for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 6220100936Snectar that was added in OpenSSL 0.9.6d. 6221100936Snectar 6222100936Snectar As the countermeasure turned out to be incompatible with some 6223100936Snectar broken SSL implementations, the new option is part of SSL_OP_ALL. 6224100936Snectar SSL_OP_ALL is usually employed when compatibility with weird SSL 6225100936Snectar implementations is desired (e.g. '-bugs' option to 's_client' and 6226100936Snectar 's_server'), so the new option is automatically set in many 6227100936Snectar applications. 6228100936Snectar [Bodo Moeller] 6229100936Snectar 6230100936Snectar *) Changes in security patch: 6231100936Snectar 6232100936Snectar Changes marked "(CHATS)" were sponsored by the Defense Advanced 6233100936Snectar Research Projects Agency (DARPA) and Air Force Research Laboratory, 6234100936Snectar Air Force Materiel Command, USAF, under agreement number 6235100936Snectar F30602-01-2-0537. 6236100936Snectar 6237100936Snectar *) Add various sanity checks to asn1_get_length() to reject 6238100936Snectar the ASN1 length bytes if they exceed sizeof(long), will appear 6239100936Snectar negative or the content length exceeds the length of the 6240160814Ssimon supplied buffer. (CVE-2002-0659) 6241100936Snectar [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6242100936Snectar 6243100936Snectar *) Assertions for various potential buffer overflows, not known to 6244100936Snectar happen in practice. 6245100936Snectar [Ben Laurie (CHATS)] 6246100936Snectar 6247100936Snectar *) Various temporary buffers to hold ASCII versions of integers were 6248160814Ssimon too small for 64 bit platforms. (CVE-2002-0655) 6249100936Snectar [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 6250100936Snectar 6251100936Snectar *) Remote buffer overflow in SSL3 protocol - an attacker could 6252160814Ssimon supply an oversized session ID to a client. (CVE-2002-0656) 6253100936Snectar [Ben Laurie (CHATS)] 6254100936Snectar 6255100936Snectar *) Remote buffer overflow in SSL2 protocol - an attacker could 6256160814Ssimon supply an oversized client master key. (CVE-2002-0656) 6257100936Snectar [Ben Laurie (CHATS)] 6258100936Snectar 6259100928Snectar Changes between 0.9.6c and 0.9.6d [9 May 2002] 6260100928Snectar 6261100928Snectar *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 6262100928Snectar encoded as NULL) with id-dsa-with-sha1. 6263100928Snectar [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 6264100928Snectar 6265100928Snectar *) Check various X509_...() return values in apps/req.c. 6266100928Snectar [Nils Larsch <nla@trustcenter.de>] 6267100928Snectar 6268100928Snectar *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 6269100928Snectar an end-of-file condition would erronously be flagged, when the CRLF 6270100928Snectar was just at the end of a processed block. The bug was discovered when 6271100928Snectar processing data through a buffering memory BIO handing the data to a 6272100928Snectar BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 6273100928Snectar <ptsekov@syntrex.com> and Nedelcho Stanev. 6274100928Snectar [Lutz Jaenicke] 6275100928Snectar 6276100928Snectar *) Implement a countermeasure against a vulnerability recently found 6277100928Snectar in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 6278100928Snectar before application data chunks to avoid the use of known IVs 6279100928Snectar with data potentially chosen by the attacker. 6280100928Snectar [Bodo Moeller] 6281100928Snectar 6282100928Snectar *) Fix length checks in ssl3_get_client_hello(). 6283100928Snectar [Bodo Moeller] 6284100928Snectar 6285100928Snectar *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 6286100928Snectar to prevent ssl3_read_internal() from incorrectly assuming that 6287100928Snectar ssl3_read_bytes() found application data while handshake 6288100928Snectar processing was enabled when in fact s->s3->in_read_app_data was 6289100928Snectar merely automatically cleared during the initial handshake. 6290100928Snectar [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 6291100928Snectar 6292100928Snectar *) Fix object definitions for Private and Enterprise: they were not 6293100928Snectar recognized in their shortname (=lowercase) representation. Extend 6294100928Snectar obj_dat.pl to issue an error when using undefined keywords instead 6295100928Snectar of silently ignoring the problem (Svenning Sorensen 6296100928Snectar <sss@sss.dnsalias.net>). 6297100928Snectar [Lutz Jaenicke] 6298100928Snectar 6299100928Snectar *) Fix DH_generate_parameters() so that it works for 'non-standard' 6300100928Snectar generators, i.e. generators other than 2 and 5. (Previously, the 6301100928Snectar code did not properly initialise the 'add' and 'rem' values to 6302100928Snectar BN_generate_prime().) 6303100928Snectar 6304100928Snectar In the new general case, we do not insist that 'generator' is 6305100928Snectar actually a primitive root: This requirement is rather pointless; 6306100928Snectar a generator of the order-q subgroup is just as good, if not 6307100928Snectar better. 6308100928Snectar [Bodo Moeller] 6309100928Snectar 6310100928Snectar *) Map new X509 verification errors to alerts. Discovered and submitted by 6311100928Snectar Tom Wu <tom@arcot.com>. 6312100928Snectar [Lutz Jaenicke] 6313100928Snectar 6314100928Snectar *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 6315100928Snectar returning non-zero before the data has been completely received 6316100928Snectar when using non-blocking I/O. 6317100928Snectar [Bodo Moeller; problem pointed out by John Hughes] 6318100928Snectar 6319100928Snectar *) Some of the ciphers missed the strength entry (SSL_LOW etc). 6320100928Snectar [Ben Laurie, Lutz Jaenicke] 6321100928Snectar 6322100928Snectar *) Fix bug in SSL_clear(): bad sessions were not removed (found by 6323100928Snectar Yoram Zahavi <YoramZ@gilian.com>). 6324100928Snectar [Lutz Jaenicke] 6325100928Snectar 6326100928Snectar *) Add information about CygWin 1.3 and on, and preserve proper 6327100928Snectar configuration for the versions before that. 6328100928Snectar [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 6329100928Snectar 6330100928Snectar *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 6331100928Snectar check whether we deal with a copy of a session and do not delete from 6332100928Snectar the cache in this case. Problem reported by "Izhar Shoshani Levi" 6333100928Snectar <izhar@checkpoint.com>. 6334100928Snectar [Lutz Jaenicke] 6335100928Snectar 6336100928Snectar *) Do not store session data into the internal session cache, if it 6337100928Snectar is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 6338100928Snectar flag is set). Proposed by Aslam <aslam@funk.com>. 6339100928Snectar [Lutz Jaenicke] 6340100928Snectar 6341100928Snectar *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 6342100928Snectar value is 0. 6343100928Snectar [Richard Levitte] 6344100928Snectar 6345109998Smarkm *) [In 0.9.6d-engine release:] 6346109998Smarkm Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 6347100928Snectar [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 6348100928Snectar 6349100928Snectar *) Add the configuration target linux-s390x. 6350100928Snectar [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 6351100928Snectar 6352100928Snectar *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 6353100928Snectar ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 6354100928Snectar variable as an indication that a ClientHello message has been 6355100928Snectar received. As the flag value will be lost between multiple 6356100928Snectar invocations of ssl3_accept when using non-blocking I/O, the 6357100928Snectar function may not be aware that a handshake has actually taken 6358100928Snectar place, thus preventing a new session from being added to the 6359100928Snectar session cache. 6360100928Snectar 6361100928Snectar To avoid this problem, we now set s->new_session to 2 instead of 6362100928Snectar using a local variable. 6363100928Snectar [Lutz Jaenicke, Bodo Moeller] 6364100928Snectar 6365100928Snectar *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 6366100928Snectar if the SSL_R_LENGTH_MISMATCH error is detected. 6367100928Snectar [Geoff Thorpe, Bodo Moeller] 6368100928Snectar 6369100928Snectar *) New 'shared_ldflag' column in Configure platform table. 6370100928Snectar [Richard Levitte] 6371100928Snectar 6372100928Snectar *) Fix EVP_CIPHER_mode macro. 6373100928Snectar ["Dan S. Camper" <dan@bti.net>] 6374100928Snectar 6375100928Snectar *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 6376100928Snectar type, we must throw them away by setting rr->length to 0. 6377100928Snectar [D P Chang <dpc@qualys.com>] 6378100928Snectar 637989837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 638089837Skris 638189837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 638289837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 638389837Skris worked incorrectly for those cases where range = 10..._2 and 638489837Skris 3*range is two bits longer than range.) 638589837Skris [Bodo Moeller] 638689837Skris 638789837Skris *) Only add signing time to PKCS7 structures if it is not already 638889837Skris present. 638989837Skris [Steve Henson] 639089837Skris 639189837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 639289837Skris OBJ_ld_ce should be OBJ_id_ce. 639389837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 639489837Skris incorrect (cf. RFC 3039). 639589837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 639689837Skris 639789837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 639889837Skris returns early because it has nothing to do. 639989837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 640089837Skris 640189837Skris *) [In 0.9.6c-engine release:] 640289837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 640389837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 640489837Skris 640589837Skris *) [In 0.9.6c-engine release:] 640689837Skris Add support for Cryptographic Appliance's keyserver technology. 640789837Skris (Use engine 'keyclient') 640889837Skris [Cryptographic Appliances and Geoff Thorpe] 640989837Skris 641089837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 641189837Skris is called via tools/c89.sh because arguments have to be 641289837Skris rearranged (all '-L' options must appear before the first object 641389837Skris modules). 641489837Skris [Richard Shapiro <rshapiro@abinitio.com>] 641589837Skris 641689837Skris *) [In 0.9.6c-engine release:] 641789837Skris Add support for Broadcom crypto accelerator cards, backported 641889837Skris from 0.9.7. 641989837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 642089837Skris 642189837Skris *) [In 0.9.6c-engine release:] 642289837Skris Add support for SureWare crypto accelerator cards from 642389837Skris Baltimore Technologies. (Use engine 'sureware') 642489837Skris [Baltimore Technologies and Mark Cox] 642589837Skris 642689837Skris *) [In 0.9.6c-engine release:] 642789837Skris Add support for crypto accelerator cards from Accelerated 642889837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 642989837Skris [AEP Inc. and Mark Cox] 643089837Skris 643189837Skris *) Add a configuration entry for gcc on UnixWare. 643289837Skris [Gary Benson <gbenson@redhat.com>] 643389837Skris 643489837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 643589837Skris messages are stored in a single piece (fixed-length part and 643689837Skris variable-length part combined) and fix various bugs found on the way. 643789837Skris [Bodo Moeller] 643889837Skris 643989837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 644089837Skris instead. BIO_gethostbyname() does not know what timeouts are 644189837Skris appropriate, so entries would stay in cache even when they have 644289837Skris become invalid. 644389837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 644489837Skris 644589837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 644689837Skris faced with a pathologically small ClientHello fragment that does 644789837Skris not contain client_version: Instead of aborting with an error, 644889837Skris simply choose the highest available protocol version (i.e., 644989837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 645089837Skris messages are never sent like this, but this change gives us 645189837Skris strictly correct behaviour at least for TLS. 645289837Skris [Bodo Moeller] 645389837Skris 645489837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 645589837Skris never resets s->method to s->ctx->method when called from within 645689837Skris one of the SSL handshake functions. 645789837Skris [Bodo Moeller; problem pointed out by Niko Baric] 645889837Skris 645989837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 646089837Skris (sent using the client's version number) if client_version is 646189837Skris smaller than the protocol version in use. Also change 646289837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 646389837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 646489837Skris the client will at least see that alert. 646589837Skris [Bodo Moeller] 646689837Skris 646789837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 646889837Skris correctly. 646989837Skris [Bodo Moeller] 647089837Skris 647189837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 647289837Skris client receives HelloRequest while in a handshake. 647389837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 647489837Skris 647589837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 647689837Skris should end in 'break', not 'goto end' which circuments various 647789837Skris cleanups done in state SSL_ST_OK. But session related stuff 647889837Skris must be disabled for SSL_ST_OK in the case that we just sent a 647989837Skris HelloRequest. 648089837Skris 648189837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 648289837Skris before just sending a HelloRequest. 648389837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 648489837Skris 648589837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 648689837Skris reveal whether illegal block cipher padding was found or a MAC 648789837Skris verification error occured. (Neither SSLerr() codes nor alerts 648889837Skris are directly visible to potential attackers, but the information 648989837Skris may leak via logfiles.) 649089837Skris 649189837Skris Similar changes are not required for the SSL 2.0 implementation 649289837Skris because the number of padding bytes is sent in clear for SSL 2.0, 649389837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 649489837Skris failed to verify that the purported number of padding bytes is in 649589837Skris the legal range. 649689837Skris [Bodo Moeller] 649789837Skris 649889837Skris *) Add OpenUNIX-8 support including shared libraries 649989837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 650089837Skris [Lutz Jaenicke] 650189837Skris 650289837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 650389837Skris 'wristwatch attack' using huge encoding parameters (cf. 650489837Skris James H. Manger's CRYPTO 2001 paper). Note that the 650589837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 650689837Skris encoding parameters and hence was not vulnerable. 650789837Skris [Bodo Moeller] 650889837Skris 650989837Skris *) BN_sqr() bug fix. 651089837Skris [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 651189837Skris 651289837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 651389837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 651489837Skris followed by modular reduction. 651589837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 651689837Skris 651789837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 651889837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 651989837Skris [Bodo Moeller] 652089837Skris 652189837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 652289837Skris This function was broken, as the check for a new client hello message 652389837Skris to handle SGC did not allow these large messages. 652489837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 652589837Skris [Lutz Jaenicke] 652689837Skris 652789837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 652889837Skris [Lutz Jaenicke] 652989837Skris 653089837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 653189837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 653289837Skris [Lutz Jaenicke] 653389837Skris 653489837Skris *) Rework the configuration and shared library support for Tru64 Unix. 653589837Skris The configuration part makes use of modern compiler features and 653689837Skris still retains old compiler behavior for those that run older versions 653789837Skris of the OS. The shared library support part includes a variant that 653889837Skris uses the RPATH feature, and is available through the special 653989837Skris configuration target "alpha-cc-rpath", which will never be selected 654089837Skris automatically. 654189837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 654289837Skris 654389837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 654489837Skris with the same message size as in ssl3_get_certificate_request(). 654589837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 654689837Skris messages might inadvertently be reject as too long. 654789837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 654889837Skris 654989837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 655089837Skris [Andy Polyakov] 655189837Skris 655289837Skris *) Modified SSL library such that the verify_callback that has been set 655389837Skris specificly for an SSL object with SSL_set_verify() is actually being 655489837Skris used. Before the change, a verify_callback set with this function was 655589837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 655689837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 655789837Skris to allow the necessary settings. 655889837Skris [Lutz Jaenicke] 655989837Skris 656089837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 656189837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 656289837Skris done automatically (in contradiction to the requirements of the C 656389837Skris standard). This made problems when used from OpenSSH. 656489837Skris [Lutz Jaenicke] 656589837Skris 656689837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 656789837Skris dh->length and always used 656889837Skris 656989837Skris BN_rand_range(priv_key, dh->p). 657089837Skris 657189837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 657289837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 657389837Skris dh->length (recommended exponent length) is much smaller than the 657489837Skris length of dh->p. We could use BN_rand_range() if the order of 657589837Skris the subgroup was stored in the DH structure, but we only have 657689837Skris dh->length. 657789837Skris 657889837Skris So switch back to 657989837Skris 658089837Skris BN_rand(priv_key, l, ...) 658189837Skris 658289837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 658389837Skris otherwise. 658489837Skris [Bodo Moeller] 658589837Skris 658689837Skris *) In 658789837Skris 658889837Skris RSA_eay_public_encrypt 658989837Skris RSA_eay_private_decrypt 659089837Skris RSA_eay_private_encrypt (signing) 659189837Skris RSA_eay_public_decrypt (signature verification) 659289837Skris 659389837Skris (default implementations for RSA_public_encrypt, 659489837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 659589837Skris always reject numbers >= n. 659689837Skris [Bodo Moeller] 659789837Skris 659889837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 659989837Skris to synchronize access to 'locking_thread'. This is necessary on 660089837Skris systems where access to 'locking_thread' (an 'unsigned long' 660189837Skris variable) is not atomic. 660289837Skris [Bodo Moeller] 660389837Skris 660489837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 660589837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 660689837Skris a race condition if 0 is a valid thread ID. 660789837Skris [Travis Vitek <vitek@roguewave.com>] 660889837Skris 660989837Skris *) Add support for shared libraries under Irix. 661089837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 661189837Skris 661289837Skris *) Add configuration option to build on Linux on both big-endian and 661389837Skris little-endian MIPS. 661489837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 661589837Skris 661689837Skris *) Add the possibility to create shared libraries on HP-UX. 661789837Skris [Richard Levitte] 661889837Skris 661979998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 662079998Skris 662179998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 662279998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 662379998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 662479998Skris PRNG state recovery was possible based on the output of 662579998Skris one PRNG request appropriately sized to gain knowledge on 662679998Skris 'md' followed by enough consecutive 1-byte PRNG requests 662779998Skris to traverse all of 'state'. 662879998Skris 662979998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 663079998Skris during PRNG output generation, hash all of the previous 663179998Skris 'md_local' value, not just the half used for PRNG output. 663279998Skris 663379998Skris 2. Make the number of bytes from 'state' included into the hash 663479998Skris independent from the number of PRNG bytes requested. 663579998Skris 663679998Skris The first measure alone would be sufficient to avoid 663779998Skris Markku-Juhani's attack. (Actually it had never occurred 663879998Skris to me that the half of 'md_local' used for chaining was the 663979998Skris half from which PRNG output bytes were taken -- I had always 664079998Skris assumed that the secret half would be used.) The second 664179998Skris measure makes sure that additional data from 'state' is never 664279998Skris mixed into 'md_local' in small portions; this heuristically 664379998Skris further strengthens the PRNG. 664479998Skris [Bodo Moeller] 664579998Skris 664679998Skris *) Fix crypto/bn/asm/mips3.s. 664779998Skris [Andy Polyakov] 664879998Skris 664979998Skris *) When only the key is given to "enc", the IV is undefined. Print out 665079998Skris an error message in this case. 665179998Skris [Lutz Jaenicke] 665279998Skris 665379998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 665479998Skris [Steve Henson] 665579998Skris 665679998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 665779998Skris positive and less than q. 665879998Skris [Bodo Moeller] 665979998Skris 666079998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 666179998Skris used: it isn't thread safe and the add_lock_callback should handle 666279998Skris that itself. 666379998Skris [Paul Rose <Paul.Rose@bridge.com>] 666479998Skris 666579998Skris *) Verify that incoming data obeys the block size in 666679998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 666779998Skris [Bodo Moeller] 666879998Skris 666979998Skris *) Fix OAEP check. 667079998Skris [Ulf M�ller, Bodo M�ller] 667179998Skris 667279998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 667389837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 667479998Skris when fixing the server behaviour for backwards-compatible 'client 667579998Skris hello' messages. (Note that the attack is impractical against 667679998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 667779998Skris means that the probability of guessing a valid ciphertext is 667879998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 667979998Skris paper.) 668079998Skris 668179998Skris Before 0.9.5, the countermeasure (hide the error by generating a 668279998Skris random 'decryption result') did not work properly because 668379998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 668479998Skris detect the supposedly ignored error. 668579998Skris 668679998Skris Both problems are now fixed. 668779998Skris [Bodo Moeller] 668879998Skris 668979998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 669079998Skris (previously it was 1024). 669179998Skris [Bodo Moeller] 669279998Skris 669379998Skris *) Fix for compatibility mode trust settings: ignore trust settings 669479998Skris unless some valid trust or reject settings are present. 669579998Skris [Steve Henson] 669679998Skris 669779998Skris *) Fix for blowfish EVP: its a variable length cipher. 669879998Skris [Steve Henson] 669979998Skris 670079998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 670179998Skris parameters in DSA public key structures and return an error in the 670279998Skris DSA routines if parameters are absent. 670379998Skris [Steve Henson] 670479998Skris 670579998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 670679998Skris in the current directory if neither $RANDFILE nor $HOME was set. 670779998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 670879998Skris caused some confusion to Windows users who haven't defined $HOME. 670979998Skris Thus RAND_file_name() is changed again: e_os.h can define a 671079998Skris DEFAULT_HOME, which will be used if $HOME is not set. 671179998Skris For Windows, we use "C:"; on other platforms, we still require 671279998Skris environment variables. 671379998Skris 671479998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 671579998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 671679998Skris having multiple threads call RAND_poll() concurrently. 671779998Skris [Bodo Moeller] 671879998Skris 671979998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 672079998Skris combination of a flag and a thread ID variable. 672179998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 672279998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 672389837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 672479998Skris that they do not hold after the first thread unsets add_do_not_lock). 672579998Skris [Bodo Moeller] 672679998Skris 672779998Skris *) Change bctest again: '-x' expressions are not available in all 672879998Skris versions of 'test'. 672979998Skris [Bodo Moeller] 673079998Skris 673176866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 673272613Skris 673376866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 673476866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 673576866Skris 673676866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 673776866Skris the default extension for executables, if any. Also, make the perl 673876866Skris scripts that use symlink() to test if it really exists and use "cp" 673976866Skris if it doesn't. All this made OpenSSL compilable and installable in 674076866Skris CygWin. 674176866Skris [Richard Levitte] 674276866Skris 674376866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 674476866Skris If SEQUENCE is length is indefinite just set c->slen to the total 674576866Skris amount of data available. 674676866Skris [Steve Henson, reported by shige@FreeBSD.org] 674776866Skris [This change does not apply to 0.9.7.] 674876866Skris 674976866Skris *) Change bctest to avoid here-documents inside command substitution 675076866Skris (workaround for FreeBSD /bin/sh bug). 675176866Skris For compatibility with Ultrix, avoid shell functions (introduced 675276866Skris in the bctest version that searches along $PATH). 