CHANGES revision 277597
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 5277597Sjkim Changes between 1.0.1k and 1.0.1l [15 Jan 2015] 6277597Sjkim 7277597Sjkim *) Build fixes for the Windows and OpenVMS platforms 8277597Sjkim [Matt Caswell and Richard Levitte] 9277597Sjkim 10276864Sjkim Changes between 1.0.1j and 1.0.1k [8 Jan 2015] 11276864Sjkim 12276864Sjkim *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS 13276864Sjkim message can cause a segmentation fault in OpenSSL due to a NULL pointer 14276864Sjkim dereference. This could lead to a Denial Of Service attack. Thanks to 15276864Sjkim Markus Stenberg of Cisco Systems, Inc. for reporting this issue. 16276864Sjkim (CVE-2014-3571) 17276864Sjkim [Steve Henson] 18276864Sjkim 19276864Sjkim *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the 20276864Sjkim dtls1_buffer_record function under certain conditions. In particular this 21276864Sjkim could occur if an attacker sent repeated DTLS records with the same 22276864Sjkim sequence number but for the next epoch. The memory leak could be exploited 23276864Sjkim by an attacker in a Denial of Service attack through memory exhaustion. 24276864Sjkim Thanks to Chris Mueller for reporting this issue. 25276864Sjkim (CVE-2015-0206) 26276864Sjkim [Matt Caswell] 27276864Sjkim 28276864Sjkim *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is 29276864Sjkim built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl 30276864Sjkim method would be set to NULL which could later result in a NULL pointer 31276864Sjkim dereference. Thanks to Frank Schmirler for reporting this issue. 32276864Sjkim (CVE-2014-3569) 33276864Sjkim [Kurt Roeckx] 34276864Sjkim 35276864Sjkim *) Abort handshake if server key exchange message is omitted for ephemeral 36276864Sjkim ECDH ciphersuites. 37276864Sjkim 38276864Sjkim Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for 39276864Sjkim reporting this issue. 40276864Sjkim (CVE-2014-3572) 41276864Sjkim [Steve Henson] 42276864Sjkim 43276864Sjkim *) Remove non-export ephemeral RSA code on client and server. This code 44276864Sjkim violated the TLS standard by allowing the use of temporary RSA keys in 45276864Sjkim non-export ciphersuites and could be used by a server to effectively 46276864Sjkim downgrade the RSA key length used to a value smaller than the server 47276864Sjkim certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at 48276864Sjkim INRIA or reporting this issue. 49276864Sjkim (CVE-2015-0204) 50276864Sjkim [Steve Henson] 51276864Sjkim 52276864Sjkim *) Fixed issue where DH client certificates are accepted without verification. 53276864Sjkim An OpenSSL server will accept a DH certificate for client authentication 54276864Sjkim without the certificate verify message. This effectively allows a client to 55276864Sjkim authenticate without the use of a private key. This only affects servers 56276864Sjkim which trust a client certificate authority which issues certificates 57276864Sjkim containing DH keys: these are extremely rare and hardly ever encountered. 58276864Sjkim Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting 59276864Sjkim this issue. 60276864Sjkim (CVE-2015-0205) 61276864Sjkim [Steve Henson] 62276864Sjkim 63276864Sjkim *) Ensure that the session ID context of an SSL is updated when its 64276864Sjkim SSL_CTX is updated via SSL_set_SSL_CTX. 65276864Sjkim 66276864Sjkim The session ID context is typically set from the parent SSL_CTX, 67276864Sjkim and can vary with the CTX. 68276864Sjkim [Adam Langley] 69276864Sjkim 70276864Sjkim *) Fix various certificate fingerprint issues. 71276864Sjkim 72276864Sjkim By using non-DER or invalid encodings outside the signed portion of a 73276864Sjkim certificate the fingerprint can be changed without breaking the signature. 