675376866Skris [Bodo Moeller] 675476866Skris 675576866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 675676866Skris with des_encrypt() defined on some operating systems, like Solaris 675776866Skris and UnixWare. 675876866Skris [Richard Levitte] 675976866Skris 676076866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 676176866Skris On the Importance of Eliminating Errors in Cryptographic 676276866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 676376866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 676476866Skris [Ulf Moeller] 676576866Skris 676676866Skris *) MIPS assembler BIGNUM division bug fix. 676776866Skris [Andy Polyakov] 676876866Skris 676976866Skris *) Disabled incorrect Alpha assembler code. 677076866Skris [Richard Levitte] 677176866Skris 677276866Skris *) Fix PKCS#7 decode routines so they correctly update the length 677376866Skris after reading an EOC for the EXPLICIT tag. 677476866Skris [Steve Henson] 677576866Skris [This change does not apply to 0.9.7.] 677676866Skris 677776866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 677876866Skris if a 3DES key was generated with a 0 initial byte. Include 677976866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 678076866Skris (but broken) behaviour. 678176866Skris [Steve Henson] 678276866Skris 678376866Skris *) Enhance bctest to search for a working bc along $PATH and print 678476866Skris it when found. 678576866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 678676866Skris 678776866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 678876866Skris don't write to the wrong index in ERR_set_error_data. 678976866Skris [Bodo Moeller] 679076866Skris 679176866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 679276866Skris did not exist. 679376866Skris [Bodo Moeller] 679476866Skris 679576866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 679676866Skris [Jeremy Cooper <jeremy@baymoo.org>] 679776866Skris 679876866Skris *) Make it possible to reuse SSLv2 sessions. 679976866Skris [Richard Levitte] 680076866Skris 680176866Skris *) In copy_email() check for >= 0 as a return value for 680276866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 680376866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 680476866Skris 680576866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 680676866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 680776866Skris PKCS7_verify() fails with non detached data. 680876866Skris [Steve Henson] 680976866Skris 681076866Skris *) Don't use getenv in library functions when run as setuid/setgid. 681176866Skris New function OPENSSL_issetugid(). 681276866Skris [Ulf Moeller] 681376866Skris 681476866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 681576866Skris due to incorrect handling of multi-threading: 681676866Skris 681776866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 681876866Skris 681976866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 682076866Skris 682176866Skris 3. Count how many times MemCheck_off() has been called so that 682276866Skris nested use can be treated correctly. This also avoids 682376866Skris inband-signalling in the previous code (which relied on the 682476866Skris assumption that thread ID 0 is impossible). 682576866Skris [Bodo Moeller] 682676866Skris 682776866Skris *) Add "-rand" option also to s_client and s_server. 682876866Skris [Lutz Jaenicke] 682976866Skris 683076866Skris *) Fix CPU detection on Irix 6.x. 683176866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 683276866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 683376866Skris 683476866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 683576866Skris was empty. 683676866Skris [Steve Henson] 683776866Skris [This change does not apply to 0.9.7.] 683876866Skris 683976866Skris *) Use the cached encoding of an X509_NAME structure rather than 684076866Skris copying it. This is apparently the reason for the libsafe "errors" 684176866Skris but the code is actually correct. 684276866Skris [Steve Henson] 684376866Skris 684472613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 684572613Skris Bleichenbacher's DSA attack. 684676866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 684776866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 684876866Skris and leaves the highest bit random. 684976866Skris [Ulf Moeller, Bodo Moeller] 685072613Skris 685172613Skris *) In the NCONF_...-based implementations for CONF_... queries 685272613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 685372613Skris a temporary CONF structure with the data component set to NULL 685472613Skris (which gives segmentation faults in lh_retrieve). 685572613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 685672613Skris CONF_get_number (which may use environment variables) and directly 685772613Skris return NULL from CONF_get_section. 685872613Skris [Bodo Moeller] 685972613Skris 686072613Skris *) Fix potential buffer overrun for EBCDIC. 686172613Skris [Ulf Moeller] 686272613Skris 686372613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 686472613Skris keyUsage if basicConstraints absent for a CA. 686572613Skris [Steve Henson] 686672613Skris 686772613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 686872613Skris is more generally accepted (no spaces before the semicolon), since 686972613Skris some programs can't parse those values properly otherwise. Also make 687072613Skris sure BIO's that break lines after each write do not create invalid 687172613Skris headers. 687272613Skris [Richard Levitte] 687372613Skris 687472613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 687572613Skris macros previously used would not encode an empty SEQUENCE OF 687672613Skris and break the signature. 687772613Skris [Steve Henson] 687876866Skris [This change does not apply to 0.9.7.] 687972613Skris 688072613Skris *) Zero the premaster secret after deriving the master secret in 688172613Skris DH ciphersuites. 688272613Skris [Steve Henson] 688372613Skris 688472613Skris *) Add some EVP_add_digest_alias registrations (as found in 688572613Skris OpenSSL_add_all_digests()) to SSL_library_init() 688672613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 688772613Skris compatibility with peers using X.509 certificates 688872613Skris with unconventional AlgorithmIdentifier OIDs. 688972613Skris [Bodo Moeller] 689072613Skris 689172613Skris *) Fix for Irix with NO_ASM. 689272613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 689372613Skris 689472613Skris *) ./config script fixes. 689572613Skris [Ulf Moeller, Richard Levitte] 689672613Skris 689772613Skris *) Fix 'openssl passwd -1'. 689872613Skris [Bodo Moeller] 689972613Skris 690072613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 690172613Skris terminated strings whose length is passed in the passlen 690272613Skris parameter, for example from PEM callbacks. This was done 690372613Skris by adding an extra length parameter to asc2uni(). 690472613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 690572613Skris 690672613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 690772613Skris call failed, free the DSA structure. 690872613Skris [Bodo Moeller] 690972613Skris 691072613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 691172613Skris These are present in some PKCS#12 files. 691272613Skris [Steve Henson] 691372613Skris 691472613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 691572613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 691672613Skris when writing a 32767 byte record. 691772613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 691872613Skris 691972613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 692072613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 692172613Skris 692272613Skris (RSA objects have a reference count access to which is protected 692372613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 692472613Skris so they are meant to be shared between threads.) 692572613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 692672613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 692772613Skris 692872613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 692972613Skris [Bodo Moeller] 693072613Skris 693176866Skris *) Use better test patterns in bntest. 693276866Skris [Ulf M�ller] 693376866Skris 693472613Skris *) rand_win.c fix for Borland C. 693572613Skris [Ulf M�ller] 693672613Skris 693772613Skris *) BN_rshift bugfix for n == 0. 693872613Skris [Bodo Moeller] 693972613Skris 694076866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 694176866Skris so that 'make test' does not abort just because 'bc' is broken. 694276866Skris [Bodo Moeller] 694376866Skris 694472613Skris *) Store verify_result within SSL_SESSION also for client side to 694572613Skris avoid potential security hole. (Re-used sessions on the client side 694672613Skris always resulted in verify_result==X509_V_OK, not using the original 694772613Skris result of the server certificate verification.) 694872613Skris [Lutz Jaenicke] 694972613Skris 695072613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 695172613Skris SSL3_RT_APPLICATION_DATA, return 0. 695272613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 695372613Skris [Bodo Moeller] 695472613Skris 695572613Skris *) Fix SSL_peek: 695672613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 695772613Skris releases, have been re-implemented by renaming the previous 695872613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 695972613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 696072613Skris to them. The new ssl[23]_{read,peek} functions are calls to 696172613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 696272613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 696372613Skris does the actual work for ssl3_read_internal. 696472613Skris [Bodo Moeller] 696572613Skris 696676866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 696776866Skris the method-specific "init()" handler. Also clean up ex_data after 696876866Skris calling the method-specific "finish()" handler. Previously, this was 696976866Skris happening the other way round. 697076866Skris [Geoff Thorpe] 697176866Skris 697272613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 697372613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 697472613Skris [Bodo Moeller] 697572613Skris 697676866Skris *) Make sure that shared libraries get the internal name engine with 697776866Skris the full version number and not just 0. This should mark the 697876866Skris shared libraries as not backward compatible. Of course, this should 697976866Skris be changed again when we can guarantee backward binary compatibility. 698076866Skris [Richard Levitte] 698176866Skris 698272613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 698372613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 698472613Skris 698576866Skris *) Rework the system to generate shared libraries: 698676866Skris 698776866Skris - Make note of the expected extension for the shared libraries and 698876866Skris if there is a need for symbolic links from for example libcrypto.so.0 698976866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 699076866Skris that. 699176866Skris 699276866Skris - Make as few rebuilds of the shared libraries as possible. 699376866Skris 699476866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 699576866Skris 699676866Skris - When installing, install the shared libraries separately from the 699776866Skris static ones. 699876866Skris [Richard Levitte] 699976866Skris 700072613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 700172613Skris 700272613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 700372613Skris and not in SSL_clear because the latter is also used by the 700472613Skris accept/connect functions; previously, the settings made by 700572613Skris SSL_set_read_ahead would be lost during the handshake. 700672613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 700772613Skris 700872613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 700972613Skris Previously, it would create entries for disableed algorithms no 701072613Skris matter what. 701172613Skris [Richard Levitte] 701272613Skris 701372613Skris *) Added several new manual pages for SSL_* function. 701472613Skris [Lutz Jaenicke] 701572613Skris 701668651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 701768651Skris 701868651Skris *) In ssl23_get_client_hello, generate an error message when faced 701968651Skris with an initial SSL 3.0/TLS record that is too small to contain the 702068651Skris first two bytes of the ClientHello message, i.e. client_version. 702168651Skris (Note that this is a pathologic case that probably has never happened 702268651Skris in real life.) The previous approach was to use the version number 702368651Skris from the record header as a substitute; but our protocol choice 702468651Skris should not depend on that one because it is not authenticated 702568651Skris by the Finished messages. 702668651Skris [Bodo Moeller] 702768651Skris 702868651Skris *) More robust randomness gathering functions for Windows. 702968651Skris [Jeffrey Altman <jaltman@columbia.edu>] 703068651Skris 703168651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 703268651Skris not set then we don't setup the error code for issuer check errors 703368651Skris to avoid possibly overwriting other errors which the callback does 703468651Skris handle. If an application does set the flag then we assume it knows 703568651Skris what it is doing and can handle the new informational codes 703668651Skris appropriately. 703768651Skris [Steve Henson] 703868651Skris 703968651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 704068651Skris a general "ANY" type, as such it should be able to decode anything 704168651Skris including tagged types. However it didn't check the class so it would 704268651Skris wrongly interpret tagged types in the same way as their universal 704368651Skris counterpart and unknown types were just rejected. Changed so that the 704468651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 704568651Skris that is the encoding is stored intact. There is also a new type 704668651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 704768651Skris case we have no idea what the actual type is so we just lump them all 704868651Skris together. 704968651Skris [Steve Henson] 705068651Skris 705168651Skris *) On VMS, stdout may very well lead to a file that is written to 705268651Skris in a record-oriented fashion. That means that every write() will 705368651Skris write a separate record, which will be read separately by the 705468651Skris programs trying to read from it. This can be very confusing. 705568651Skris 705668651Skris The solution is to put a BIO filter in the way that will buffer 705768651Skris text until a linefeed is reached, and then write everything a 705868651Skris line at a time, so every record written will be an actual line, 705968651Skris not chunks of lines and not (usually doesn't happen, but I've 706068651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 706168651Skris the answer. 706268651Skris 706368651Skris Currently, it's a VMS-only method, because that's where it has 706468651Skris been tested well enough. 706568651Skris [Richard Levitte] 706668651Skris 706768651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 706868651Skris it can return incorrect results. 706968651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 707068651Skris but it was in 0.9.6-beta[12].) 707168651Skris [Bodo Moeller] 707268651Skris 707368651Skris *) Disable the check for content being present when verifying detached 707468651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 707568651Skris include zero length content when signing messages. 707668651Skris [Steve Henson] 707768651Skris 707868651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 707968651Skris BIO_ctrl (for BIO pairs). 708068651Skris [Bodo M�ller] 708168651Skris 708268651Skris *) Add DSO method for VMS. 708368651Skris [Richard Levitte] 708468651Skris 708568651Skris *) Bug fix: Montgomery multiplication could produce results with the 708668651Skris wrong sign. 708768651Skris [Ulf M�ller] 708868651Skris 708968651Skris *) Add RPM specification openssl.spec and modify it to build three 709068651Skris packages. The default package contains applications, application 709168651Skris documentation and run-time libraries. The devel package contains 709268651Skris include files, static libraries and function documentation. The 709368651Skris doc package contains the contents of the doc directory. The original 709468651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 709568651Skris [Richard Levitte] 709668651Skris 709768651Skris *) Add a large number of documentation files for many SSL routines. 709868651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 709968651Skris 710068651Skris *) Add a configuration entry for Sony News 4. 710168651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 710268651Skris 710368651Skris *) Don't set the two most significant bits to one when generating a 710468651Skris random number < q in the DSA library. 710568651Skris [Ulf M�ller] 710668651Skris 710768651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 710868651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 710968651Skris the underlying transport is blocking) if a handshake took place. 711068651Skris (The default behaviour is needed by applications such as s_client 711168651Skris and s_server that use select() to determine when to use SSL_read; 711268651Skris but for applications that know in advance when to expect data, it 711368651Skris just makes things more complicated.) 711468651Skris [Bodo Moeller] 711568651Skris 711668651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 711768651Skris from EGD. 711868651Skris [Ben Laurie] 711968651Skris 712068651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 712168651Skris work better on such systems. 712268651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 712368651Skris 712468651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 712568651Skris Update PKCS12_parse() so it copies the friendlyName and the 712668651Skris keyid to the certificates aux info. 712768651Skris [Steve Henson] 712868651Skris 712968651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 713068651Skris if there was more than one signature. 713168651Skris [Sven Uszpelkat <su@celocom.de>] 713268651Skris 713368651Skris *) Major change in util/mkdef.pl to include extra information 713468651Skris about each symbol, as well as presentig variables as well 713568651Skris as functions. This change means that there's n more need 713668651Skris to rebuild the .num files when some algorithms are excluded. 713768651Skris [Richard Levitte] 713868651Skris 713968651Skris *) Allow the verify time to be set by an application, 714068651Skris rather than always using the current time. 714168651Skris [Steve Henson] 714268651Skris 714368651Skris *) Phase 2 verify code reorganisation. The certificate 714468651Skris verify code now looks up an issuer certificate by a 714568651Skris number of criteria: subject name, authority key id 714668651Skris and key usage. It also verifies self signed certificates 714768651Skris by the same criteria. The main comparison function is 714868651Skris X509_check_issued() which performs these checks. 714968651Skris 715068651Skris Lot of changes were necessary in order to support this 715168651Skris without completely rewriting the lookup code. 715268651Skris 715368651Skris Authority and subject key identifier are now cached. 715468651Skris 715568651Skris The LHASH 'certs' is X509_STORE has now been replaced 715668651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 715768651Skris LHASH can't store or retrieve multiple objects with 715868651Skris the same hash value. 715968651Skris 716068651Skris As a result various functions (which were all internal 716168651Skris use only) have changed to handle the new X509_STORE 716268651Skris structure. This will break anything that messed round 716368651Skris with X509_STORE internally. 716468651Skris 716568651Skris The functions X509_STORE_add_cert() now checks for an 716668651Skris exact match, rather than just subject name. 716768651Skris 716868651Skris The X509_STORE API doesn't directly support the retrieval 716968651Skris of multiple certificates matching a given criteria, however 717068651Skris this can be worked round by performing a lookup first 717168651Skris (which will fill the cache with candidate certificates) 717268651Skris and then examining the cache for matches. This is probably 717368651Skris the best we can do without throwing out X509_LOOKUP 717468651Skris entirely (maybe later...). 717568651Skris 717668651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 717768651Skris 717868651Skris All certificate lookup operations now go via a get_issuer() 717968651Skris callback. Although this currently uses an X509_STORE it 718068651Skris can be replaced by custom lookups. This is a simple way 718168651Skris to bypass the X509_STORE hackery necessary to make this 718268651Skris work and makes it possible to use more efficient techniques 718368651Skris in future. A very simple version which uses a simple 718468651Skris STACK for its trusted certificate store is also provided 718568651Skris using X509_STORE_CTX_trusted_stack(). 718668651Skris 718768651Skris The verify_cb() and verify() callbacks now have equivalents 718868651Skris in the X509_STORE_CTX structure. 718968651Skris 719068651Skris X509_STORE_CTX also has a 'flags' field which can be used 719168651Skris to customise the verify behaviour. 719268651Skris [Steve Henson] 719368651Skris 719468651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 719568651Skris excludes S/MIME capabilities. 719668651Skris [Steve Henson] 719768651Skris 719868651Skris *) When a certificate request is read in keep a copy of the 719968651Skris original encoding of the signed data and use it when outputing 720068651Skris again. Signatures then use the original encoding rather than 720168651Skris a decoded, encoded version which may cause problems if the 720268651Skris request is improperly encoded. 720368651Skris [Steve Henson] 720468651Skris 720568651Skris *) For consistency with other BIO_puts implementations, call 720668651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 720768651Skris BIO_write(b, ...). 720868651Skris 720968651Skris In BIO_puts, increment b->num_write as in BIO_write. 721068651Skris [Peter.Sylvester@EdelWeb.fr] 721168651Skris 721268651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 721368651Skris BN_zero, we may not return a BIGNUM with an array consisting of 721468651Skris words set to zero.) 721568651Skris [Bodo Moeller] 721668651Skris 721768651Skris *) Avoid calling abort() from within the library when problems are 721868651Skris detected, except if preprocessor symbols have been defined 721968651Skris (such as REF_CHECK, BN_DEBUG etc.). 722068651Skris [Bodo Moeller] 722168651Skris 722268651Skris *) New openssl application 'rsautl'. This utility can be 722368651Skris used for low level RSA operations. DER public key 722468651Skris BIO/fp routines also added. 722568651Skris [Steve Henson] 722668651Skris 722768651Skris *) New Configure entry and patches for compiling on QNX 4. 722868651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 722968651Skris 723068651Skris *) A demo state-machine implementation was sponsored by 723168651Skris Nuron (http://www.nuron.com/) and is now available in 723268651Skris demos/state_machine. 723368651Skris [Ben Laurie] 723468651Skris 723568651Skris *) New options added to the 'dgst' utility for signature 723668651Skris generation and verification. 723768651Skris [Steve Henson] 723868651Skris 723968651Skris *) Unrecognized PKCS#7 content types are now handled via a 724068651Skris catch all ASN1_TYPE structure. This allows unsupported 724168651Skris types to be stored as a "blob" and an application can 724268651Skris encode and decode it manually. 724368651Skris [Steve Henson] 724468651Skris 724568651Skris *) Fix various signed/unsigned issues to make a_strex.c 724668651Skris compile under VC++. 724768651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 724868651Skris 724968651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 725068651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 725168651Skris if passed a NULL BN and its argument was negative. 725268651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 725368651Skris 725468651Skris *) Modification to PKCS#7 encoding routines to output definite 725568651Skris length encoding. Since currently the whole structures are in 725668651Skris memory there's not real point in using indefinite length 725768651Skris constructed encoding. However if OpenSSL is compiled with 725868651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 725968651Skris [Steve Henson] 726068651Skris 726168651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 726268651Skris [Richard Levitte] 726368651Skris 726468651Skris *) Added more prefixes to parse for in the the strings written 726568651Skris through a logging bio, to cover all the levels that are available 726668651Skris through syslog. The prefixes are now: 726768651Skris 726868651Skris PANIC, EMERG, EMR => LOG_EMERG 726968651Skris ALERT, ALR => LOG_ALERT 727068651Skris CRIT, CRI => LOG_CRIT 727168651Skris ERROR, ERR => LOG_ERR 727268651Skris WARNING, WARN, WAR => LOG_WARNING 727368651Skris NOTICE, NOTE, NOT => LOG_NOTICE 727468651Skris INFO, INF => LOG_INFO 727568651Skris DEBUG, DBG => LOG_DEBUG 727668651Skris 727768651Skris and as before, if none of those prefixes are present at the 727868651Skris beginning of the string, LOG_ERR is chosen. 727968651Skris 728068651Skris On Win32, the LOG_* levels are mapped according to this: 728168651Skris 728268651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 728368651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 728468651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 728568651Skris 728668651Skris [Richard Levitte] 728768651Skris 728868651Skris *) Made it possible to reconfigure with just the configuration 728968651Skris argument "reconf" or "reconfigure". The command line arguments 729068651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 729168651Skris and are retrieved from there when reconfiguring. 729268651Skris [Richard Levitte] 729368651Skris 729468651Skris *) MD4 implemented. 729568651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 729668651Skris 729768651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 729868651Skris [Richard Levitte] 729968651Skris 730068651Skris *) The obj_dat.pl script was messing up the sorting of object 730168651Skris names. The reason was that it compared the quoted version 730268651Skris of strings as a result "OCSP" > "OCSP Signing" because 730368651Skris " > SPACE. Changed script to store unquoted versions of 730468651Skris names and add quotes on output. It was also omitting some 730568651Skris names from the lookup table if they were given a default 730668651Skris value (that is if SN is missing it is given the same 730768651Skris value as LN and vice versa), these are now added on the 730868651Skris grounds that if an object has a name we should be able to 730968651Skris look it up. Finally added warning output when duplicate 731068651Skris short or long names are found. 