74276864Sjkim Although no details of the signed portion of the certificate can be changed 75276864Sjkim this can cause problems with some applications: e.g. those using the 76276864Sjkim certificate fingerprint for blacklists. 77276864Sjkim 78276864Sjkim 1. Reject signatures with non zero unused bits. 79276864Sjkim 80276864Sjkim If the BIT STRING containing the signature has non zero unused bits reject 81276864Sjkim the signature. All current signature algorithms require zero unused bits. 82276864Sjkim 83276864Sjkim 2. Check certificate algorithm consistency. 84276864Sjkim 85276864Sjkim Check the AlgorithmIdentifier inside TBS matches the one in the 86276864Sjkim certificate signature. NB: this will result in signature failure 87276864Sjkim errors for some broken certificates. 88276864Sjkim 89276864Sjkim Thanks to Konrad Kraszewski from Google for reporting this issue. 90276864Sjkim 91276864Sjkim 3. Check DSA/ECDSA signatures use DER. 92276864Sjkim 93276864Sjkim Reencode DSA/ECDSA signatures and compare with the original received 94276864Sjkim signature. Return an error if there is a mismatch. 95276864Sjkim 96276864Sjkim This will reject various cases including garbage after signature 97276864Sjkim (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS 98276864Sjkim program for discovering this case) and use of BER or invalid ASN.1 INTEGERs 99276864Sjkim (negative or with leading zeroes). 100276864Sjkim 101276864Sjkim Further analysis was conducted and fixes were developed by Stephen Henson 102276864Sjkim of the OpenSSL core team. 103276864Sjkim 104276864Sjkim (CVE-2014-8275) 105276864Sjkim [Steve Henson] 106276864Sjkim 107276864Sjkim *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect 108276864Sjkim results on some platforms, including x86_64. This bug occurs at random 109276864Sjkim with a very low probability, and is not known to be exploitable in any 110276864Sjkim way, though its exact impact is difficult to determine. Thanks to Pieter 111276864Sjkim Wuille (Blockstream) who reported this issue and also suggested an initial 112276864Sjkim fix. Further analysis was conducted by the OpenSSL development team and 113276864Sjkim Adam Langley of Google. The final fix was developed by Andy Polyakov of 114276864Sjkim the OpenSSL core team. 115276864Sjkim (CVE-2014-3570) 116276864Sjkim [Andy Polyakov] 117276864Sjkim 118276864Sjkim *) Do not resume sessions on the server if the negotiated protocol 119276864Sjkim version does not match the session's version. Resuming with a different 120276864Sjkim version, while not strictly forbidden by the RFC, is of questionable 121276864Sjkim sanity and breaks all known clients. 122276864Sjkim [David Benjamin, Emilia K�sper] 123276864Sjkim 124276864Sjkim *) Tighten handling of the ChangeCipherSpec (CCS) message: reject 125276864Sjkim early CCS messages during renegotiation. (Note that because 126276864Sjkim renegotiation is encrypted, this early CCS was not exploitable.) 127276864Sjkim [Emilia K�sper] 128276864Sjkim 129276864Sjkim *) Tighten client-side session ticket handling during renegotiation: 130276864Sjkim ensure that the client only accepts a session ticket if the server sends 131276864Sjkim the extension anew in the ServerHello. Previously, a TLS client would 132276864Sjkim reuse the old extension state and thus accept a session ticket if one was 133276864Sjkim announced in the initial ServerHello. 134276864Sjkim 135276864Sjkim Similarly, ensure that the client requires a session ticket if one 136276864Sjkim was advertised in the ServerHello. Previously, a TLS client would 137276864Sjkim ignore a missing NewSessionTicket message. 