731168651Skris [Steve Henson] 731268651Skris 731368651Skris *) Changes needed for Tandem NSK. 731468651Skris [Scott Uroff <scott@xypro.com>] 731568651Skris 731668651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 731768651Skris RSA_padding_check_SSLv23(), special padding was never detected 731868651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 731968651Skris version rollback attacks was not effective. 732068651Skris 732168651Skris In s23_clnt.c, don't use special rollback-attack detection padding 732268651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 732368651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 732468651Skris SSL 2.0 is the only protocol enabled in the server. 732568651Skris [Bodo Moeller] 732668651Skris 732768651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 732868651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 732968651Skris BIO_dump_indent() are added. 733068651Skris [Richard Levitte] 733168651Skris 733268651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 733368651Skris these print out strings and name structures based on various 733468651Skris flags including RFC2253 support and proper handling of 733568651Skris multibyte characters. Added options to the 'x509' utility 733668651Skris to allow the various flags to be set. 733768651Skris [Steve Henson] 733868651Skris 733968651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 734068651Skris Also change the functions X509_cmp_current_time() and 734168651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 734268651Skris this will enable certificates using GeneralizedTime in validity 734368651Skris dates to be checked. 734468651Skris [Steve Henson] 734568651Skris 734668651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 734768651Skris negative public key encodings) on by default, 734868651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 734968651Skris [Steve Henson] 735068651Skris 735168651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 735268651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 735368651Skris the encoding can be trivially obtained from the structure. 735468651Skris [Steve Henson] 735568651Skris 735668651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 735768651Skris not read locks (CRYPTO_r_[un]lock). 735868651Skris [Bodo Moeller] 735968651Skris 736068651Skris *) A first attempt at creating official support for shared 736168651Skris libraries through configuration. I've kept it so the 736268651Skris default is static libraries only, and the OpenSSL programs 736368651Skris are always statically linked for now, but there are 736468651Skris preparations for dynamic linking in place. 736589837Skris This has been tested on Linux and Tru64. 736668651Skris [Richard Levitte] 736768651Skris 736868651Skris *) Randomness polling function for Win9x, as described in: 736968651Skris Peter Gutmann, Software Generation of Practically Strong 737068651Skris Random Numbers. 737168651Skris [Ulf M�ller] 737268651Skris 737368651Skris *) Fix so PRNG is seeded in req if using an already existing 737468651Skris DSA key. 737568651Skris [Steve Henson] 737668651Skris 737768651Skris *) New options to smime application. -inform and -outform 737868651Skris allow alternative formats for the S/MIME message including 737968651Skris PEM and DER. The -content option allows the content to be 738068651Skris specified separately. This should allow things like Netscape 738168651Skris form signing output easier to verify. 738268651Skris [Steve Henson] 738368651Skris 738468651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 738568651Skris [Steve Henson] 738668651Skris 738768651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 738868651Skris STRING types. These convert content octets to and from the 738968651Skris underlying type. The actual tag and length octets are 739068651Skris already assumed to have been read in and checked. These 739168651Skris are needed because all other string types have virtually 739268651Skris identical handling apart from the tag. By having versions 739368651Skris of the ASN1 functions that just operate on content octets 739468651Skris IMPLICIT tagging can be handled properly. It also allows 739568651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 739668651Skris and ASN1_INTEGER are identical apart from the tag. 739768651Skris [Steve Henson] 739868651Skris 739968651Skris *) Change the handling of OID objects as follows: 740068651Skris 740168651Skris - New object identifiers are inserted in objects.txt, following 740268651Skris the syntax given in objects.README. 740368651Skris - objects.pl is used to process obj_mac.num and create a new 740468651Skris obj_mac.h. 740568651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 740668651Skris obj_mac.h. 740768651Skris 740868651Skris This is currently kind of a hack, and the perl code in objects.pl 740968651Skris isn't very elegant, but it works as I intended. The simplest way 741068651Skris to check that it worked correctly is to look in obj_dat.h and 741168651Skris check the array nid_objs and make sure the objects haven't moved 741268651Skris around (this is important!). Additions are OK, as well as 741368651Skris consistent name changes. 741468651Skris [Richard Levitte] 741568651Skris 741668651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 741768651Skris [Bodo Moeller] 741868651Skris 741968651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 742068651Skris The given file adds to whatever has already been seeded into the 742168651Skris random pool through the RANDFILE configuration file option or 742268651Skris environment variable, or the default random state file. 742368651Skris [Richard Levitte] 742468651Skris 742568651Skris *) mkstack.pl now sorts each macro group into lexical order. 742668651Skris Previously the output order depended on the order the files 742768651Skris appeared in the directory, resulting in needless rewriting 742868651Skris of safestack.h . 742968651Skris [Steve Henson] 743068651Skris 743168651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 743268651Skris work arounds for the VC++ problem that it treats func() as 743368651Skris func(void). Also stripped out the parts of mkdef.pl that 743468651Skris added extra typesafe functions: these no longer exist. 743568651Skris [Steve Henson] 743668651Skris 743768651Skris *) Reorganisation of the stack code. The macros are now all 743868651Skris collected in safestack.h . Each macro is defined in terms of 743968651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 744068651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 744168651Skris this has the advantage of retaining type safety without the 744268651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 744368651Skris then the non typesafe macros are used instead. Also modified the 744468651Skris mkstack.pl script to handle the new form. Needs testing to see 744568651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 744668651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 744768651Skris and PKCS12_STACK_OF. 744868651Skris [Steve Henson] 744968651Skris 745068651Skris *) When some versions of IIS use the 'NET' form of private key the 745168651Skris key derivation algorithm is different. Normally MD5(password) is 745268651Skris used as a 128 bit RC4 key. In the modified case 745368651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 745468651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 745568651Skris as the old Netscape_RSA functions except they have an additional 745668651Skris 'sgckey' parameter which uses the modified algorithm. Also added 745768651Skris an -sgckey command line option to the rsa utility. Thanks to 745868651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 745968651Skris algorithm to openssl-dev. 746068651Skris [Steve Henson] 746168651Skris 746268651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 746368651Skris invalid expansion on some systems (SCO 5.0.5 for example). 746468651Skris Corrected to 'c.kname'. 746568651Skris [Phillip Porch <root@theporch.com>] 746668651Skris 746768651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 746868651Skris a STACK of email addresses from a certificate or request, these look 746968651Skris in the subject name and the subject alternative name extensions and 747068651Skris omit any duplicate addresses. 747168651Skris [Steve Henson] 747268651Skris 747368651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 747468651Skris This makes DSA verification about 2 % faster. 747568651Skris [Bodo Moeller] 747668651Skris 747768651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 747868651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 747968651Skris plus overhead for 1024 bit moduli). 748068651Skris This makes exponentiations about 0.5 % faster for 1024 bit 748168651Skris exponents (as measured by "openssl speed rsa2048"). 748268651Skris [Bodo Moeller] 748368651Skris 748468651Skris *) Rename memory handling macros to avoid conflicts with other 748568651Skris software: 748668651Skris Malloc => OPENSSL_malloc 748768651Skris Malloc_locked => OPENSSL_malloc_locked 748868651Skris Realloc => OPENSSL_realloc 748968651Skris Free => OPENSSL_free 749068651Skris [Richard Levitte] 749168651Skris 749268651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 749368651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 749468651Skris [Bodo Moeller] 749568651Skris 749668651Skris *) CygWin32 support. 749768651Skris [John Jarvie <jjarvie@newsguy.com>] 749868651Skris 749968651Skris *) The type-safe stack code has been rejigged. It is now only compiled 750068651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 750168651Skris by default all type-specific stack functions are "#define"d back to 750268651Skris standard stack functions. This results in more streamlined output 750368651Skris but retains the type-safety checking possibilities of the original 750468651Skris approach. 750568651Skris [Geoff Thorpe] 750668651Skris 750768651Skris *) The STACK code has been cleaned up, and certain type declarations 750868651Skris that didn't make a lot of sense have been brought in line. This has 750968651Skris also involved a cleanup of sorts in safestack.h to more correctly 751068651Skris map type-safe stack functions onto their plain stack counterparts. 751168651Skris This work has also resulted in a variety of "const"ifications of 751268651Skris lots of the code, especially "_cmp" operations which should normally 751368651Skris be prototyped with "const" parameters anyway. 751468651Skris [Geoff Thorpe] 751568651Skris 751668651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 751768651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 751868651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 751968651Skris where all of 'md' is used each time the PRNG is used, but 'state' 752068651Skris is used only indexed by a cyclic counter. As entropy may not be 752168651Skris well distributed from the beginning, 'md' is important as a 752268651Skris chaining variable. However, the output function chains only half 752368651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 752468651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 752568651Skris in all of 'state' being rewritten, with the new values depending 752668651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 752768651Skris [Bodo Moeller] 752868651Skris 752968651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 753068651Skris the handshake is continued after ssl_verify_cert_chain(); 753168651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 753268651Skris can lead to 'unexplainable' connection aborts later. 753368651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 753468651Skris 753568651Skris *) Major EVP API cipher revision. 753668651Skris Add hooks for extra EVP features. This allows various cipher 753768651Skris parameters to be set in the EVP interface. Support added for variable 753868651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 753968651Skris setting of RC2 and RC5 parameters. 754068651Skris 754168651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 754268651Skris ciphers. 754368651Skris 754468651Skris Remove lots of duplicated code from the EVP library. For example *every* 754568651Skris cipher init() function handles the 'iv' in the same way according to the 754668651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 754768651Skris for CFB and OFB modes they zero ctx->num. 754868651Skris 754968651Skris New functionality allows removal of S/MIME code RC2 hack. 755068651Skris 755168651Skris Most of the routines have the same form and so can be declared in terms 755268651Skris of macros. 755368651Skris 755468651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 755568651Skris all individual ciphers. If the cipher wants to handle IVs or keys 755668651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 755768651Skris flags. 755868651Skris 755968651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 756068651Skris value: although software versions of the algorithms cannot fail 756168651Skris any installed hardware versions can. 756268651Skris [Steve Henson] 756368651Skris 756468651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 756568651Skris this option is set, tolerate broken clients that send the negotiated 756668651Skris protocol version number instead of the requested protocol version 756768651Skris number. 756868651Skris [Bodo Moeller] 756968651Skris 757068651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 757168651Skris i.e. non-zero for export ciphersuites, zero otherwise. 757268651Skris Previous versions had this flag inverted, inconsistent with 757368651Skris rsa_tmp_cb (..._TMP_RSA_CB). 757468651Skris [Bodo Moeller; problem reported by Amit Chopra] 757568651Skris 757668651Skris *) Add missing DSA library text string. Work around for some IIS 757768651Skris key files with invalid SEQUENCE encoding. 757868651Skris [Steve Henson] 757968651Skris 758068651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 758168651Skris and so on that are implemented in OpenSSL. 758268651Skris [Richard Levitte] 758368651Skris 758468651Skris *) Enhance c_rehash script. Old version would mishandle certificates 758568651Skris with the same subject name hash and wouldn't handle CRLs at all. 758668651Skris Added -fingerprint option to crl utility, to support new c_rehash 758768651Skris features. 758868651Skris [Steve Henson] 758968651Skris 759068651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 759168651Skris [Ulf M�ller] 759268651Skris 759368651Skris *) Fix for SSL server purpose checking. Server checking was 759468651Skris rejecting certificates which had extended key usage present 759568651Skris but no ssl client purpose. 759668651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 759768651Skris 759868651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 759968651Skris is a little unclear about how a blank password is handled. 760068651Skris Since the password in encoded as a BMPString with terminating 760168651Skris double NULL a zero length password would end up as just the 760268651Skris double NULL. However no password at all is different and is 760368651Skris handled differently in the PKCS#12 key generation code. NS 760468651Skris treats a blank password as zero length. MSIE treats it as no 760568651Skris password on export: but it will try both on import. We now do 760668651Skris the same: PKCS12_parse() tries zero length and no password if 760768651Skris the password is set to "" or NULL (NULL is now a valid password: 760868651Skris it wasn't before) as does the pkcs12 application. 760968651Skris [Steve Henson] 761068651Skris 761168651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 761268651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 761368651Skris be obtained from the error queue. 761468651Skris [Bodo Moeller] 761568651Skris 761668651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 761768651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 761868651Skris accordingly to avoid race conditions (this is necessary because 761968651Skris thread_hash is no longer constant once set). 762068651Skris [Bodo Moeller] 762168651Skris 762268651Skris *) Bugfix for linux-elf makefile.one. 762368651Skris [Ulf M�ller] 762468651Skris 762568651Skris *) RSA_get_default_method() will now cause a default 762668651Skris RSA_METHOD to be chosen if one doesn't exist already. 762768651Skris Previously this was only set during a call to RSA_new() 762868651Skris or RSA_new_method(NULL) meaning it was possible for 762968651Skris RSA_get_default_method() to return NULL. 763068651Skris [Geoff Thorpe] 763168651Skris 763268651Skris *) Added native name translation to the existing DSO code 763368651Skris that will convert (if the flag to do so is set) filenames 763468651Skris that are sufficiently small and have no path information 763568651Skris into a canonical native form. Eg. "blah" converted to 763668651Skris "libblah.so" or "blah.dll" etc. 763768651Skris [Geoff Thorpe] 763868651Skris 763968651Skris *) New function ERR_error_string_n(e, buf, len) which is like 764068651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 764168651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 764268651Skris may not be NULL. 764368651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 764468651Skris 764568651Skris *) CONF library reworked to become more general. A new CONF 764668651Skris configuration file reader "class" is implemented as well as a 764768651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 764868651Skris old CONF_* functions are still there, but are reimplemented to 764968651Skris work in terms of the new functions. Also, a set of functions 765068651Skris to handle the internal storage of the configuration data is 765168651Skris provided to make it easier to write new configuration file 765268651Skris reader "classes" (I can definitely see something reading a 765368651Skris configuration file in XML format, for example), called _CONF_*, 765468651Skris or "the configuration storage API"... 765568651Skris 765668651Skris The new configuration file reading functions are: 765768651Skris 765868651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 765968651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 766068651Skris 766168651Skris NCONF_default, NCONF_WIN32 766268651Skris 766368651Skris NCONF_dump_fp, NCONF_dump_bio 766468651Skris 766568651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 766668651Skris NCONF_new creates a new CONF object. This works in the same way 766768651Skris as other interfaces in OpenSSL, like the BIO interface. 766868651Skris NCONF_dump_* dump the internal storage of the configuration file, 766968651Skris which is useful for debugging. All other functions take the same 767068651Skris arguments as the old CONF_* functions wth the exception of the 767168651Skris first that must be a `CONF *' instead of a `LHASH *'. 767268651Skris 767368651Skris To make it easer to use the new classes with the old CONF_* functions, 767468651Skris the function CONF_set_default_method is provided. 767568651Skris [Richard Levitte] 767668651Skris 767768651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 767868651Skris mentioned in the documentation but had not been implemented. 767968651Skris (This option is not yet really useful because even the additional 768068651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 768168651Skris [Bodo Moeller] 768268651Skris 768368651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 768468651Skris OpenSSL-based applications) load shared libraries and bind to 768568651Skris them in a portable way. 768668651Skris [Geoff Thorpe, with contributions from Richard Levitte] 768768651Skris 768859191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 768959191Skris 769059191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 769159191Skris 769259191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 769359191Skris (the default implementation of RAND_status). 769459191Skris 769559191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 769659191Skris to '-clrext' (= clear extensions), as intended and documented. 769759191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 769859191Skris <attili@amaxo.com>] 769959191Skris 770059191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 770159191Skris was larger than the MD block size. 770259191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 770359191Skris 770459191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 770559191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 770659191Skris using the passed key: if the passed key was a private key the result 770759191Skris of X509_print(), for example, would be to print out all the private key 770859191Skris components. 770959191Skris [Steve Henson] 771059191Skris 771159191Skris *) des_quad_cksum() byte order bug fix. 771259191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 771359191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 771459191Skris 771559191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 771659191Skris discouraged. 771759191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 771859191Skris 771959191Skris *) For easily testing in shell scripts whether some command 772059191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 772159191Skris returns with exit code 0 iff no command of the given name is available. 772259191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 772359191Skris the output goes to stdout and nothing is printed to stderr. 772459191Skris Additional arguments are always ignored. 772559191Skris 772659191Skris Since for each cipher there is a command of the same name, 772759191Skris the 'no-cipher' compilation switches can be tested this way. 772859191Skris 772959191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 773059191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 773159191Skris [Bodo Moeller] 773259191Skris 773359191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 773459191Skris [Bodo Moeller] 773559191Skris 773659191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 773759191Skris is set; it will be thrown away anyway because each handshake creates 773859191Skris its own key. 773959191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 774059191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 774159191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 774259191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 774359191Skris [Bodo Moeller] 774459191Skris 774559191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 774659191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 774759191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 774859191Skris does not suppress any output. 774959191Skris [Richard Levitte] 775059191Skris 775159191Skris *) Add compatibility options to the purpose and trust code. The 775259191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 775359191Skris accepts a certificate or CA, this was the previous behaviour, 775459191Skris with all the associated security issues. 775559191Skris 775659191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 775759191Skris automatically trust self signed roots in certificate store. A 775859191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 775959191Skris a purpose has no associated trust setting and it should instead 776059191Skris use the value in the default purpose. 776159191Skris [Steve Henson] 776259191Skris 776359191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 776459191Skris and fix a memory leak. 776559191Skris [Steve Henson] 776659191Skris 776759191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 776859191Skris reason strings from the previous version of the .c file, as 776959191Skris the default to have only downcase letters (and digits) in 777059191Skris automatically generated reasons codes is not always appropriate. 777159191Skris [Bodo Moeller] 777259191Skris 777359191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 777459191Skris using strerror. Previously, ERR_reason_error_string() returned 777559191Skris library names as reason strings for SYSerr; but SYSerr is a special 777659191Skris case where small numbers are errno values, not library numbers. 777759191Skris [Bodo Moeller] 777859191Skris 777959191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 778059191Skris converts DSA parameters into DH parameters. (When creating parameters, 778159191Skris DSA_generate_parameters is used.) 778259191Skris [Bodo Moeller] 778359191Skris 778459191Skris *) Include 'length' (recommended exponent length) in C code generated 778559191Skris by 'openssl dhparam -C'. 778659191Skris [Bodo Moeller] 778759191Skris 778859191Skris *) The second argument to set_label in perlasm was already being used 778959191Skris so couldn't be used as a "file scope" flag. Moved to third argument 779059191Skris which was free. 779159191Skris [Steve Henson] 779259191Skris 779359191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 779459191Skris instead of RAND_bytes for encryption IVs and salts. 779559191Skris [Bodo Moeller] 779659191Skris 779759191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 779859191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 779959191Skris RAND_set_rand_method would be impossible. 780059191Skris [Bodo Moeller] 780159191Skris 780259191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 780359191Skris number generation fails. 780459191Skris [Bodo Moeller] 780559191Skris 780659191Skris *) New 'rand' application for creating pseudo-random output. 780759191Skris [Bodo Moeller] 780859191Skris 780959191Skris *) Added configuration support for Linux/IA64 781059191Skris [Rolf Haberrecker <rolf@suse.de>] 781159191Skris 781259191Skris *) Assembler module support for Mingw32. 781359191Skris [Ulf M�ller] 781459191Skris 781559191Skris *) Shared library support for HPUX (in shlib/). 781659191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 781759191Skris 781859191Skris *) Shared library support for Solaris gcc. 781959191Skris [Lutz Behnke <behnke@trustcenter.de>] 782059191Skris 782159191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 782259191Skris 782359191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 782459191Skris were added manually and by SMIME_crlf_copy(). 782559191Skris [Steve Henson] 782659191Skris 782759191Skris *) In bntest.c don't call BN_rand with zero bits argument. 782859191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 782959191Skris 783059191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 783159191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 783259191Skris [Ulf M�ller] 783359191Skris 783459191Skris *) Add an optional second argument to the set_label() in the perl 783559191Skris assembly language builder. If this argument exists and is set 783659191Skris to 1 it signals that the assembler should use a symbol whose 783759191Skris scope is the entire file, not just the current function. This 783859191Skris is needed with MASM which uses the format label:: for this scope. 783959191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 784059191Skris 784159191Skris *) Change the ASN1 types so they are typedefs by default. Before 784259191Skris almost all types were #define'd to ASN1_STRING which was causing 784359191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 784459191Skris for example. 784559191Skris [Steve Henson] 784659191Skris 784759191Skris *) Change names of new functions to the new get1/get0 naming 784859191Skris convention: After 'get1', the caller owns a reference count 784959191Skris and has to call ..._free; 'get0' returns a pointer to some 785059191Skris data structure without incrementing reference counters. 785159191Skris (Some of the existing 'get' functions increment a reference 785259191Skris counter, some don't.) 785359191Skris Similarly, 'set1' and 'add1' functions increase reference 785459191Skris counters or duplicate objects. 785559191Skris [Steve Henson] 785659191Skris 785759191Skris *) Allow for the possibility of temp RSA key generation failure: 785859191Skris the code used to assume it always worked and crashed on failure. 785959191Skris [Steve Henson] 786059191Skris 786159191Skris *) Fix potential buffer overrun problem in BIO_printf(). 786259191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 786359191Skris pointed out by David Sacerdote <das33@cornell.edu>] 786459191Skris 786559191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 786659191Skris RAND_egd() and RAND_status(). In the command line application, 786759191Skris the EGD socket can be specified like a seed file using RANDFILE 786859191Skris or -rand. 786959191Skris [Ulf M�ller] 787059191Skris 787159191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 787259191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 787359191Skris [Steve Henson] 787459191Skris 787559191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 787659191Skris list to exclude them. This means that no special compilation option 787759191Skris is needed to use anonymous DH: it just needs to be included in the 787859191Skris cipher list. 787959191Skris [Steve Henson] 788059191Skris 788159191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 788259191Skris EVP_MD_type. The old functionality is available in a new macro called 788359191Skris EVP_MD_md(). Change code that uses it and update docs. 788459191Skris [Steve Henson] 788559191Skris 788659191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 788759191Skris where the 'void *' argument is replaced by a function pointer argument. 788859191Skris Previously 'void *' was abused to point to functions, which works on 788959191Skris many platforms, but is not correct. As these functions are usually 789059191Skris called by macros defined in OpenSSL header files, most source code 789159191Skris should work without changes. 789259191Skris [Richard Levitte] 789359191Skris 789459191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 789559191Skris sections with information on -D... compiler switches used for 789659191Skris compiling the library so that applications can see them. To enable 789759191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 789859191Skris must be defined. E.g., 789959191Skris #define OPENSSL_ALGORITHM_DEFINES 790059191Skris #include <openssl/opensslconf.h> 790159191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 790259191Skris [Richard Levitte, Ulf and Bodo M�ller] 790359191Skris 790459191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 790559191Skris record layer. 790659191Skris [Bodo Moeller] 790759191Skris 790859191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 790959191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 791059191Skris the required ASN1 format: arbitrary types determined by an OID. 791159191Skris [Steve Henson] 791259191Skris 791359191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 791459191Skris argument to 'req'. This is not because the function is newer or 791559191Skris better than others it just uses the work 'NEW' in the certificate 791659191Skris request header lines. Some software needs this. 791759191Skris [Steve Henson] 791859191Skris 791959191Skris *) Reorganise password command line arguments: now passwords can be 792059191Skris obtained from various sources. Delete the PEM_cb function and make 792159191Skris it the default behaviour: i.e. if the callback is NULL and the 792259191Skris usrdata argument is not NULL interpret it as a null terminated pass 792359191Skris phrase. If usrdata and the callback are NULL then the pass phrase 792459191Skris is prompted for as usual. 792559191Skris [Steve Henson] 792659191Skris 792759191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 792859191Skris the support is automatically enabled. The resulting binaries will 792959191Skris autodetect the card and use it if present. 793059191Skris [Ben Laurie and Compaq Inc.] 793159191Skris 793259191Skris *) Work around for Netscape hang bug. This sends certificate request 793359191Skris and server done in one record. Since this is perfectly legal in the 793459191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 793559191Skris the bugs/SSLv3 entry for more info. 793659191Skris [Steve Henson] 793759191Skris 793859191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 793959191Skris [Andy Polyakov] 794059191Skris 794159191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 794259191Skris of seed file. 794359191Skris [Steve Henson] 794459191Skris 794559191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 794659191Skris [Bodo Moeller] 794759191Skris 794859191Skris *) Add command line password options to the remaining applications. 794959191Skris [Steve Henson] 795059191Skris 795159191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 795259191Skris bits. 795359191Skris [Ulf M�ller] 795459191Skris 795559191Skris *) More tests in bntest.c, and changed test_bn output. 795659191Skris [Ulf M�ller] 795759191Skris 795859191Skris *) ./config recognizes MacOS X now. 795959191Skris [Andy Polyakov] 796059191Skris 796159191Skris *) Bug fix for BN_div() when the first words of num and divsor are 796259191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 796359191Skris [Ulf M�ller] 796459191Skris 796559191Skris *) Add support for various broken PKCS#8 formats, and command line 796659191Skris options to produce them. 796759191Skris [Steve Henson] 796859191Skris 796959191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 797059191Skris get temporary BIGNUMs from a BN_CTX. 797159191Skris [Ulf M�ller] 797259191Skris 797359191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 797459191Skris for p == 0. 797559191Skris [Ulf M�ller] 797659191Skris 797759191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 797859191Skris include a #define from the old name to the new. The original intent 797959191Skris was that statically linked binaries could for example just call 798059191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 798159191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 798259191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 798359191Skris one would link with the other. They are now in separate source files. 798459191Skris [Steve Henson] 798559191Skris 798659191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 798759191Skris [Steve Henson] 798859191Skris 798959191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 799059191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 799159191Skris loop, our standard modexp algorithms are faster). 799259191Skris [Bodo Moeller] 799359191Skris 799459191Skris *) Support for the EBCDIC character set completed. 799559191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 799659191Skris 799759191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 799859191Skris use void * instead of char * in lhash. 799959191Skris [Ulf M�ller] 800059191Skris 800159191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 800259191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 800359191Skris this the server could overwrite ephemeral keys that the client 800459191Skris has already seen). 800559191Skris [Bodo Moeller] 800659191Skris 800759191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 800859191Skris using 50 iterations of the Rabin-Miller test. 800959191Skris 801059191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 801159191Skris iterations of the Rabin-Miller test as required by the appendix 801259191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 801359191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 801459191Skris generation becomes much faster. 801559191Skris 801659191Skris This implies a change for the callback functions in DSA_is_prime 801759191Skris and DSA_generate_parameters: The callback function is called once 801859191Skris for each positive witness in the Rabin-Miller test, not just 801959191Skris occasionally in the inner loop; and the parameters to the 802059191Skris callback function now provide an iteration count for the outer 802159191Skris loop rather than for the current invocation of the inner loop. 802259191Skris DSA_generate_parameters additionally can call the callback 802359191Skris function with an 'iteration count' of -1, meaning that a 802459191Skris candidate has passed the trial division test (when q is generated 802559191Skris from an application-provided seed, trial division is skipped). 802659191Skris [Bodo Moeller] 802759191Skris 802859191Skris *) New function BN_is_prime_fasttest that optionally does trial 802959191Skris division before starting the Rabin-Miller test and has 803059191Skris an additional BN_CTX * argument (whereas BN_is_prime always 803159191Skris has to allocate at least one BN_CTX). 803259191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 803359191Skris trial division stage. 803459191Skris [Bodo Moeller] 803559191Skris 803659191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 803759191Skris as ASN1_TIME. 803859191Skris [Steve Henson] 803959191Skris 804059191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 804159191Skris [Steve Henson] 804259191Skris 804359191Skris *) New function BN_pseudo_rand(). 804459191Skris [Ulf M�ller] 804559191Skris 804659191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 804759191Skris bignum version of BN_from_montgomery() with the working code from 804859191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 804959191Skris the comments. 805059191Skris [Ulf M�ller] 805159191Skris 805259191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 805359191Skris made it impossible to use the same SSL_SESSION data structure in 805459191Skris SSL2 clients in multiple threads. 805559191Skris [Bodo Moeller] 805659191Skris 805759191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 805859191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 805959191Skris to seed the PRNG (previously an explicit byte count was required). 806059191Skris [Ulf M�ller, Bodo M�ller] 806159191Skris 806259191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 806359191Skris used (char *) instead of (void *) and had casts all over the place. 806459191Skris [Steve Henson] 806559191Skris 806659191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 806759191Skris [Ulf M�ller] 806859191Skris 806959191Skris *) Retain source code compatibility for BN_prime_checks macro: 807059191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 807159191Skris BN_prime_checks_for_size to determine the appropriate number of 807259191Skris Rabin-Miller iterations. 807359191Skris [Ulf M�ller] 807459191Skris 807559191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 807659191Skris DH_CHECK_P_NOT_SAFE_PRIME. 807759191Skris (Check if this is true? OpenPGP calls them "strong".) 807859191Skris [Ulf M�ller] 807959191Skris 808059191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 808159191Skris "dhparam". The old programs are retained for now but will handle DH keys 808259191Skris (instead of parameters) in future. 808359191Skris [Steve Henson] 808459191Skris 808559191Skris *) Make the ciphers, s_server and s_client programs check the return values 808659191Skris when a new cipher list is set. 808759191Skris [Steve Henson] 808859191Skris 808959191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 809059191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 809159191Skris wrong. 809259191Skris 809359191Skris The syntax for the cipher sorting has been extended to support sorting by 809459191Skris cipher-strength (using the strength_bits hard coded in the tables). 809559191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 809659191Skris 809759191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 809859191Skris string with an "undefined" symbol (neither command nor alphanumeric 809959191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 810059191Skris an error is flagged. 810159191Skris 810259191Skris Due to the strength-sorting extension, the code of the 810359191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 810459191Skris the readability was also increased :-) 810559191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 810659191Skris 810759191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 810859191Skris for the first serial number and places 2 in the serial number file. This 810959191Skris avoids problems when the root CA is created with serial number zero and 811059191Skris the first user certificate has the same issuer name and serial number 811159191Skris as the root CA. 811259191Skris [Steve Henson] 811359191Skris 811459191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 811559191Skris the new code. Add documentation for this stuff. 811659191Skris [Steve Henson] 811759191Skris 811859191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 811959191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 812059191Skris structures and behave in an analagous way to the X509v3 functions: 812159191Skris they shouldn't be called directly but wrapper functions should be used 812259191Skris instead. 812359191Skris 812459191Skris So we also now have some wrapper functions that call the X509at functions 812559191Skris when passed certificate requests. (TO DO: similar things can be done with 812659191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 812759191Skris things. Some of these need some d2i or i2d and print functionality 812859191Skris because they handle more complex structures.) 812959191Skris [Steve Henson] 813059191Skris 813159191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 813259191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 813359191Skris NO_RSA in ssl/s2*.c. 813459191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 813559191Skris 813659191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 813759191Skris has a return value which indicates the quality of the random data 813859191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 813959191Skris error queue. New function RAND_pseudo_bytes() generates output that is 814059191Skris guaranteed to be unique but not unpredictable. RAND_add is like 814159191Skris RAND_seed, but takes an extra argument for an entropy estimate 814259191Skris (RAND_seed always assumes full entropy). 814359191Skris [Ulf M�ller] 814459191Skris 814559191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 814659191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 814759191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 814859191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 814959191Skris false-positive rate of at most 2^-80 for random input. 815059191Skris [Bodo Moeller] 815159191Skris 815259191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 815359191Skris [Bodo Moeller] 815459191Skris 815559191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 815659191Skris in the 0.9.5 release), this returns the chain 815759191Skris from an X509_CTX structure with a dup of the stack and all 815859191Skris the X509 reference counts upped: so the stack will exist 815959191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 816059191Skris to use this. 816159191Skris 816259191Skris Also make SSL_SESSION_print() print out the verify return 816359191Skris code. 816459191Skris [Steve Henson] 816559191Skris 816659191Skris *) Add manpage for the pkcs12 command. Also change the default 816759191Skris behaviour so MAC iteration counts are used unless the new 816859191Skris -nomaciter option is used. This improves file security and 816959191Skris only older versions of MSIE (4.0 for example) need it. 817059191Skris [Steve Henson] 817159191Skris 817259191Skris *) Honor the no-xxx Configure options when creating .DEF files. 817359191Skris [Ulf M�ller] 817459191Skris 817559191Skris *) Add PKCS#10 attributes to field table: challengePassword, 817659191Skris unstructuredName and unstructuredAddress. These are taken from 817759191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 817859191Skris international characters are used. 817959191Skris 818059191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 818159191Skris based on strings. Remove the 'loc' parameter when adding 818259191Skris attributes because these will be a SET OF encoding which is sorted 818359191Skris in ASN1 order. 818459191Skris [Steve Henson] 818559191Skris 818659191Skris *) Initial changes to the 'req' utility to allow request generation 818759191Skris automation. This will allow an application to just generate a template 818859191Skris file containing all the field values and have req construct the 818959191Skris request. 819059191Skris 819159191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 819259191Skris used all over the place including certificate requests and PKCS#7 819359191Skris structures. They are currently handled manually where necessary with 819459191Skris some primitive wrappers for PKCS#7. The new functions behave in a 819559191Skris manner analogous to the X509 extension functions: they allow 819659191Skris attributes to be looked up by NID and added. 819759191Skris 819859191Skris Later something similar to the X509V3 code would be desirable to 819959191Skris automatically handle the encoding, decoding and printing of the 820059191Skris more complex types. The string types like challengePassword can 820159191Skris be handled by the string table functions. 820259191Skris 820359191Skris Also modified the multi byte string table handling. Now there is 820459191Skris a 'global mask' which masks out certain types. The table itself 820559191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 820659191Skris is useful when for example there is only one permissible type 820759191Skris (as in countryName) and using the mask might result in no valid 820859191Skris types at all. 820959191Skris [Steve Henson] 821059191Skris 821159191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 821259191Skris SSL_get_peer_finished to allow applications to obtain the latest 821359191Skris Finished messages sent to the peer or expected from the peer, 821459191Skris respectively. (SSL_get_peer_finished is usually the Finished message 821559191Skris actually received from the peer, otherwise the protocol will be aborted.) 821659191Skris 821759191Skris As the Finished message are message digests of the complete handshake 821859191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 821959191Skris be used for external authentication procedures when the authentication 822059191Skris provided by SSL/TLS is not desired or is not enough. 822159191Skris [Bodo Moeller] 822259191Skris 822359191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 822459191Skris the host supports BWX extension and if Compaq C is present on the 822559191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 822659191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 822759191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 822859191Skris SHA1. 822959191Skris [Andy Polyakov] 823059191Skris 823159191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 823259191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 823359191Skris weak crypto and after checking the certificate is SGC a second one 823459191Skris with strong crypto. MS SGC stops the first handshake after receiving 823559191Skris the server certificate message and sends a second client hello. Since 823659191Skris a server will typically do all the time consuming operations before 823759191Skris expecting any further messages from the client (server key exchange 823859191Skris is the most expensive) there is little difference between the two. 823959191Skris 824059191Skris To get OpenSSL to support MS SGC we have to permit a second client 824159191Skris hello message after we have sent server done. In addition we have to 824259191Skris reset the MAC if we do get this second client hello. 824359191Skris [Steve Henson] 824459191Skris 824559191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 824659191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 824759191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 824859191Skris format DER encoded private key. Newer code should use PKCS#8 format which 824959191Skris has the key type encoded in the ASN1 structure. Added DER private key 825059191Skris support to pkcs8 application. 825159191Skris [Steve Henson] 825259191Skris 825359191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 825459191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 825559191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 825659191Skris is set, we interpret this as a request to violate the specification 825759191Skris (the worst that can happen is a handshake failure, and 'correct' 825859191Skris behaviour would result in a handshake failure anyway). 825959191Skris [Bodo Moeller] 826059191Skris 826159191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 826259191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 826359191Skris concurrently obtain them from an external cache). 826459191Skris The internal cache can handle only one SSL_SESSION with a given ID, 826559191Skris so if there's a conflict, we now throw out the old one to achieve 826659191Skris consistency. 826759191Skris [Bodo Moeller] 826859191Skris 826959191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 827059191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 827159191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 827259191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 827359191Skris example. 827459191Skris [Steve Henson] 827559191Skris 827659191Skris *) Simplify the trust setting structure and code. Now we just have 827759191Skris two sequences of OIDs for trusted and rejected settings. These will 827859191Skris typically have values the same as the extended key usage extension 827959191Skris and any application specific purposes. 828059191Skris 828159191Skris The trust checking code now has a default behaviour: it will just 828259191Skris check for an object with the same NID as the passed id. Functions can 828359191Skris be provided to override either the default behaviour or the behaviour 828459191Skris for a given id. SSL client, server and email already have functions 828559191Skris in place for compatibility: they check the NID and also return "trusted" 828659191Skris if the certificate is self signed. 828759191Skris [Steve Henson] 828859191Skris 828959191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 829059191Skris traditional format into an EVP_PKEY structure. 829159191Skris [Steve Henson] 829259191Skris 829359191Skris *) Add a password callback function PEM_cb() which either prompts for 829459191Skris a password if usr_data is NULL or otherwise assumes it is a null 829559191Skris terminated password. Allow passwords to be passed on command line 829659191Skris environment or config files in a few more utilities. 829759191Skris [Steve Henson] 829859191Skris 829959191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 830059191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 830159191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 830259191Skris Update documentation. 830359191Skris [Steve Henson] 830459191Skris 830559191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 830659191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 830759191Skris and produce an error if it couldn't. For compatibility we also have 830859191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 830959191Skris don't allocate anything because they don't need to. 831059191Skris [Steve Henson] 831159191Skris 831259191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 831359191Skris for details. 831459191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 831559191Skris 831659191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 831759191Skris possibly others as well. The purpose is to make an interface that 831859191Skris provide hooks so anyone can build a separate set of allocation and 831959191Skris deallocation routines to be used by OpenSSL, for example memory 832059191Skris pool implementations, or something else, which was previously hard 832159191Skris since Malloc(), Realloc() and Free() were defined as macros having 832259191Skris the values malloc, realloc and free, respectively (except for Win32 832359191Skris compilations). The same is provided for memory debugging code. 832459191Skris OpenSSL already comes with functionality to find memory leaks, but 832559191Skris this gives people a chance to debug other memory problems. 832659191Skris 832759191Skris With these changes, a new set of functions and macros have appeared: 832859191Skris 832968651Skris CRYPTO_set_mem_debug_functions() [F] 833068651Skris CRYPTO_get_mem_debug_functions() [F] 833168651Skris CRYPTO_dbg_set_options() [F] 833268651Skris CRYPTO_dbg_get_options() [F] 833368651Skris CRYPTO_malloc_debug_init() [M] 833459191Skris 833559191Skris The memory debug functions are NULL by default, unless the library 833659191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 833759191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 833859191Skris gives the standard debugging functions that come with OpenSSL) or 833959191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 834059191Skris provided by the library user) must be used. When the standard 834159191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 834259191Skris request additional information: 834359191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 834459191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 834559191Skris 834659191Skris Also, things like CRYPTO_set_mem_functions will always give the 834759191Skris expected result (the new set of functions is used for allocation 834859191Skris and deallocation) at all times, regardless of platform and compiler 834959191Skris options. 835059191Skris 835159191Skris To finish it up, some functions that were never use in any other 835259191Skris way than through macros have a new API and new semantic: 835359191Skris 835459191Skris CRYPTO_dbg_malloc() 835559191Skris CRYPTO_dbg_realloc() 835659191Skris CRYPTO_dbg_free() 835759191Skris 835859191Skris All macros of value have retained their old syntax. 835959191Skris [Richard Levitte and Bodo Moeller] 836059191Skris 836159191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 836259191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 836359191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 836459191Skris algorithm. 