138276864Sjkim [Emilia K�sper] 139276864Sjkim 140273149Sjkim Changes between 1.0.1i and 1.0.1j [15 Oct 2014] 141273149Sjkim 142273149Sjkim *) SRTP Memory Leak. 143273149Sjkim 144273149Sjkim A flaw in the DTLS SRTP extension parsing code allows an attacker, who 145273149Sjkim sends a carefully crafted handshake message, to cause OpenSSL to fail 146273149Sjkim to free up to 64k of memory causing a memory leak. This could be 147273149Sjkim exploited in a Denial Of Service attack. This issue affects OpenSSL 148273149Sjkim 1.0.1 server implementations for both SSL/TLS and DTLS regardless of 149273149Sjkim whether SRTP is used or configured. Implementations of OpenSSL that 150273149Sjkim have been compiled with OPENSSL_NO_SRTP defined are not affected. 151273149Sjkim 152273149Sjkim The fix was developed by the OpenSSL team. 153273149Sjkim (CVE-2014-3513) 154273149Sjkim [OpenSSL team] 155273149Sjkim 156273149Sjkim *) Session Ticket Memory Leak. 157273149Sjkim 158273149Sjkim When an OpenSSL SSL/TLS/DTLS server receives a session ticket the 159273149Sjkim integrity of that ticket is first verified. In the event of a session 160273149Sjkim ticket integrity check failing, OpenSSL will fail to free memory 161273149Sjkim causing a memory leak. By sending a large number of invalid session 162273149Sjkim tickets an attacker could exploit this issue in a Denial Of Service 163273149Sjkim attack. 164273149Sjkim (CVE-2014-3567) 165273149Sjkim [Steve Henson] 166273149Sjkim 167273149Sjkim *) Build option no-ssl3 is incomplete. 168273149Sjkim 169273149Sjkim When OpenSSL is configured with "no-ssl3" as a build option, servers 170273149Sjkim could accept and complete a SSL 3.0 handshake, and clients could be 171273149Sjkim configured to send them. 172273149Sjkim (CVE-2014-3568) 173273149Sjkim [Akamai and the OpenSSL team] 174273149Sjkim 175273149Sjkim *) Add support for TLS_FALLBACK_SCSV. 176273149Sjkim Client applications doing fallback retries should call 177273149Sjkim SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV). 178273149Sjkim (CVE-2014-3566) 179273149Sjkim [Adam Langley, Bodo Moeller] 180273149Sjkim 181273149Sjkim *) Add additional DigestInfo checks. 182273149Sjkim 183273149Sjkim Reencode DigestInto in DER and check against the original when 184273149Sjkim verifying RSA signature: this will reject any improperly encoded 185273149Sjkim DigestInfo structures. 186273149Sjkim 187273149Sjkim Note: this is a precautionary measure and no attacks are currently known. 188273149Sjkim 189273149Sjkim [Steve Henson] 190273149Sjkim 191269686Sjkim Changes between 1.0.1h and 1.0.1i [6 Aug 2014] 192269686Sjkim 193269686Sjkim *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the 194269686Sjkim SRP code can be overrun an internal buffer. Add sanity check that 195269686Sjkim g, A, B < N to SRP code. 196269686Sjkim 197269686Sjkim Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC 198269686Sjkim Group for discovering this issue. 199269686Sjkim (CVE-2014-3512) 200269686Sjkim [Steve Henson] 201269686Sjkim 202269686Sjkim *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate 203269686Sjkim TLS 1.0 instead of higher protocol versions when the ClientHello message 204269686Sjkim is badly fragmented. This allows a man-in-the-middle attacker to force a 205269686Sjkim downgrade to TLS 1.0 even if both the server and the client support a 206269686Sjkim higher protocol version, by modifying the client's TLS records. 207269686Sjkim 208269686Sjkim Thanks to David Benjamin and Adam Langley (Google) for discovering and 209269686Sjkim researching this issue. 