836559191Skris [Steve Henson] 836659191Skris 836759191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 836859191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 836959191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 837059191Skris 837159191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 837259191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 837359191Skris functionality to handle multipart/signed properly) and a utility 837459191Skris called 'smime' to call all this stuff. This is based on code I 837559191Skris originally wrote for Celo who have kindly allowed it to be 837659191Skris included in OpenSSL. 837759191Skris [Steve Henson] 837859191Skris 837959191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 838059191Skris des_set_key (aka des_key_sched). Global variable des_check_key 838159191Skris decides which of these is called by des_set_key; this way 838259191Skris des_check_key behaves as it always did, but applications and 838359191Skris the library itself, which was buggy for des_check_key == 1, 838459191Skris have a cleaner way to pick the version they need. 838559191Skris [Bodo Moeller] 838659191Skris 838759191Skris *) New function PKCS12_newpass() which changes the password of a 838859191Skris PKCS12 structure. 838959191Skris [Steve Henson] 839059191Skris 839159191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 839259191Skris dynamic mix. In both cases the ids can be used as an index into the 839359191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 839459191Skris functions so they accept a list of the field values and the 839559191Skris application doesn't need to directly manipulate the X509_TRUST 839659191Skris structure. 839759191Skris [Steve Henson] 839859191Skris 839959191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 840059191Skris need initialising. 840159191Skris [Steve Henson] 840259191Skris 840359191Skris *) Modify the way the V3 extension code looks up extensions. This now 840459191Skris works in a similar way to the object code: we have some "standard" 840559191Skris extensions in a static table which is searched with OBJ_bsearch() 840659191Skris and the application can add dynamic ones if needed. The file 840759191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 840859191Skris updated whenever a new extension is added to the core code and kept 840959191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 841059191Skris this. New extensions are not added too often so this file can readily 841159191Skris be maintained manually. 841259191Skris 841359191Skris There are two big advantages in doing things this way. The extensions 841459191Skris can be looked up immediately and no longer need to be "added" using 841559191Skris X509V3_add_standard_extensions(): this function now does nothing. 841659191Skris [Side note: I get *lots* of email saying the extension code doesn't 841759191Skris work because people forget to call this function] 841859191Skris Also no dynamic allocation is done unless new extensions are added: 841959191Skris so if we don't add custom extensions there is no need to call 842059191Skris X509V3_EXT_cleanup(). 842159191Skris [Steve Henson] 842259191Skris 842359191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 842459191Skris magic header, so unsalted files fail gracefully instead of just decrypting 842559191Skris to garbage. This is because not salting is a big security hole, so people 842659191Skris should be discouraged from doing it. 842759191Skris [Ben Laurie] 842859191Skris 842959191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 843059191Skris digest to be passed on the command line but it only used this 843159191Skris parameter when signing a certificate. Modified so all relevant 843259191Skris operations are affected by the digest parameter including the 843359191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 843459191Skris DSA key was used because it didn't fix the digest. 843559191Skris [Steve Henson] 843659191Skris 843759191Skris *) Initial certificate chain verify code. Currently tests the untrusted 843859191Skris certificates for consistency with the verify purpose (which is set 843959191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 844059191Skris 844159191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 844259191Skris this is because it will reject chains with invalid extensions whereas 844359191Skris every previous version of OpenSSL and SSLeay made no checks at all. 844459191Skris 844559191Skris Trust code: checks the root CA for the relevant trust settings. Trust 844659191Skris settings have an initial value consistent with the verify purpose: e.g. 844759191Skris if the verify purpose is for SSL client use it expects the CA to be 844859191Skris trusted for SSL client use. However the default value can be changed to 844959191Skris permit custom trust settings: one example of this would be to only trust 845059191Skris certificates from a specific "secure" set of CAs. 845159191Skris 845259191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 845359191Skris which should be used for version portability: especially since the 845459191Skris verify structure is likely to change more often now. 845559191Skris 845659191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 845759191Skris to set them. If not set then assume SSL clients will verify SSL servers 845859191Skris and vice versa. 845959191Skris 846059191Skris Two new options to the verify program: -untrusted allows a set of 846159191Skris untrusted certificates to be passed in and -purpose which sets the 846259191Skris intended purpose of the certificate. If a purpose is set then the 846359191Skris new chain verify code is used to check extension consistency. 846459191Skris [Steve Henson] 846559191Skris 846659191Skris *) Support for the authority information access extension. 846759191Skris [Steve Henson] 846859191Skris 846959191Skris *) Modify RSA and DSA PEM read routines to transparently handle 847059191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 847159191Skris public keys in a format compatible with certificate 847259191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 847359191Skris functions called *_PublicKey_* which used various odd formats so 847459191Skris these are retained for compatibility: however the DSA variants were 847559191Skris never in a public release so they have been deleted. Changed dsa/rsa 847659191Skris utilities to handle the new format: note no releases ever handled public 847759191Skris keys so we should be OK. 847859191Skris 847959191Skris The primary motivation for this change is to avoid the same fiasco 848059191Skris that dogs private keys: there are several incompatible private key 848159191Skris formats some of which are standard and some OpenSSL specific and 848259191Skris require various evil hacks to allow partial transparent handling and 848359191Skris even then it doesn't work with DER formats. Given the option anything 848459191Skris other than PKCS#8 should be dumped: but the other formats have to 848559191Skris stay in the name of compatibility. 848659191Skris 848759191Skris With public keys and the benefit of hindsight one standard format 848859191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 848959191Skris it clearly returns an error if you try to read the wrong kind of key. 849059191Skris 849159191Skris Added a -pubkey option to the 'x509' utility to output the public key. 849259191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 849359191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 849459191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 849559191Skris that do the same as the EVP_PKEY_assign_*() except they up the 849659191Skris reference count of the added key (they don't "swallow" the 849759191Skris supplied key). 849859191Skris [Steve Henson] 849959191Skris 850059191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 850159191Skris CRLs would fail if the file contained no certificates or no CRLs: 850259191Skris added a new function to read in both types and return the number 850359191Skris read: this means that if none are read it will be an error. The 850459191Skris DER versions of the certificate and CRL reader would always fail 850559191Skris because it isn't possible to mix certificates and CRLs in DER format 850659191Skris without choking one or the other routine. Changed this to just read 850759191Skris a certificate: this is the best we can do. Also modified the code 850859191Skris in apps/verify.c to take notice of return codes: it was previously 850959191Skris attempting to read in certificates from NULL pointers and ignoring 851059191Skris any errors: this is one reason why the cert and CRL reader seemed 851159191Skris to work. It doesn't check return codes from the default certificate 851259191Skris routines: these may well fail if the certificates aren't installed. 851359191Skris [Steve Henson] 851459191Skris 851559191Skris *) Code to support otherName option in GeneralName. 851659191Skris [Steve Henson] 851759191Skris 851859191Skris *) First update to verify code. Change the verify utility 851959191Skris so it warns if it is passed a self signed certificate: 852059191Skris for consistency with the normal behaviour. X509_verify 852159191Skris has been modified to it will now verify a self signed 852259191Skris certificate if *exactly* the same certificate appears 852359191Skris in the store: it was previously impossible to trust a 852459191Skris single self signed certificate. This means that: 852559191Skris openssl verify ss.pem 852659191Skris now gives a warning about a self signed certificate but 852759191Skris openssl verify -CAfile ss.pem ss.pem 852859191Skris is OK. 852959191Skris [Steve Henson] 853059191Skris 853159191Skris *) For servers, store verify_result in SSL_SESSION data structure 853259191Skris (and add it to external session representation). 853359191Skris This is needed when client certificate verifications fails, 853459191Skris but an application-provided verification callback (set by 853559191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 853659191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 853759191Skris but returns 1): When the session is reused, we have to set 853859191Skris ssl->verify_result to the appropriate error code to avoid 853959191Skris security holes. 854059191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 854159191Skris 854259191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 854359191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 854459191Skris didn't contain any existing data because it was being created. 854559191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 854659191Skris 854759191Skris *) Add a salt to the key derivation routines in enc.c. This 854859191Skris forms the first 8 bytes of the encrypted file. Also add a 854959191Skris -S option to allow a salt to be input on the command line. 855059191Skris [Steve Henson] 855159191Skris 855259191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 855359191Skris to compare two certificates. We do this by working out the SHA1 855459191Skris hash and comparing that. X509_cmp() will be needed by the trust 855559191Skris code. 855659191Skris [Steve Henson] 855759191Skris 855859191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 855959191Skris the reference count in the SSL_SESSION returned. 856059191Skris [Geoff Thorpe <geoff@eu.c2.net>] 856159191Skris 856259191Skris *) Fix for 'req': it was adding a null to request attributes. 856359191Skris Also change the X509_LOOKUP and X509_INFO code to handle 856459191Skris certificate auxiliary information. 856559191Skris [Steve Henson] 856659191Skris 856759191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 856859191Skris the 'enc' command. 856959191Skris [Steve Henson] 857059191Skris 857159191Skris *) Add the possibility to add extra information to the memory leak 857259191Skris detecting output, to form tracebacks, showing from where each 857359191Skris allocation was originated: CRYPTO_push_info("constant string") adds 857459191Skris the string plus current file name and line number to a per-thread 857559191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 857659191Skris is like calling CYRPTO_pop_info() until the stack is empty. 857759191Skris Also updated memory leak detection code to be multi-thread-safe. 857859191Skris [Richard Levitte] 857959191Skris 858059191Skris *) Add options -text and -noout to pkcs7 utility and delete the 858159191Skris encryption options which never did anything. Update docs. 858259191Skris [Steve Henson] 858359191Skris 858459191Skris *) Add options to some of the utilities to allow the pass phrase 858559191Skris to be included on either the command line (not recommended on 858659191Skris OSes like Unix) or read from the environment. Update the 858759191Skris manpages and fix a few bugs. 858859191Skris [Steve Henson] 858959191Skris 859059191Skris *) Add a few manpages for some of the openssl commands. 859159191Skris [Steve Henson] 859259191Skris 859359191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 859459191Skris leaking and not finding already revoked certificates. 859559191Skris [Steve Henson] 859659191Skris 859759191Skris *) Extensive changes to support certificate auxiliary information. 859859191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 859959191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 860059191Skris can still read in a certificate file in the usual way but it 860159191Skris will also read in any additional "auxiliary information". By 860259191Skris doing things this way a fair degree of compatibility can be 860359191Skris retained: existing certificates can have this information added 860459191Skris using the new 'x509' options. 860559191Skris 860659191Skris Current auxiliary information includes an "alias" and some trust 860759191Skris settings. The trust settings will ultimately be used in enhanced 860859191Skris certificate chain verification routines: currently a certificate 860959191Skris can only be trusted if it is self signed and then it is trusted 861059191Skris for all purposes. 861159191Skris [Steve Henson] 861259191Skris 861359191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 861459191Skris The problem was that one of the replacement routines had not been working 861559191Skris since SSLeay releases. For now the offending routine has been replaced 861659191Skris with non-optimised assembler. Even so, this now gives around 95% 861759191Skris performance improvement for 1024 bit RSA signs. 861859191Skris [Mark Cox] 861959191Skris 862059191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 862159191Skris handling. Most clients have the effective key size in bits equal to 862259191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 862359191Skris A few however don't do this and instead use the size of the decrypted key 862459191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 862559191Skris the effective key length. In this case the effective key length can still 862659191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 862759191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 862859191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 862959191Skris the key length and effective key length are equal. 863059191Skris [Steve Henson] 863159191Skris 863259191Skris *) Add a bunch of functions that should simplify the creation of 863359191Skris X509_NAME structures. Now you should be able to do: 863459191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 863559191Skris and have it automatically work out the correct field type and fill in 863659191Skris the structures. The more adventurous can try: 863759191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 863859191Skris and it will (hopefully) work out the correct multibyte encoding. 863959191Skris [Steve Henson] 864059191Skris 864159191Skris *) Change the 'req' utility to use the new field handling and multibyte 864259191Skris copy routines. Before the DN field creation was handled in an ad hoc 864359191Skris way in req, ca, and x509 which was rather broken and didn't support 864459191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 864559191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 864659191Skris using the dirstring_type option. See the new comment in the default 864759191Skris openssl.cnf for more info. 864859191Skris [Steve Henson] 864959191Skris 865059191Skris *) Make crypto/rand/md_rand.c more robust: 865159191Skris - Assure unique random numbers after fork(). 865259191Skris - Make sure that concurrent threads access the global counter and 865359191Skris md serializably so that we never lose entropy in them 865459191Skris or use exactly the same state in multiple threads. 865559191Skris Access to the large state is not always serializable because 865659191Skris the additional locking could be a performance killer, and 865759191Skris md should be large enough anyway. 865859191Skris [Bodo Moeller] 865959191Skris 866059191Skris *) New file apps/app_rand.c with commonly needed functionality 866159191Skris for handling the random seed file. 866259191Skris 866359191Skris Use the random seed file in some applications that previously did not: 866459191Skris ca, 866559191Skris dsaparam -genkey (which also ignored its '-rand' option), 866659191Skris s_client, 866759191Skris s_server, 866859191Skris x509 (when signing). 866959191Skris Except on systems with /dev/urandom, it is crucial to have a random 867059191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 867159191Skris for RSA signatures we could do without one. 867259191Skris 867359191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 867459191Skris of each file listed in the '-rand' option. The function as previously 867559191Skris found in genrsa is now in app_rand.c and is used by all programs 867659191Skris that support '-rand'. 867759191Skris [Bodo Moeller] 867859191Skris 867959191Skris *) In RAND_write_file, use mode 0600 for creating files; 868059191Skris don't just chmod when it may be too late. 868159191Skris [Bodo Moeller] 868259191Skris 868359191Skris *) Report an error from X509_STORE_load_locations 868459191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 868559191Skris [Bill Perry] 868659191Skris 868759191Skris *) New function ASN1_mbstring_copy() this copies a string in either 868859191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 868959191Skris into an ASN1_STRING type. A mask of permissible types is passed 869059191Skris and it chooses the "minimal" type to use or an error if not type 869159191Skris is suitable. 869259191Skris [Steve Henson] 869359191Skris 869459191Skris *) Add function equivalents to the various macros in asn1.h. The old 869559191Skris macros are retained with an M_ prefix. Code inside the library can 869659191Skris use the M_ macros. External code (including the openssl utility) 869759191Skris should *NOT* in order to be "shared library friendly". 869859191Skris [Steve Henson] 869959191Skris 870059191Skris *) Add various functions that can check a certificate's extensions 870159191Skris to see if it usable for various purposes such as SSL client, 870259191Skris server or S/MIME and CAs of these types. This is currently 870359191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 870459191Skris verification. Also added a -purpose flag to x509 utility to 870559191Skris print out all the purposes. 870659191Skris [Steve Henson] 870759191Skris 870859191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 870959191Skris functions. 871059191Skris [Steve Henson] 871159191Skris 871259191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 871359191Skris for, obtain and decode and extension and obtain its critical flag. 871459191Skris This allows all the necessary extension code to be handled in a 871559191Skris single function call. 871659191Skris [Steve Henson] 871759191Skris 871859191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 871959191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 872059191Skris [Andy Polyakov] 872159191Skris 872259191Skris *) New -noout option to asn1parse. This causes no output to be produced 872359191Skris its main use is when combined with -strparse and -out to extract data 872459191Skris from a file (which may not be in ASN.1 format). 872559191Skris [Steve Henson] 872659191Skris 872759191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 872859191Skris when producing the local key id. 872959191Skris [Richard Levitte <levitte@stacken.kth.se>] 873059191Skris 873159191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 873259191Skris stated explicitly. If it is not stated then it tries the first server 873359191Skris certificate file. The previous behaviour hard coded the filename 873459191Skris "server.pem". 873559191Skris [Steve Henson] 873659191Skris 873759191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 873859191Skris a public key to be input or output. For example: 873959191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 874059191Skris Also added necessary DSA public key functions to handle this. 874159191Skris [Steve Henson] 874259191Skris 874359191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 874459191Skris in the message. This was handled by allowing 874559191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 874659191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 874759191Skris 874859191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 874959191Skris to the end of the strings whereas this didn't. This would cause problems 875059191Skris if strings read with d2i_ASN1_bytes() were later modified. 875159191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 875259191Skris 875359191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 875459191Skris data and it contains EOF it will end up returning an error. This is 875559191Skris caused by input 46 bytes long. The cause is due to the way base64 875659191Skris BIOs find the start of base64 encoded data. They do this by trying a 875759191Skris trial decode on each line until they find one that works. When they 875859191Skris do a flag is set and it starts again knowing it can pass all the 875959191Skris data directly through the decoder. Unfortunately it doesn't reset 876059191Skris the context it uses. This means that if EOF is reached an attempt 876159191Skris is made to pass two EOFs through the context and this causes the 876259191Skris resulting error. This can also cause other problems as well. As is 876359191Skris usual with these problems it takes *ages* to find and the fix is 876459191Skris trivial: move one line. 876559191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 876659191Skris 876759191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 876859191Skris old code wouldn't work because it needed to select() on sockets and the 876959191Skris tty (for keypresses and to see if data could be written). Win32 only 877059191Skris supports select() on sockets so we select() with a 1s timeout on the 877159191Skris sockets and then see if any characters are waiting to be read, if none 877259191Skris are present then we retry, we also assume we can always write data to 877359191Skris the tty. This isn't nice because the code then blocks until we've 877459191Skris received a complete line of data and it is effectively polling the 877559191Skris keyboard at 1s intervals: however it's quite a bit better than not 877659191Skris working at all :-) A dedicated Windows application might handle this 877759191Skris with an event loop for example. 877859191Skris [Steve Henson] 877959191Skris 878059191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 878159191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 878259191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 878359191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 878459191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 878559191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 878659191Skris This necessitated the support of an extra signature type NID_md5_sha1 878759191Skris for SSL signatures and modifications to the SSL library to use it instead 878859191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 878959191Skris [Steve Henson] 879059191Skris 879159191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 879259191Skris will lookup a CRL issuers certificate and verify the signature in a 879359191Skris similar way to the verify program. Tidy up the crl program so it 879459191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 879559191Skris less strict. It will now permit CRL extensions even if it is not 879659191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 879759191Skris [Steve Henson] 879859191Skris 879959191Skris *) Initialize all non-automatic variables each time one of the openssl 880059191Skris sub-programs is started (this is necessary as they may be started 880159191Skris multiple times from the "OpenSSL>" prompt). 880259191Skris [Lennart Bang, Bodo Moeller] 880359191Skris 880459191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 880559191Skris removing all other RSA functionality (this is what NO_RSA does). This 880659191Skris is so (for example) those in the US can disable those operations covered 880759191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 880859191Skris key generation. 880959191Skris [Steve Henson] 881059191Skris 881159191Skris *) Non-copying interface to BIO pairs. 881259191Skris (still largely untested) 881359191Skris [Bodo Moeller] 881459191Skris 881559191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 881659191Skris ASCII string. This was handled independently in various places before. 881759191Skris [Steve Henson] 881859191Skris 881959191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 882059191Skris UTF8 strings a character at a time. 882159191Skris [Steve Henson] 882259191Skris 882359191Skris *) Use client_version from client hello to select the protocol 882459191Skris (s23_srvr.c) and for RSA client key exchange verification 882559191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 882659191Skris [Bodo Moeller] 882759191Skris 882859191Skris *) Add various utility functions to handle SPKACs, these were previously 882959191Skris handled by poking round in the structure internals. Added new function 883059191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 883159191Skris print, verify and generate SPKACs. Based on an original idea from 883259191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 883359191Skris [Steve Henson] 883459191Skris 883559191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 883659191Skris [Andy Polyakov] 883759191Skris 883859191Skris *) Allow the config file extension section to be overwritten on the 883959191Skris command line. Based on an original idea from Massimiliano Pala 884059191Skris <madwolf@comune.modena.it>. The new option is called -extensions 884159191Skris and can be applied to ca, req and x509. Also -reqexts to override 884259191Skris the request extensions in req and -crlexts to override the crl extensions 884359191Skris in ca. 884459191Skris [Steve Henson] 884559191Skris 884659191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 884759191Skris the same field multiple times by preceding it by "XXXX." for example: 884859191Skris 1.OU="Unit name 1" 884959191Skris 2.OU="Unit name 2" 885059191Skris this is the same syntax as used in the req config file. 885159191Skris [Steve Henson] 885259191Skris 885359191Skris *) Allow certificate extensions to be added to certificate requests. These 885459191Skris are specified in a 'req_extensions' option of the req section of the 885559191Skris config file. They can be printed out with the -text option to req but 885659191Skris are otherwise ignored at present. 