210269686Sjkim (CVE-2014-3511) 211269686Sjkim [David Benjamin] 212269686Sjkim 213269686Sjkim *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject 214269686Sjkim to a denial of service attack. A malicious server can crash the client 215269686Sjkim with a null pointer dereference (read) by specifying an anonymous (EC)DH 216269686Sjkim ciphersuite and sending carefully crafted handshake messages. 217269686Sjkim 218269686Sjkim Thanks to Felix Gr�bert (Google) for discovering and researching this 219269686Sjkim issue. 220269686Sjkim (CVE-2014-3510) 221269686Sjkim [Emilia K�sper] 222269686Sjkim 223269686Sjkim *) By sending carefully crafted DTLS packets an attacker could cause openssl 224269686Sjkim to leak memory. This can be exploited through a Denial of Service attack. 225269686Sjkim Thanks to Adam Langley for discovering and researching this issue. 226269686Sjkim (CVE-2014-3507) 227269686Sjkim [Adam Langley] 228269686Sjkim 229269686Sjkim *) An attacker can force openssl to consume large amounts of memory whilst 230269686Sjkim processing DTLS handshake messages. This can be exploited through a 231269686Sjkim Denial of Service attack. 232269686Sjkim Thanks to Adam Langley for discovering and researching this issue. 233269686Sjkim (CVE-2014-3506) 234269686Sjkim [Adam Langley] 235269686Sjkim 236269686Sjkim *) An attacker can force an error condition which causes openssl to crash 237269686Sjkim whilst processing DTLS packets due to memory being freed twice. This 238269686Sjkim can be exploited through a Denial of Service attack. 239269686Sjkim Thanks to Adam Langley and Wan-Teh Chang for discovering and researching 240269686Sjkim this issue. 241269686Sjkim (CVE-2014-3505) 242269686Sjkim [Adam Langley] 243269686Sjkim 244269686Sjkim *) If a multithreaded client connects to a malicious server using a resumed 245269686Sjkim session and the server sends an ec point format extension it could write 246269686Sjkim up to 255 bytes to freed memory. 247269686Sjkim 248269686Sjkim Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this 249269686Sjkim issue. 250269686Sjkim (CVE-2014-3509) 251269686Sjkim [Gabor Tyukasz] 252269686Sjkim 253269686Sjkim *) A malicious server can crash an OpenSSL client with a null pointer 254269686Sjkim dereference (read) by specifying an SRP ciphersuite even though it was not 255269686Sjkim properly negotiated with the client. This can be exploited through a 256269686Sjkim Denial of Service attack. 257269686Sjkim 258269686Sjkim Thanks to Joonas Kuorilehto and Riku Hietam�ki (Codenomicon) for 259269686Sjkim discovering and researching this issue. 260269686Sjkim (CVE-2014-5139) 261269686Sjkim [Steve Henson] 262269686Sjkim 263269686Sjkim *) A flaw in OBJ_obj2txt may cause pretty printing functions such as 264269686Sjkim X509_name_oneline, X509_name_print_ex et al. to leak some information 265269686Sjkim from the stack. Applications may be affected if they echo pretty printing 266269686Sjkim output to the attacker. 267269686Sjkim 268269686Sjkim Thanks to Ivan Fratric (Google) for discovering this issue. 269269686Sjkim (CVE-2014-3508) 270269686Sjkim [Emilia K�sper, and Steve Henson] 271269686Sjkim 272269686Sjkim *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) 273269686Sjkim for corner cases. (Certain input points at infinity could lead to 274269686Sjkim bogus results, with non-infinity inputs mapped to infinity too.) 275269686Sjkim [Bodo Moeller] 276269686Sjkim 277267258Sjkim Changes between 1.0.1g and 1.0.1h [5 Jun 2014] 278267258Sjkim 279267258Sjkim *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted 280267258Sjkim handshake can force the use of weak keying material in OpenSSL 281267258Sjkim SSL/TLS clients and servers. 282267258Sjkim 283267258Sjkim Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and 284267258Sjkim researching this issue. (CVE-2014-0224) 285267258Sjkim [KIKUCHI Masashi, Steve Henson] 286267258Sjkim 287267258Sjkim *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an 288267258Sjkim OpenSSL DTLS client the code can be made to recurse eventually crashing 289267258Sjkim in a DoS attack. 