885759191Skris [Steve Henson] 885859191Skris 885959191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 886059191Skris data read consists of only the final block it would not decrypted because 886159191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 886259191Skris A misplaced 'break' also meant the decrypted final block might not be 886359191Skris copied until the next read. 886459191Skris [Steve Henson] 886559191Skris 886659191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 886759191Skris a few extra parameters to the DH structure: these will be useful if 886859191Skris for example we want the value of 'q' or implement X9.42 DH. 886959191Skris [Steve Henson] 887059191Skris 887159191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 887259191Skris provides hooks that allow the default DSA functions or functions on a 887359191Skris "per key" basis to be replaced. This allows hardware acceleration and 887459191Skris hardware key storage to be handled without major modification to the 887559191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 887659191Skris associated functions. 887759191Skris [Steve Henson] 887859191Skris 887959191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 888059191Skris as "read only": it can't be written to and the buffer it points to will 888159191Skris not be freed. Reading from a read only BIO is much more efficient than 888259191Skris a normal memory BIO. This was added because there are several times when 888359191Skris an area of memory needs to be read from a BIO. The previous method was 888459191Skris to create a memory BIO and write the data to it, this results in two 888559191Skris copies of the data and an O(n^2) reading algorithm. There is a new 888659191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 888759191Skris an area of memory. Also modified the PKCS#7 routines to use read only 888859191Skris memory BIOs. 888959191Skris [Steve Henson] 889059191Skris 889159191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 889259191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 889359191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 889459191Skris but a retry condition occured while trying to read the rest. 889559191Skris [Bodo Moeller] 889659191Skris 889759191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 889859191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 889959191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 890059191Skris the encrypted data type: this is a more sensible place to put it and it 890159191Skris allows the PKCS#12 code to be tidied up that duplicated this 890259191Skris functionality. 890359191Skris [Steve Henson] 890459191Skris 890559191Skris *) Changed obj_dat.pl script so it takes its input and output files on 890659191Skris the command line. This should avoid shell escape redirection problems 890759191Skris under Win32. 890859191Skris [Steve Henson] 890959191Skris 891059191Skris *) Initial support for certificate extension requests, these are included 891159191Skris in things like Xenroll certificate requests. Included functions to allow 891259191Skris extensions to be obtained and added. 891359191Skris [Steve Henson] 891459191Skris 891559191Skris *) -crlf option to s_client and s_server for sending newlines as 891659191Skris CRLF (as required by many protocols). 891759191Skris [Bodo Moeller] 891859191Skris 891955714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 892055714Skris 892155714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 892255714Skris [Ralf S. Engelschall] 892355714Skris 892455714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 892555714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 892655714Skris 892755714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 892855714Skris program. 892955714Skris [Steve Henson] 893055714Skris 893155714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 893255714Skris DH parameters/keys (q is lost during that conversion, but the resulting 893355714Skris DH parameters contain its length). 893455714Skris 893555714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 893655714Skris much faster than DH_generate_parameters (which creates parameters 893755714Skris where p = 2*q + 1), and also the smaller q makes DH computations 893855714Skris much more efficient (160-bit exponentiation instead of 1024-bit 893955714Skris exponentiation); so this provides a convenient way to support DHE 894055714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 894155714Skris utter importance to use 894255714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 894355714Skris or 894455714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 894555714Skris when such DH parameters are used, because otherwise small subgroup 894655714Skris attacks may become possible! 894755714Skris [Bodo Moeller] 894855714Skris 894955714Skris *) Avoid memory leak in i2d_DHparams. 895055714Skris [Bodo Moeller] 895155714Skris 895255714Skris *) Allow the -k option to be used more than once in the enc program: 895355714Skris this allows the same encrypted message to be read by multiple recipients. 895455714Skris [Steve Henson] 895555714Skris 895655714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 895755714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 895855714Skris it will always use the numerical form of the OID, even if it has a short 895955714Skris or long name. 896055714Skris [Steve Henson] 896155714Skris 896255714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 896355714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 896455714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 896555714Skris no private key components need be present and it might store extra data 896659191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 896759191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 896859191Skris private key operations. 896955714Skris [Steve Henson] 897055714Skris 897155714Skris *) Added support for SPARC Linux. 897255714Skris [Andy Polyakov] 897355714Skris 897455714Skris *) pem_password_cb function type incompatibly changed from 897555714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 897655714Skris to 897755714Skris ....(char *buf, int size, int rwflag, void *userdata); 897855714Skris so that applications can pass data to their callbacks: 897955714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 898055714Skris additional void * argument, which is just handed through whenever 898155714Skris the password callback is called. 898259191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 898355714Skris 898455714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 898555714Skris 898655714Skris Compatibility note: As many C implementations push function arguments 898755714Skris onto the stack in reverse order, the new library version is likely to 898855714Skris interoperate with programs that have been compiled with the old 898955714Skris pem_password_cb definition (PEM_whatever takes some data that 899055714Skris happens to be on the stack as its last argument, and the callback 899155714Skris just ignores this garbage); but there is no guarantee whatsoever that 899255714Skris this will work. 899355714Skris 899455714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 899555714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 899655714Skris problems not only on Windows, but also on some Unix platforms. 899755714Skris To avoid problematic command lines, these definitions are now in an 899855714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 899955714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 900055714Skris [Bodo Moeller] 900155714Skris 900255714Skris *) MIPS III/IV assembler module is reimplemented. 900355714Skris [Andy Polyakov] 900455714Skris 900555714Skris *) More DES library cleanups: remove references to srand/rand and 900655714Skris delete an unused file. 900755714Skris [Ulf M�ller] 900855714Skris 900955714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 901055714Skris since not many people have MASM (ml) and it can be hard to obtain. 901155714Skris This is currently experimental but it seems to work OK and pass all 901255714Skris the tests. Check out INSTALL.W32 for info. 901355714Skris [Steve Henson] 901455714Skris 901555714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 901655714Skris without temporary keys kept an extra copy of the server key, 901755714Skris and connections with temporary keys did not free everything in case 901855714Skris of an error. 901955714Skris [Bodo Moeller] 902055714Skris 902155714Skris *) New function RSA_check_key and new openssl rsa option -check 902255714Skris for verifying the consistency of RSA keys. 902355714Skris [Ulf Moeller, Bodo Moeller] 902455714Skris 902555714Skris *) Various changes to make Win32 compile work: 902655714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 902755714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 902855714Skris comparison" warnings. 902955714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 903055714Skris [Steve Henson] 903155714Skris 903255714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 903355714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 903455714Skris derived keys are printed to stderr. 903555714Skris [Steve Henson] 903655714Skris 903755714Skris *) Copy the flags in ASN1_STRING_dup(). 903855714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 903955714Skris 904055714Skris *) The x509 application mishandled signing requests containing DSA 904155714Skris keys when the signing key was also DSA and the parameters didn't match. 904255714Skris 904355714Skris It was supposed to omit the parameters when they matched the signing key: 904455714Skris the verifying software was then supposed to automatically use the CA's 904555714Skris parameters if they were absent from the end user certificate. 904655714Skris 904755714Skris Omitting parameters is no longer recommended. The test was also 904855714Skris the wrong way round! This was probably due to unusual behaviour in 904955714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 905055714Skris This meant that parameters were omitted when they *didn't* match and 905155714Skris the certificate was useless. Certificates signed with 'ca' didn't have 905255714Skris this bug. 905355714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 905455714Skris 905555714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 905655714Skris The interface is as follows: 905755714Skris Applications can use 905855714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 905955714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 906055714Skris "off" is now the default. 906155714Skris The library internally uses 906255714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 906355714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 906455714Skris to disable memory-checking temporarily. 906555714Skris 906655714Skris Some inconsistent states that previously were possible (and were 906755714Skris even the default) are now avoided. 906855714Skris 906955714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 907055714Skris with each memory chunk allocated; this is occasionally more helpful 907155714Skris than just having a counter. 907255714Skris 907355714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 907455714Skris 907555714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 907655714Skris extensions. 907755714Skris [Bodo Moeller] 907855714Skris 907955714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 908055714Skris which largely parallels "options", but is for changing API behaviour, 908155714Skris whereas "options" are about protocol behaviour. 908255714Skris Initial "mode" flags are: 908355714Skris 908455714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 908555714Skris a single record has been written. 908655714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 908755714Skris retries use the same buffer location. 908855714Skris (But all of the contents must be 908955714Skris copied!) 909055714Skris [Bodo Moeller] 909155714Skris 909279998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 909355714Skris worked. 909455714Skris 909555714Skris *) Fix problems with no-hmac etc. 909655714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 909755714Skris 909855714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 909955714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 910055714Skris to mess around with the internals of an RSA structure. 910155714Skris [Steve Henson] 910255714Skris 910355714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 910455714Skris Also really enable memory leak checks in openssl.c and in some 910555714Skris test programs. 910655714Skris [Chad C. Mulligan, Bodo Moeller] 910755714Skris 910855714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 910955714Skris up the length of negative integers. This has now been simplified to just 911055714Skris store the length when it is first determined and use it later, rather 911155714Skris than trying to keep track of where data is copied and updating it to 911255714Skris point to the end. 911355714Skris [Steve Henson, reported by Brien Wheeler 911455714Skris <bwheeler@authentica-security.com>] 911555714Skris 911655714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 911755714Skris of a PKCS#7 signature but with the signing certificate passed to the 911855714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 911955714Skris certificate is present in the PKCS#7 structure. This isn't always the 912055714Skris case: certificates can be omitted from a PKCS#7 structure and be 912155714Skris distributed by "out of band" means (such as a certificate database). 912255714Skris [Steve Henson] 912355714Skris 912455714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 912555714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 912655714Skris necessary function names. 912755714Skris [Steve Henson] 912855714Skris 912955714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 913055714Skris options set by Configure in the top level Makefile, and Configure 913155714Skris was not even able to write more than one option correctly. 913255714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 913355714Skris [Bodo Moeller] 913455714Skris 913555714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 913655714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 913755714Skris for example allow memory BIOs to contain config info. 913855714Skris [Steve Henson] 913955714Skris 914055714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 914155714Skris Whoever hopes to achieve shared-library compatibility across versions 914255714Skris must use this, not the compile-time macro. 914355714Skris (Exercise 0.9.4: Which is the minimum library version required by 914455714Skris such programs?) 914555714Skris Note: All this applies only to multi-threaded programs, others don't 914655714Skris need locks. 914755714Skris [Bodo Moeller] 914855714Skris 914955714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 915055714Skris through a BIO pair triggered the default case, i.e. 915155714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 915255714Skris [Bodo Moeller] 915355714Skris 915455714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 915555714Skris can use the SSL library even if none of the specific BIOs is 915655714Skris appropriate. 915755714Skris [Bodo Moeller] 915855714Skris 915955714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 916055714Skris for the encoded length. 916155714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 916255714Skris 916355714Skris *) Add initial documentation of the X509V3 functions. 916455714Skris [Steve Henson] 916555714Skris 916655714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 916755714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 916855714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 916955714Skris secure PKCS#8 private key format with a high iteration count. 917055714Skris [Steve Henson] 917155714Skris 917255714Skris *) Fix determination of Perl interpreter: A perl or perl5 917355714Skris _directory_ in $PATH was also accepted as the interpreter. 917455714Skris [Ralf S. Engelschall] 917555714Skris 917655714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 917755714Skris wrong with it but it was very old and did things like calling 917855714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 917955714Skris unusual formatting. 918055714Skris [Steve Henson] 918155714Skris 918255714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 918355714Skris to use the new extension code. 918455714Skris [Steve Henson] 918555714Skris 918655714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 918755714Skris with macros. This should make it easier to change their form, add extra 918855714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 918955714Skris constant. 919055714Skris [Steve Henson] 919155714Skris 919255714Skris *) Add to configuration table a new entry that can specify an alternative 919355714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 919455714Skris according to Mark Crispin <MRC@Panda.COM>. 919555714Skris [Bodo Moeller] 919655714Skris 919755714Skris#if 0 919855714Skris *) DES CBC did not update the IV. Weird. 919955714Skris [Ben Laurie] 920055714Skris#else 920155714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 920255714Skris Changing the behaviour of the former might break existing programs -- 920355714Skris where IV updating is needed, des_ncbc_encrypt can be used. 920455714Skris#endif 920555714Skris 920655714Skris *) When bntest is run from "make test" it drives bc to check its 920755714Skris calculations, as well as internally checking them. If an internal check 920855714Skris fails, it needs to cause bc to give a non-zero result or make test carries 920955714Skris on without noticing the failure. Fixed. 921055714Skris [Ben Laurie] 921155714Skris 921255714Skris *) DES library cleanups. 921355714Skris [Ulf M�ller] 921455714Skris 921555714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 921655714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 921755714Skris ciphers. NOTE: although the key derivation function has been verified 921855714Skris against some published test vectors it has not been extensively tested 921955714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 922055714Skris of v2.0. 922155714Skris [Steve Henson] 922255714Skris 922355714Skris *) Instead of "mkdir -p", which is not fully portable, use new 922455714Skris Perl script "util/mkdir-p.pl". 922555714Skris [Bodo Moeller] 922655714Skris 922755714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 922855714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 922955714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 923055714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 923155714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 923255714Skris underlying key generation function so it must do its own ASN1 parsing. 923355714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 923455714Skris 'parameter' argument instead of literal salt and iteration count values 923555714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 923655714Skris [Steve Henson] 923755714Skris 923855714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 923955714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 924055714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 924155714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 924255714Skris value was just used as a "magic string" and not used directly its 924355714Skris value doesn't matter. 924455714Skris [Steve Henson] 924555714Skris 924655714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 924755714Skris support mutable. 924855714Skris [Ben Laurie] 924955714Skris 925055714Skris *) "linux-sparc64" configuration (ultrapenguin). 925155714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 925255714Skris "linux-sparc" configuration. 925355714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 925455714Skris 925555714Skris *) config now generates no-xxx options for missing ciphers. 925655714Skris [Ulf M�ller] 925755714Skris 925855714Skris *) Support the EBCDIC character set (work in progress). 925955714Skris File ebcdic.c not yet included because it has a different license. 926055714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 926155714Skris 926255714Skris *) Support BS2000/OSD-POSIX. 926355714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 926455714Skris 926555714Skris *) Make callbacks for key generation use void * instead of char *. 926655714Skris [Ben Laurie] 926755714Skris 926855714Skris *) Make S/MIME samples compile (not yet tested). 926955714Skris [Ben Laurie] 927055714Skris 927155714Skris *) Additional typesafe stacks. 927255714Skris [Ben Laurie] 927355714Skris 927455714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 927555714Skris [Bodo Moeller] 927655714Skris 927755714Skris 927855714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 927955714Skris 928055714Skris *) New configuration variant "sco5-gcc". 928155714Skris 928255714Skris *) Updated some demos. 928355714Skris [Sean O Riordain, Wade Scholine] 928455714Skris 928555714Skris *) Add missing BIO_free at exit of pkcs12 application. 928655714Skris [Wu Zhigang] 928755714Skris 928855714Skris *) Fix memory leak in conf.c. 928955714Skris [Steve Henson] 929055714Skris 929155714Skris *) Updates for Win32 to assembler version of MD5. 929255714Skris [Steve Henson] 929355714Skris 929455714Skris *) Set #! path to perl in apps/der_chop to where we found it 929555714Skris instead of using a fixed path. 929655714Skris [Bodo Moeller] 929755714Skris 929855714Skris *) SHA library changes for irix64-mips4-cc. 929955714Skris [Andy Polyakov] 930055714Skris 930155714Skris *) Improvements for VMS support. 930255714Skris [Richard Levitte] 930355714Skris 930455714Skris 930555714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 930655714Skris 930755714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 930855714Skris This also avoids the problems with SC4.2 and unpatched SC5. 930955714Skris [Andy Polyakov <appro@fy.chalmers.se>] 931055714Skris 931155714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 931255714Skris These are required because of the typesafe stack would otherwise break 931355714Skris existing code. If old code used a structure member which used to be STACK 931455714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 931555714Skris sk_num or sk_value it would produce an error because the num, data members 931655714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 931755714Skris replaces the old method of assigning a value to sk_value 931855714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 931955714Skris that does this will no longer work (and should use sk_set instead) but 932055714Skris this could be regarded as a "questionable" behaviour anyway. 932155714Skris [Steve Henson] 932255714Skris 932355714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 932455714Skris correctly handle encrypted S/MIME data. 932555714Skris [Steve Henson] 932655714Skris 932755714Skris *) Change type of various DES function arguments from des_cblock 932855714Skris (which means, in function argument declarations, pointer to char) 932955714Skris to des_cblock * (meaning pointer to array with 8 char elements), 933055714Skris which allows the compiler to do more typechecking; it was like 933155714Skris that back in SSLeay, but with lots of ugly casts. 933255714Skris 933355714Skris Introduce new type const_des_cblock. 933455714Skris [Bodo Moeller] 933555714Skris 933655714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 933755714Skris problems: find RecipientInfo structure that matches recipient certificate 933855714Skris and initialise the ASN1 structures properly based on passed cipher. 933955714Skris [Steve Henson] 934055714Skris 934155714Skris *) Belatedly make the BN tests actually check the results. 934255714Skris [Ben Laurie] 934355714Skris 934455714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 934555714Skris to and from BNs: it was completely broken. New compilation option 934655714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 934755714Skris key elements as negative integers. 934855714Skris [Steve Henson] 934955714Skris 935055714Skris *) Reorganize and speed up MD5. 935155714Skris [Andy Polyakov <appro@fy.chalmers.se>] 935255714Skris 935355714Skris *) VMS support. 935455714Skris [Richard Levitte <richard@levitte.org>] 935555714Skris 935655714Skris *) New option -out to asn1parse to allow the parsed structure to be 935755714Skris output to a file. This is most useful when combined with the -strparse 935855714Skris option to examine the output of things like OCTET STRINGS. 935955714Skris [Steve Henson] 936055714Skris 936155714Skris *) Make SSL library a little more fool-proof by not requiring any longer 936255714Skris that SSL_set_{accept,connect}_state be called before 936355714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 936455714Skris in many applications because usually everything *appeared* to work as 936555714Skris intended anyway -- now it really works as intended). 936655714Skris [Bodo Moeller] 936755714Skris 936855714Skris *) Move openssl.cnf out of lib/. 936955714Skris [Ulf M�ller] 937055714Skris 937155714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 937255714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 937355714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 937455714Skris [Ralf S. Engelschall] 937555714Skris 937655714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 937755714Skris handle PKCS#7 enveloped data properly. 937855714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 937955714Skris 938055714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 938155714Skris copying pointers. The cert_st handling is changed by this in 938255714Skris various ways (and thus what used to be known as ctx->default_cert 938355714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 938455714Skris any longer when s->cert does not give us what we need). 938555714Skris ssl_cert_instantiate becomes obsolete by this change. 938655714Skris As soon as we've got the new code right (possibly it already is?), 938755714Skris we have solved a couple of bugs of the earlier code where s->cert 938855714Skris was used as if it could not have been shared with other SSL structures. 