290267258Sjkim 291267258Sjkim Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue. 292267258Sjkim (CVE-2014-0221) 293267258Sjkim [Imre Rad, Steve Henson] 294267258Sjkim 295267258Sjkim *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can 296267258Sjkim be triggered by sending invalid DTLS fragments to an OpenSSL DTLS 297267258Sjkim client or server. This is potentially exploitable to run arbitrary 298267258Sjkim code on a vulnerable client or server. 299267258Sjkim 300267258Sjkim Thanks to J�ri Aedla for reporting this issue. (CVE-2014-0195) 301267258Sjkim [J�ri Aedla, Steve Henson] 302267258Sjkim 303267258Sjkim *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites 304267258Sjkim are subject to a denial of service attack. 305267258Sjkim 306267258Sjkim Thanks to Felix Gr�bert and Ivan Fratric at Google for discovering 307267258Sjkim this issue. (CVE-2014-3470) 308267258Sjkim [Felix Gr�bert, Ivan Fratric, Steve Henson] 309267258Sjkim 310267258Sjkim *) Harmonize version and its documentation. -f flag is used to display 311267258Sjkim compilation flags. 312267258Sjkim [mancha <mancha1@zoho.com>] 313267258Sjkim 314267258Sjkim *) Fix eckey_priv_encode so it immediately returns an error upon a failure 315267258Sjkim in i2d_ECPrivateKey. 316267258Sjkim [mancha <mancha1@zoho.com>] 317267258Sjkim 318267258Sjkim *) Fix some double frees. These are not thought to be exploitable. 319267258Sjkim [mancha <mancha1@zoho.com>] 320267258Sjkim 321264331Sjkim Changes between 1.0.1f and 1.0.1g [7 Apr 2014] 322264331Sjkim 323264331Sjkim *) A missing bounds check in the handling of the TLS heartbeat extension 324264331Sjkim can be used to reveal up to 64k of memory to a connected client or 325264331Sjkim server. 326264331Sjkim 327264331Sjkim Thanks for Neel Mehta of Google Security for discovering this bug and to 328264331Sjkim Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for 329264331Sjkim preparing the fix (CVE-2014-0160) 330264331Sjkim [Adam Langley, Bodo Moeller] 331264331Sjkim 332264331Sjkim *) Fix for the attack described in the paper "Recovering OpenSSL 333264331Sjkim ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack" 334264331Sjkim by Yuval Yarom and Naomi Benger. Details can be obtained from: 335264331Sjkim http://eprint.iacr.org/2014/140 336264331Sjkim 337264331Sjkim Thanks to Yuval Yarom and Naomi Benger for discovering this 338264331Sjkim flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076) 339264331Sjkim [Yuval Yarom and Naomi Benger] 340264331Sjkim 341264331Sjkim *) TLS pad extension: draft-agl-tls-padding-03 342264331Sjkim 343264331Sjkim Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the 344264331Sjkim TLS client Hello record length value would otherwise be > 255 and 345264331Sjkim less that 512 pad with a dummy extension containing zeroes so it 346264331Sjkim is at least 512 bytes long. 347264331Sjkim 348264331Sjkim [Adam Langley, Steve Henson] 349264331Sjkim 350264331Sjkim Changes between 1.0.1e and 1.0.1f [6 Jan 2014] 351264331Sjkim 352264331Sjkim *) Fix for TLS record tampering bug. A carefully crafted invalid 353264331Sjkim handshake could crash OpenSSL with a NULL pointer exception. 354264331Sjkim Thanks to Anton Johansson for reporting this issues. 