938955714Skris 939055714Skris Note that using the SSL API in certain dirty ways now will result 939155714Skris in different behaviour than observed with earlier library versions: 939255714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 939355714Skris does not influence s as it used to. 939455714Skris 939555714Skris In order to clean up things more thoroughly, inside SSL_SESSION 939655714Skris we don't use CERT any longer, but a new structure SESS_CERT 939755714Skris that holds per-session data (if available); currently, this is 939855714Skris the peer's certificate chain and, for clients, the server's certificate 939955714Skris and temporary key. CERT holds only those values that can have 940055714Skris meaningful defaults in an SSL_CTX. 940155714Skris [Bodo Moeller] 940255714Skris 940355714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 940455714Skris from the internal representation. Various PKCS#7 fixes: remove some 940555714Skris evil casts and set the enc_dig_alg field properly based on the signing 940655714Skris key type. 940755714Skris [Steve Henson] 940855714Skris 940955714Skris *) Allow PKCS#12 password to be set from the command line or the 941055714Skris environment. Let 'ca' get its config file name from the environment 941155714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 941255714Skris and 'x509'). 941355714Skris [Steve Henson] 941455714Skris 941555714Skris *) Allow certificate policies extension to use an IA5STRING for the 941655714Skris organization field. This is contrary to the PKIX definition but 941755714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 941855714Skris extension option. 941955714Skris [Steve Henson] 942055714Skris 942155714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 942255714Skris without disallowing inline assembler and the like for non-pedantic builds. 942355714Skris [Ben Laurie] 942455714Skris 942555714Skris *) Support Borland C++ builder. 942655714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 942755714Skris 942855714Skris *) Support Mingw32. 942955714Skris [Ulf M�ller] 943055714Skris 943155714Skris *) SHA-1 cleanups and performance enhancements. 943255714Skris [Andy Polyakov <appro@fy.chalmers.se>] 943355714Skris 943455714Skris *) Sparc v8plus assembler for the bignum library. 943555714Skris [Andy Polyakov <appro@fy.chalmers.se>] 943655714Skris 943755714Skris *) Accept any -xxx and +xxx compiler options in Configure. 943855714Skris [Ulf M�ller] 943955714Skris 944055714Skris *) Update HPUX configuration. 944155714Skris [Anonymous] 944255714Skris 944355714Skris *) Add missing sk_<type>_unshift() function to safestack.h 944455714Skris [Ralf S. Engelschall] 944555714Skris 944655714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 944755714Skris "extra_cert"s in addition to the certificate. (This makes sense 944855714Skris only for "PEM" format files, as chains as a whole are not 944955714Skris DER-encoded.) 945055714Skris [Bodo Moeller] 945155714Skris 945255714Skris *) Support verify_depth from the SSL API. 945355714Skris x509_vfy.c had what can be considered an off-by-one-error: 945455714Skris Its depth (which was not part of the external interface) 945555714Skris was actually counting the number of certificates in a chain; 945655714Skris now it really counts the depth. 945755714Skris [Bodo Moeller] 945855714Skris 945955714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 946055714Skris instead of X509err, which often resulted in confusing error 946155714Skris messages since the error codes are not globally unique 946255714Skris (e.g. an alleged error in ssl3_accept when a certificate 946355714Skris didn't match the private key). 946455714Skris 946555714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 946655714Skris value (so that you don't need SSL_set_session_id_context for each 946755714Skris connection using the SSL_CTX). 946855714Skris [Bodo Moeller] 946955714Skris 947055714Skris *) OAEP decoding bug fix. 947155714Skris [Ulf M�ller] 947255714Skris 947355714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 947455714Skris David Harris. 947555714Skris [Bodo Moeller] 947655714Skris 947755714Skris *) New Configure options "threads" and "no-threads". For systems 947855714Skris where the proper compiler options are known (currently Solaris 947955714Skris and Linux), "threads" is the default. 948055714Skris [Bodo Moeller] 948155714Skris 948255714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 948355714Skris [Bodo Moeller] 948455714Skris 948555714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 948655714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 948755714Skris such as /usr/local/bin. 948855714Skris [Bodo Moeller] 948955714Skris 949055714Skris *) "make linux-shared" to build shared libraries. 949155714Skris [Niels Poppe <niels@netbox.org>] 949255714Skris 949355714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 949455714Skris [Ulf M�ller] 949555714Skris 949655714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 949755714Skris extension adding in x509 utility. 949855714Skris [Steve Henson] 949955714Skris 950055714Skris *) Remove NOPROTO sections and error code comments. 950155714Skris [Ulf M�ller] 950255714Skris 950355714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 950455714Skris prototypes. 950555714Skris [Steve Henson] 950655714Skris 950755714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 950855714Skris [Ulf M�ller] 950955714Skris 951055714Skris *) Complete rewrite of the error code script(s). It is all now handled 951155714Skris by one script at the top level which handles error code gathering, 951255714Skris header rewriting and C source file generation. It should be much better 951355714Skris than the old method: it now uses a modified version of Ulf's parser to 951455714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 951555714Skris aren't needed for error creation any more) and do a better job of 951655714Skris translating function codes into names. The old 'ASN1 error code imbedded 951755714Skris in a comment' is no longer necessary and it doesn't use .err files which 951855714Skris have now been deleted. Also the error code call doesn't have to appear all 951955714Skris on one line (which resulted in some large lines...). 952055714Skris [Steve Henson] 952155714Skris 952255714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 952355714Skris [Bodo Moeller] 952455714Skris 952555714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 952655714Skris 0 (which usually indicates a closed connection), but continue reading. 952755714Skris [Bodo Moeller] 952855714Skris 952955714Skris *) Fix some race conditions. 953055714Skris [Bodo Moeller] 953155714Skris 953255714Skris *) Add support for CRL distribution points extension. Add Certificate 953355714Skris Policies and CRL distribution points documentation. 953455714Skris [Steve Henson] 953555714Skris 953655714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 953755714Skris [Ulf M�ller] 953855714Skris 953955714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 954055714Skris 8 of keying material. Merlin has also confirmed interop with this fix 954155714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 954255714Skris [Merlin Hughes <merlin@baltimore.ie>] 954355714Skris 954455714Skris *) Fix lots of warnings. 954555714Skris [Richard Levitte <levitte@stacken.kth.se>] 954655714Skris 954755714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 954855714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 954955714Skris [Richard Levitte <levitte@stacken.kth.se>] 955055714Skris 955155714Skris *) Fix problems with sizeof(long) == 8. 955255714Skris [Andy Polyakov <appro@fy.chalmers.se>] 955355714Skris 955455714Skris *) Change functions to ANSI C. 955555714Skris [Ulf M�ller] 955655714Skris 955755714Skris *) Fix typos in error codes. 955855714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 955955714Skris 956055714Skris *) Remove defunct assembler files from Configure. 956155714Skris [Ulf M�ller] 956255714Skris 956355714Skris *) SPARC v8 assembler BIGNUM implementation. 956455714Skris [Andy Polyakov <appro@fy.chalmers.se>] 956555714Skris 956655714Skris *) Support for Certificate Policies extension: both print and set. 956755714Skris Various additions to support the r2i method this uses. 956855714Skris [Steve Henson] 956955714Skris 957055714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 957155714Skris return a const string when you are expecting an allocated buffer. 957255714Skris [Ben Laurie] 957355714Skris 957455714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 957555714Skris types DirectoryString and DisplayText. 957655714Skris [Steve Henson] 957755714Skris 957855714Skris *) Add code to allow r2i extensions to access the configuration database, 957955714Skris add an LHASH database driver and add several ctx helper functions. 958055714Skris [Steve Henson] 958155714Skris 958255714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 958355714Skris fail when they extended the size of a BIGNUM. 958455714Skris [Steve Henson] 958555714Skris 958655714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 958755714Skris support typesafe stack. 958855714Skris [Steve Henson] 958955714Skris 959055714Skris *) Fix typo in SSL_[gs]et_options(). 959155714Skris [Nils Frostberg <nils@medcom.se>] 959255714Skris 959355714Skris *) Delete various functions and files that belonged to the (now obsolete) 959455714Skris old X509V3 handling code. 959555714Skris [Steve Henson] 959655714Skris 959755714Skris *) New Configure option "rsaref". 959855714Skris [Ulf M�ller] 959955714Skris 960055714Skris *) Don't auto-generate pem.h. 960155714Skris [Bodo Moeller] 960255714Skris 960355714Skris *) Introduce type-safe ASN.1 SETs. 960455714Skris [Ben Laurie] 960555714Skris 960655714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 960755714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 960855714Skris 960955714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 961055714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 961155714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 961255714Skris few STACKed things have been converted already. Feel free to convert more. 961355714Skris In the fullness of time, I'll do away with the STACK type altogether. 961455714Skris [Ben Laurie] 961555714Skris 961655714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 961755714Skris specified in <certfile> by updating the entry in the index.txt file. 961855714Skris This way one no longer has to edit the index.txt file manually for 961955714Skris revoking a certificate. The -revoke option does the gory details now. 962055714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 962155714Skris 962255714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 962355714Skris `-text' option at all and this way the `-noout -text' combination was 962455714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 962555714Skris [Ralf S. Engelschall] 962655714Skris 962755714Skris *) Make sure a corresponding plain text error message exists for the 962855714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 962955714Skris verify callback function determined that a certificate was revoked. 963055714Skris [Ralf S. Engelschall] 963155714Skris 963255714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 963355714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 963455714Skris all available cipers including rc5, which was forgotten until now. 963555714Skris In order to let the testing shell script know which algorithms 963655714Skris are available, a new (up to now undocumented) command 963755714Skris "openssl list-cipher-commands" is used. 963855714Skris [Bodo Moeller] 963955714Skris 964055714Skris *) Bugfix: s_client occasionally would sleep in select() when 964155714Skris it should have checked SSL_pending() first. 964255714Skris [Bodo Moeller] 964355714Skris 964455714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 964555714Skris the raw DSA values prior to ASN.1 encoding. 964655714Skris [Ulf M�ller] 964755714Skris 964855714Skris *) Tweaks to Configure 964955714Skris [Niels Poppe <niels@netbox.org>] 965055714Skris 965155714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 965255714Skris yet... 965355714Skris [Steve Henson] 965455714Skris 965555714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 965655714Skris [Ulf M�ller] 965755714Skris 965855714Skris *) New config option to avoid instructions that are illegal on the 80386. 965955714Skris The default code is faster, but requires at least a 486. 966055714Skris [Ulf M�ller] 966155714Skris 966255714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 966355714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 966455714Skris same as SSL2_VERSION anyway. 966555714Skris [Bodo Moeller] 966655714Skris 966755714Skris *) New "-showcerts" option for s_client. 966855714Skris [Bodo Moeller] 966955714Skris 967055714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 967155714Skris application. Various cleanups and fixes. 967255714Skris [Steve Henson] 967355714Skris 967455714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 967555714Skris modify error routines to work internally. Add error codes and PBE init 967655714Skris to library startup routines. 967755714Skris [Steve Henson] 967855714Skris 967955714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 968055714Skris packing functions to asn1 and evp. Changed function names and error 968155714Skris codes along the way. 968255714Skris [Steve Henson] 968355714Skris 968455714Skris *) PKCS12 integration: and so it begins... First of several patches to 968555714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 968655714Skris objects to objects.h 968755714Skris [Steve Henson] 968855714Skris 968955714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 969055714Skris and display support for Thawte strong extranet extension. 969155714Skris [Steve Henson] 969255714Skris 969355714Skris *) Add LinuxPPC support. 969455714Skris [Jeff Dubrule <igor@pobox.org>] 969555714Skris 969655714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 969755714Skris bn_div_words in alpha.s. 969855714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 969955714Skris 970055714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 970155714Skris OAEP isn't supported when OpenSSL is built with RSAref. 970255714Skris [Ulf Moeller <ulf@fitug.de>] 970355714Skris 970455714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 970555714Skris so they no longer are missing under -DNOPROTO. 970655714Skris [Soren S. Jorvang <soren@t.dk>] 970755714Skris 970855714Skris 970955714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 971055714Skris 971155714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 971255714Skris doesn't work when the session is reused. Coming soon! 971355714Skris [Ben Laurie] 971455714Skris 971555714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 971655714Skris context thus bypassing client cert protection! All software that uses 971755714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 971855714Skris allow session reuse! A fuller solution is in the works. 971955714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 972055714Skris 972155714Skris *) Some more source tree cleanups (removed obsolete files 972255714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 972355714Skris permission on "config" script to be executable) and a fix for the INSTALL 972455714Skris document. 972555714Skris [Ulf Moeller <ulf@fitug.de>] 972655714Skris 972755714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 972855714Skris Malloc, Free. 972955714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 973055714Skris 973155714Skris *) Make rsa_oaep_test return non-zero on error. 973255714Skris [Ulf Moeller <ulf@fitug.de>] 973355714Skris 973455714Skris *) Add support for native Solaris shared libraries. Configure 973555714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 973655714Skris if someone would make that last step automatic. 973755714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 973855714Skris 973955714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 974055714Skris [Ben Laurie] 974155714Skris 974255714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 974355714Skris except NULL ciphers". This means the default cipher list will no longer 974455714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 974555714Skris the string "DEFAULT:eNULL". 974655714Skris [Steve Henson] 974755714Skris 974855714Skris *) Fix to RSA private encryption routines: if p < q then it would 974955714Skris occasionally produce an invalid result. This will only happen with 975055714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 975155714Skris [Steve Henson] 975255714Skris 975355714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 975455714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 975555714Skris because this way one can also use an interpreter named `perl5' (which is 975655714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 975755714Skris installed as `perl'). 975855714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 975955714Skris 976055714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 976155714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 976255714Skris 976355714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 976455714Skris advapi32.lib to Win32 build and change the pem test comparision 976555714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 976655714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 976755714Skris and crypto/des/ede_cbcm_enc.c. 976855714Skris [Steve Henson] 976955714Skris 977055714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 977155714Skris [Ben Laurie] 977255714Skris 977355714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 977455714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 977555714Skris is horrible: I feel ill.... 977655714Skris [Steve Henson] 977755714Skris 977855714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 977955714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 978055714Skris sections: 10 functions were absent from non ANSI section and not exported 978155714Skris from Windows DLLs. Fixed up libeay.num for new functions. 978255714Skris [Steve Henson] 978355714Skris 978455714Skris *) Make `openssl version' output lines consistent. 978555714Skris [Ralf S. Engelschall] 978655714Skris 978755714Skris *) Fix Win32 symbol export lists for BIO functions: Added 978855714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 978955714Skris to ms/libeay{16,32}.def. 979055714Skris [Ralf S. Engelschall] 979155714Skris 979255714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 979355714Skris fine under Unix and passes some trivial tests I've now added. But the 979455714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 979555714Skris added to make sure no one expects that this stuff really works in the 979655714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 979755714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 979855714Skris openssl_bio.xs. 979955714Skris [Ralf S. Engelschall] 980055714Skris 980155714Skris *) Fix the generation of two part addresses in perl. 980255714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 980355714Skris 980455714Skris *) Add config entry for Linux on MIPS. 980555714Skris [John Tobey <jtobey@channel1.com>] 980655714Skris 980755714Skris *) Make links whenever Configure is run, unless we are on Windoze. 980855714Skris [Ben Laurie] 980955714Skris 981055714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 981155714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 981255714Skris in CRLs. 981355714Skris [Steve Henson] 981455714Skris 981555714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 981655714Skris other platforms details on the command line without having to patch the 981755714Skris Configure script everytime: One now can use ``perl Configure 981855714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 981955714Skris to them (seperated by colons). This is treated as there would be a static 982055714Skris pre-configured entry in Configure's %table under key <id> with value 982155714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 982255714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 982355714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 982455714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 982555714Skris [Ralf S. Engelschall] 982655714Skris 982755714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 982855714Skris [Ben Laurie] 982955714Skris 983055714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 983155714Skris on the `perl Configure ...' command line. This way one can compile 983255714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 983355714Skris for linking it into DSOs. 983455714Skris [Ralf S. Engelschall] 983555714Skris 983655714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 983755714Skris Fixed. 983855714Skris [Ben Laurie] 983955714Skris 984055714Skris *) Cleaned up the LICENSE document: The official contact for any license 984155714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 984255714Skris And add a paragraph about the dual-license situation to make sure people 984355714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 984455714Skris to the OpenSSL toolkit. 984555714Skris [Ralf S. Engelschall] 984655714Skris 984755714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 984855714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 984955714Skris Additonally cleaned up the `make links' target: Remove unnecessary 985055714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 985155714Skris to speed processing and no longer clutter the display with confusing 985255714Skris stuff. Instead only the actually done links are displayed. 985355714Skris [Ralf S. Engelschall] 985455714Skris 985555714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 985655714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 985755714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 985855714Skris encryption. 985955714Skris [Ben Laurie] 986055714Skris 986155714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 986255714Skris signed attributes when verifying signatures (this would break them), 986355714Skris the detached data encoding was wrong and public keys obtained using 986455714Skris X509_get_pubkey() weren't freed. 986555714Skris [Steve Henson] 986655714Skris 986755714Skris *) Add text documentation for the BUFFER functions. Also added a work around 986855714Skris to a Win95 console bug. This was triggered by the password read stuff: the 986955714Skris last character typed gets carried over to the next fread(). If you were 987055714Skris generating a new cert request using 'req' for example then the last 987155714Skris character of the passphrase would be CR which would then enter the first 987255714Skris field as blank. 987355714Skris [Steve Henson] 987455714Skris 987555714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 987655714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 987755714Skris button and can be used by applications based on OpenSSL to show the 987855714Skris relationship to the OpenSSL project. 987955714Skris [Ralf S. Engelschall] 988055714Skris 988155714Skris *) Remove confusing variables in function signatures in files 988255714Skris ssl/ssl_lib.c and ssl/ssl.h. 988355714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 988455714Skris 988555714Skris *) Don't install bss_file.c under PREFIX/include/ 988655714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 988755714Skris 988855714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 988955714Skris functions that return function pointers and has support for NT specific 989055714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 989155714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 989255714Skris unsigned to signed types: this was killing the Win32 compile. 989355714Skris [Steve Henson] 989455714Skris 989555714Skris *) Add new certificate file to stack functions, 989655714Skris SSL_add_dir_cert_subjects_to_stack() and 989755714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 989855714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 989955714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 990055714Skris This means that Apache-SSL and similar packages don't have to mess around 990155714Skris to add as many CAs as they want to the preferred list. 990255714Skris [Ben Laurie] 990355714Skris 990455714Skris *) Experiment with doxygen documentation. Currently only partially applied to 990555714Skris ssl/ssl_lib.c. 990655714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 990755714Skris openssl.doxy as the configuration file. 990855714Skris [Ben Laurie] 990955714Skris 991055714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 991155714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 991255714Skris 991355714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 991455714Skris compiled in by default: it has problems with large keys. 991555714Skris [Steve Henson] 991655714Skris 991755714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 991855714Skris DH private keys and/or callback functions which directly correspond to 991955714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 992055714Skris is needed for applications which have to configure certificates on a 992155714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 992255714Skris (e.g. s_server). 992355714Skris For the RSA certificate situation is makes no difference, but 992455714Skris for the DSA certificate situation this fixes the "no shared cipher" 992555714Skris problem where the OpenSSL cipher selection procedure failed because the 992655714Skris temporary keys were not overtaken from the context and the API provided 992755714Skris no way to reconfigure them. 992855714Skris The new functions now let applications reconfigure the stuff and they 992955714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 993055714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 993155714Skris non-public-API function ssl_cert_instantiate() is used as a helper 993255714Skris function and also to reduce code redundancy inside ssl_rsa.c. 993355714Skris [Ralf S. Engelschall] 993455714Skris 993555714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 993655714Skris area because they are useful for the DSA situation and should be 993755714Skris recognized by the users. 993855714Skris [Ralf S. Engelschall] 993955714Skris 994055714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 994155714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 994255714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 994355714Skris already masked variable. 994455714Skris [Richard Levitte <levitte@stacken.kth.se>] 994555714Skris 994655714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 994755714Skris [Richard Levitte <levitte@stacken.kth.se>] 994855714Skris 994955714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 995055714Skris from `int' to `unsigned int' because it's a length and initialized by 995155714Skris EVP_DigestFinal() which expects an `unsigned int *'. 995255714Skris [Richard Levitte <levitte@stacken.kth.se>] 995355714Skris 995455714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 995555714Skris script. Instead use the usual Shell->Perl transition trick. 995655714Skris [Ralf S. Engelschall] 995755714Skris 995855714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 995955714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 996055714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 996155714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 996255714Skris currently the public key is printed (a decision which was already done by 996355714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 996455714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 996555714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 996655714Skris now, too. 996755714Skris [Ralf S. Engelschall] 996855714Skris 996955714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 997055714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 997155714Skris [Arne Ansper <arne@ats.cyber.ee>] 997255714Skris 997355714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 997455714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 997555714Skris config file. 997655714Skris [Steve Henson] 997755714Skris 997855714Skris *) Add cool BIO that does syslog (or event log on NT). 997955714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 998055714Skris 998155714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 998255714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 998355714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 998455714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 998555714Skris [Ben Laurie] 998655714Skris 998755714Skris *) Add preliminary config info for new extension code. 998855714Skris [Steve Henson] 998955714Skris 999055714Skris *) Make RSA_NO_PADDING really use no padding. 999155714Skris [Ulf Moeller <ulf@fitug.de>] 999255714Skris 999355714Skris *) Generate errors when private/public key check is done. 999455714Skris [Ben Laurie] 999555714Skris 999655714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 999755714Skris for some CRL extensions and new objects added. 999855714Skris [Steve Henson] 999955714Skris 1000055714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 1000155714Skris key usage extension and fuller support for authority key id. 1000255714Skris [Steve Henson] 1000355714Skris 1000455714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 1000555714Skris padding method for RSA, which is recommended for new applications in PKCS 1000655714Skris #1 v2.0 (RFC 2437, October 1998). 1000755714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 1000855714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 1000955714Skris against Bleichbacher's attack on RSA. 1001055714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 1001155714Skris Ben Laurie] 1001255714Skris 1001355714Skris *) Updates to the new SSL compression code 1001455714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 1001555714Skris 1001655714Skris *) Fix so that the version number in the master secret, when passed 1001755714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 1001855714Skris (because the server will not accept higher), that the version number 1001955714Skris is 0x03,0x01, not 0x03,0x00 1002055714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 1002155714Skris 1002255714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 1002355714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 1002455714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 1002555714Skris [Steve Henson] 1002655714Skris 1002755714Skris *) Support for RAW extensions where an arbitrary extension can be 1002855714Skris created by including its DER encoding. See apps/openssl.cnf for 1002955714Skris an example. 1003055714Skris [Steve Henson] 1003155714Skris 1003255714Skris *) Make sure latest Perl versions don't interpret some generated C array 1003355714Skris code as Perl array code in the crypto/err/err_genc.pl script. 1003455714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 1003555714Skris 1003655714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 1003755714Skris not many people have the assembler. Various Win32 compilation fixes and 1003855714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 1003955714Skris build instructions. 1004055714Skris [Steve Henson] 1004155714Skris 1004255714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 1004355714Skris file under Win32 and also build pem.h from pem.org. New script 1004455714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 1004555714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 1004655714Skris [Steve Henson] 1004755714Skris 1004855714Skris *) Major rework of DES function declarations, in the pursuit of correctness 1004955714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 1005055714Skris too. You may find this causes warnings in your code. Zapping your evil 1005155714Skris casts will probably fix them. Mostly. 1005255714Skris [Ben Laurie] 1005355714Skris 1005455714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 1005555714Skris obj_dat.pl. It considered a zero in an object definition to mean 1005655714Skris "end of object": none of the objects in objects.h have any zeros 1005755714Skris so it wasn't spotted. 1005855714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 1005955714Skris 1006055714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 1006155714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 1006255714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 1006355714Skris vectors if you have them. 1006455714Skris [Ben Laurie] 1006555714Skris 1006655714Skris *) Correct calculation of key length for export ciphers (too much space was 1006755714Skris allocated for null ciphers). This has not been tested! 1006855714Skris [Ben Laurie] 1006955714Skris 1007055714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 1007155714Skris message is now correct (it understands "crypto" and "ssl" on its 1007255714Skris command line). There is also now an "update" option. This will update 1007355714Skris the util/ssleay.num and util/libeay.num files with any new functions. 1007455714Skris If you do a: 1007555714Skris perl util/mkdef.pl crypto ssl update 1007655714Skris it will update them. 1007755714Skris [Steve Henson] 1007855714Skris 1007955714Skris *) Overhauled the Perl interface (perl/*): 1008055714Skris - ported BN stuff to OpenSSL's different BN library 1008155714Skris - made the perl/ source tree CVS-aware 1008255714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 1008355714Skris their history because I've copied them in the repository) 1008455714Skris - removed obsolete files (the test scripts will be replaced 1008555714Skris by better Test::Harness variants in the future) 1008655714Skris [Ralf S. Engelschall] 1008755714Skris 1008855714Skris *) First cut for a very conservative source tree cleanup: 1008955714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 1009055714Skris where we collect the old documents and readme texts. 1009155714Skris 2. remove the first part of files where I'm already sure that we no 1009255714Skris longer need them because of three reasons: either they are just temporary 1009355714Skris files which were left by Eric or they are preserved original files where 1009455714Skris I've verified that the diff is also available in the CVS via "cvs diff 1009555714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 1009655714Skris the crypto/md/ stuff). 1009755714Skris [Ralf S. Engelschall] 1009855714Skris 1009955714Skris *) More extension code. Incomplete support for subject and issuer alt 1010055714Skris name, issuer and authority key id. Change the i2v function parameters 1010155714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 1010255714Skris what that's for :-) Fix to ASN1 macro which messed up 1010355714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 1010455714Skris [Steve Henson] 1010555714Skris 1010655714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 1010755714Skris INTEGER code. 1010855714Skris [Steve Henson] 1010955714Skris 1011055714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 1011155714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 1011255714Skris 1011355714Skris *) Make sure `make rehash' target really finds the `openssl' program. 1011455714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 1011555714Skris 1011655714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 1011755714Skris like to hear about it if this slows down other processors. 1011855714Skris [Ben Laurie] 1011955714Skris 1012055714Skris *) Add CygWin32 platform information to Configure script. 1012155714Skris [Alan Batie <batie@aahz.jf.intel.com>] 1012255714Skris 1012355714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 1012455714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 1012555714Skris 1012655714Skris *) New program nseq to manipulate netscape certificate sequences 1012755714Skris [Steve Henson] 1012855714Skris 1012955714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 1013055714Skris few typos. 1013155714Skris [Steve Henson] 1013255714Skris 1013355714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 1013455714Skris but the BN code had some problems that would cause failures when 1013555714Skris doing certificate verification and some other functions. 1013655714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 1013755714Skris 1013855714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 1013955714Skris [Steve Henson] 1014055714Skris 1014155714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 1014255714Skris [Steve Henson] 1014355714Skris 1014455714Skris *) Add several PKIX and private extended key usage OIDs. 1014555714Skris [Steve Henson] 1014655714Skris 1014755714Skris *) Modify the 'ca' program to handle the new extension code. Modify 1014855714Skris openssl.cnf for new extension format, add comments. 1014955714Skris [Steve Henson] 1015055714Skris 1015155714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 1015255714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 1015355714Skris CA extensions. 1015455714Skris [Steve Henson] 1015555714Skris 1015655714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 1015755714Skris error code, add initial support to X509_print() and x509 application. 1015855714Skris [Steve Henson] 1015955714Skris 1016055714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 1016155714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 1016255714Skris stuff is currently isolated and isn't even compiled yet. 1016355714Skris [Steve Henson] 1016455714Skris 1016555714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 1016655714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 1016755714Skris Removed the versions check from X509 routines when loading extensions: 1016855714Skris this allows certain broken certificates that don't set the version 1016955714Skris properly to be processed. 1017055714Skris [Steve Henson] 1017155714Skris 1017255714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 1017355714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 1017455714Skris can still be regenerated with "make depend". 1017555714Skris [Ben Laurie] 1017655714Skris 1017755714Skris *) Spelling mistake in C version of CAST-128. 1017855714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 1017955714Skris 1018055714Skris *) Changes to the error generation code. The perl script err-code.pl 1018155714Skris now reads in the old error codes and retains the old numbers, only 1018255714Skris adding new ones if necessary. It also only changes the .err files if new 1018355714Skris codes are added. The makefiles have been modified to only insert errors 1018455714Skris when needed (to avoid needlessly modifying header files). This is done 1018555714Skris by only inserting errors if the .err file is newer than the auto generated 1018655714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 1018755714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 1018855714Skris or delete all the .err files. 1018955714Skris [Steve Henson] 1019055714Skris 1019155714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 1019255714Skris been fixed, but is untested. The assembler versions are also fixed, but 1019355714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 1019455714Skris to regenerate it if needed. 1019555714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 1019655714Skris Hagino <itojun@kame.net>] 1019755714Skris 1019855714Skris *) File was opened incorrectly in randfile.c. 1019955714Skris [Ulf M�ller <ulf@fitug.de>] 1020055714Skris 1020155714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 1020255714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 1020355714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 1020455714Skris al: it's just almost always a UTCTime. Note this patch adds new error 1020555714Skris codes so do a "make errors" if there are problems. 1020655714Skris [Steve Henson] 1020755714Skris 1020855714Skris *) Correct Linux 1 recognition in config. 1020955714Skris [Ulf M�ller <ulf@fitug.de>] 1021055714Skris 1021155714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 1021255714Skris [Anonymous <nobody@replay.com>] 1021355714Skris 1021455714Skris *) Generate an error if given an empty string as a cert directory. Also 1021555714Skris generate an error if handed NULL (previously returned 0 to indicate an 1021655714Skris error, but didn't set one). 1021755714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 1021855714Skris 1021955714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 1022055714Skris [Ben Laurie] 1022155714Skris 1022255714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 1022355714Skris parameters. This was causing a warning which killed off the Win32 compile. 1022455714Skris [Steve Henson] 1022555714Skris 1022655714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 1022755714Skris [Neil Costigan <neil.costigan@celocom.com>] 1022855714Skris 1022955714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 1023055714Skris based on a text string, looking up short and long names and finally 1023155714Skris "dot" format. The "dot" format stuff didn't work. Added new function 1023255714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 1023355714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 1023455714Skris OID is not part of the table. 1023555714Skris [Steve Henson] 1023655714Skris 1023755714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 1023855714Skris X509_LOOKUP_by_alias(). 1023955714Skris [Ben Laurie] 1024055714Skris 1024155714Skris *) Sort openssl functions by name. 1024255714Skris [Ben Laurie] 1024355714Skris 1024455714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 1024555714Skris encryption from sample DSA keys (in case anyone is interested the password 1024655714Skris was "1234"). 1024755714Skris [Steve Henson] 1024855714Skris 1024955714Skris *) Make _all_ *_free functions accept a NULL pointer. 1025055714Skris [Frans Heymans <fheymans@isaserver.be>] 1025155714Skris 1025255714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 1025355714Skris NULL pointers. 1025455714Skris [Anonymous <nobody@replay.com>] 1025555714Skris 1025655714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 1025755714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 1025855714Skris 1025955714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 1026055714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 1026155714Skris 1026255714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 1026355714Skris [Anonymous <nobody@replay.com>] 1026455714Skris 1026555714Skris *) Add prototype for temp key callback functions 1026655714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 1026755714Skris [Ben Laurie] 1026855714Skris 1026955714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 1027055714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 1027155714Skris [Steve Henson] 1027255714Skris 1027355714Skris *) X509_name_add_entry() freed the wrong thing after an error. 1027455714Skris [Arne Ansper <arne@ats.cyber.ee>] 1027555714Skris 1027655714Skris *) rsa_eay.c would attempt to free a NULL context. 1027755714Skris [Arne Ansper <arne@ats.cyber.ee>] 1027855714Skris 1027955714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 1028055714Skris [Arne Ansper <arne@ats.cyber.ee>] 1028155714Skris 1028255714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 1028355714Skris [Arne Ansper <arne@ats.cyber.ee>] 1028455714Skris 1028555714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 1028655714Skris in X509_STORE_new(), but document the fact that this variable is still 1028755714Skris unused in the certificate verification process. 1028855714Skris [Ralf S. Engelschall] 1028955714Skris 1029055714Skris *) Fix the various library and apps files to free up pkeys obtained from 1029155714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 1029255714Skris [Steve Henson] 1029355714Skris 1029455714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 1029555714Skris demos/maurice/example2.c work, amongst others, probably. 1029655714Skris [Steve Henson and Ben Laurie] 1029755714Skris 1029855714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 1029955714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 1030055714Skris are no longer created. This way we have a single and consistent command 1030155714Skris line interface `openssl <command>', similar to `cvs <command>'. 1030255714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 1030355714Skris 1030455714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 1030555714Skris BIT STRING wrapper always have zero unused bits. 1030655714Skris [Steve Henson] 1030755714Skris 1030855714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 1030955714Skris [Steve Henson] 1031055714Skris 1031155714Skris *) Make the top-level INSTALL documentation easier to understand. 1031255714Skris [Paul Sutton] 1031355714Skris 1031455714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 1031555714Skris make (including if user presses ^C) [Paul Sutton] 1031655714Skris 1031755714Skris *) Make Montgomery context stuff explicit in RSA data structure. 1031855714Skris [Ben Laurie] 1031955714Skris 1032055714Skris *) Fix build order of pem and err to allow for generated pem.h. 1032155714Skris [Ben Laurie] 1032255714Skris 1032355714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 1032455714Skris [Ben Laurie] 1032555714Skris 1032655714Skris *) Enhanced the err-ins.pl script so it makes the error library number 1032755714Skris global and can add a library name. This is needed for external ASN1 and 1032855714Skris other error libraries. 1032955714Skris [Steve Henson] 1033055714Skris 1033155714Skris *) Fixed sk_insert which never worked properly. 1033255714Skris [Steve Henson] 1033355714Skris 1033455714Skris *) Fix ASN1 macros so they can handle indefinite length construted 1033555714Skris EXPLICIT tags. Some non standard certificates use these: they can now 1033655714Skris be read in. 1033755714Skris [Steve Henson] 1033855714Skris 1033955714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 1034055714Skris into a single doc/ssleay.txt bundle. This way the information is still 1034155714Skris preserved but no longer messes up this directory. Now it's new room for 1034255714Skris the new set of documenation files. 1034355714Skris [Ralf S. Engelschall] 1034455714Skris 1034555714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 1034655714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 1034755714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 1034855714Skris number of arguments. 1034955714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 1035055714Skris 1035155714Skris *) Fix test data to work with the above. 1035255714Skris [Ben Laurie] 1035355714Skris 1035455714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 1035555714Skris was already fixed by Eric for 0.9.1 it seems. 1035655714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 1035755714Skris 1035855714Skris *) Autodetect FreeBSD3. 1035955714Skris [Ben Laurie] 1036055714Skris 1036155714Skris *) Fix various bugs in Configure. This affects the following platforms: 1036255714Skris nextstep 1036355714Skris ncr-scde 1036455714Skris unixware-2.0 1036555714Skris unixware-2.0-pentium 1036655714Skris sco5-cc. 1036755714Skris [Ben Laurie] 1036855714Skris 1036955714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 1037055714Skris before they are needed. 1037155714Skris [Ben Laurie] 1037255714Skris 1037355714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 1037455714Skris [Ben Laurie] 1037555714Skris 1037655714Skris 1037755714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 1037855714Skris 1037955714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 1038055714Skris changed SSLeay to OpenSSL in version strings. 1038155714Skris [Ralf S. Engelschall] 1038255714Skris 1038355714Skris *) Some fixups to the top-level documents. 1038455714Skris [Paul Sutton] 1038555714Skris 1038655714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 1038755714Skris because the symlink to include/ was missing. 1038855714Skris [Ralf S. Engelschall] 1038955714Skris 1039055714Skris *) Incorporated the popular no-RSA/DSA-only patches 1039155714Skris which allow to compile a RSA-free SSLeay. 1039255714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 1039355714Skris 1039455714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 1039555714Skris when "ssleay" is still not found. 1039655714Skris [Ralf S. Engelschall] 1039755714Skris 1039855714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 1039955714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 1040055714Skris 1040155714Skris *) Updated the README file. 1040255714Skris [Ralf S. Engelschall] 1040355714Skris 1040455714Skris *) Added various .cvsignore files in the CVS repository subdirs 1040555714Skris to make a "cvs update" really silent. 1040655714Skris [Ralf S. Engelschall] 1040755714Skris 1040855714Skris *) Recompiled the error-definition header files and added 1040955714Skris missing symbols to the Win32 linker tables. 1041055714Skris [Ralf S. Engelschall] 1041155714Skris 1041255714Skris *) Cleaned up the top-level documents; 1041355714Skris o new files: CHANGES and LICENSE 1041455714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 1041555714Skris o merged COPYRIGHT into LICENSE 1041655714Skris o removed obsolete TODO file 1041755714Skris o renamed MICROSOFT to INSTALL.W32 1041855714Skris [Ralf S. Engelschall] 1041955714Skris 1042055714Skris *) Removed dummy files from the 0.9.1b source tree: 1042155714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 1042255714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 1042355714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 1042455714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 1042555714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 1042655714Skris [Ralf S. Engelschall] 1042755714Skris 1042855714Skris *) Added various platform portability fixes. 1042955714Skris [Mark J. Cox] 1043055714Skris 1043155714Skris *) The Genesis of the OpenSSL rpject: 1043255714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 1043355714Skris Young and Tim J. Hudson created while they were working for C2Net until 1043455714Skris summer 1998. 1043555714Skris [The OpenSSL Project] 1043655714Skris 1043755714Skris 1043855714Skris Changes between 0.9.0b and 0.9.1b [not released] 1043955714Skris 1044055714Skris *) Updated a few CA certificates under certs/ 1044155714Skris [Eric A. Young] 1044255714Skris 1044355714Skris *) Changed some BIGNUM api stuff. 1044455714Skris [Eric A. Young] 1044555714Skris 1044655714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 1044755714Skris DGUX x86, Linux Alpha, etc. 1044855714Skris [Eric A. Young] 1044955714Skris 1045055714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 1045155714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 1045255714Skris available). 1045355714Skris [Eric A. Young] 1045455714Skris 1045555714Skris *) Add -strparse option to asn1pars program which parses nested 1045655714Skris binary structures 1045755714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 1045855714Skris 1045955714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 1046055714Skris [Eric A. Young] 1046155714Skris 1046255714Skris *) DSA fix for "ca" program. 1046355714Skris [Eric A. Young] 1046455714Skris 1046555714Skris *) Added "-genkey" option to "dsaparam" program. 1046655714Skris [Eric A. Young] 1046755714Skris 1046855714Skris *) Added RIPE MD160 (rmd160) message digest. 1046955714Skris [Eric A. Young] 1047055714Skris 1047155714Skris *) Added -a (all) option to "ssleay version" command. 1047255714Skris [Eric A. Young] 1047355714Skris 1047455714Skris *) Added PLATFORM define which is the id given to Configure. 1047555714Skris [Eric A. Young] 1047655714Skris 1047755714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 1047855714Skris [Eric A. Young] 1047955714Skris 1048055714Skris *) Extended the ASN.1 parser routines. 1048155714Skris [Eric A. Young] 1048255714Skris 1048355714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 1048455714Skris [Eric A. Young] 1048555714Skris 1048655714Skris *) Added a BN_CTX to the BN library. 1048755714Skris [Eric A. Young] 1048855714Skris 1048955714Skris *) Fixed the weak key values in DES library 1049055714Skris [Eric A. Young] 1049155714Skris 1049255714Skris *) Changed API in EVP library for cipher aliases. 1049355714Skris [Eric A. Young] 1049455714Skris 1049555714Skris *) Added support for RC2/64bit cipher. 1049655714Skris [Eric A. Young] 1049755714Skris 1049855714Skris *) Converted the lhash library to the crypto/mem.c functions. 1049955714Skris [Eric A. Young] 1050055714Skris 1050155714Skris *) Added more recognized ASN.1 object ids. 1050255714Skris [Eric A. Young] 1050355714Skris 1050455714Skris *) Added more RSA padding checks for SSL/TLS. 1050555714Skris [Eric A. Young] 1050655714Skris 1050755714Skris *) Added BIO proxy/filter functionality. 1050855714Skris [Eric A. Young] 1050955714Skris 1051055714Skris *) Added extra_certs to SSL_CTX which can be used 1051155714Skris send extra CA certificates to the client in the CA cert chain sending 1051255714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 1051355714Skris [Eric A. Young] 1051455714Skris 1051555714Skris *) Now Fortezza is denied in the authentication phase because 1051655714Skris this is key exchange mechanism is not supported by SSLeay at all. 1051755714Skris [Eric A. Young] 1051855714Skris 1051955714Skris *) Additional PKCS1 checks. 1052055714Skris [Eric A. Young] 1052155714Skris 1052255714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 1052355714Skris [Eric A. Young] 1052455714Skris 1052555714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 1052655714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 1052755714Skris [Eric A. Young] 1052855714Skris 1052955714Skris *) Fixed a few memory leaks. 1053055714Skris [Eric A. Young] 1053155714Skris 1053255714Skris *) Fixed various code and comment typos. 1053355714Skris [Eric A. Young] 1053455714Skris 1053555714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 1053655714Skris bytes sent in the client random. 1053755714Skris [Edward Bishop <ebishop@spyglass.com>] 1053855714Skris 10539