355264331Sjkim (CVE-2013-4353) 356264331Sjkim 357264331Sjkim *) Keep original DTLS digest and encryption contexts in retransmission 358264331Sjkim structures so we can use the previous session parameters if they need 359264331Sjkim to be resent. (CVE-2013-6450) 360264331Sjkim [Steve Henson] 361264331Sjkim 362264331Sjkim *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which 363264331Sjkim avoids preferring ECDHE-ECDSA ciphers when the client appears to be 364264331Sjkim Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for 365264331Sjkim several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug 366264331Sjkim is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing 367264331Sjkim 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer. 368264331Sjkim [Rob Stradling, Adam Langley] 369264331Sjkim 370246772Sjkim Changes between 1.0.1d and 1.0.1e [11 Feb 2013] 371246772Sjkim 372264331Sjkim *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI 373264331Sjkim supporting platforms or when small records were transferred. 374264331Sjkim [Andy Polyakov, Steve Henson] 375246772Sjkim 376246772Sjkim Changes between 1.0.1c and 1.0.1d [5 Feb 2013] 377246772Sjkim 378246772Sjkim *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 379246772Sjkim 380246772Sjkim This addresses the flaw in CBC record processing discovered by 381246772Sjkim Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 382246772Sjkim at: http://www.isg.rhul.ac.uk/tls/ 383246772Sjkim 384246772Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 385246772Sjkim Security Group at Royal Holloway, University of London 386246772Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 387246772Sjkim Emilia K�sper for the initial patch. 388246772Sjkim (CVE-2013-0169) 389246772Sjkim [Emilia K�sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 390246772Sjkim 391246772Sjkim *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode 392246772Sjkim ciphersuites which can be exploited in a denial of service attack. 393246772Sjkim Thanks go to and to Adam Langley <agl@chromium.org> for discovering 394246772Sjkim and detecting this bug and to Wolfgang Ettlinger 395246772Sjkim <wolfgang.ettlinger@gmail.com> for independently discovering this issue. 396246772Sjkim (CVE-2012-2686) 397246772Sjkim [Adam Langley] 398246772Sjkim 399246772Sjkim *) Return an error when checking OCSP signatures when key is NULL. 400246772Sjkim This fixes a DoS attack. (CVE-2013-0166) 401246772Sjkim [Steve Henson] 402246772Sjkim 403246772Sjkim *) Make openssl verify return errors. 404246772Sjkim [Chris Palmer <palmer@google.com> and Ben Laurie] 405246772Sjkim 406246772Sjkim *) Call OCSP Stapling callback after ciphersuite has been chosen, so 407246772Sjkim the right response is stapled. Also change SSL_get_certificate() 408246772Sjkim so it returns the certificate actually sent. 409246772Sjkim See http://rt.openssl.org/Ticket/Display.html?id=2836. 410246772Sjkim [Rob Stradling <rob.stradling@comodo.com>] 411246772Sjkim 412246772Sjkim *) Fix possible deadlock when decoding public keys. 413246772Sjkim [Steve Henson] 414246772Sjkim 415246772Sjkim *) Don't use TLS 1.0 record version number in initial client hello 416246772Sjkim if renegotiating. 417246772Sjkim [Steve Henson] 418246772Sjkim 419238405Sjkim Changes between 1.0.1b and 1.0.1c [10 May 2012] 420237657Sjkim 421238405Sjkim *) Sanity check record length before skipping explicit IV in TLS 422238405Sjkim 1.2, 1.1 and DTLS to fix DoS attack. 423237657Sjkim 424237657Sjkim Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 425237657Sjkim fuzzing as a service testing platform. 426237657Sjkim (CVE-2012-2333) 427237657Sjkim [Steve Henson] 428237657Sjkim 429237657Sjkim *) Initialise tkeylen properly when encrypting CMS messages. 430237657Sjkim Thanks to Solar Designer of Openwall for reporting this issue. 431237657Sjkim [Steve Henson] 432237657Sjkim 433238405Sjkim *) In FIPS mode don't try to use composite ciphers as they are not 434238405Sjkim approved. 435238405Sjkim [Steve Henson] 436237657Sjkim 437238405Sjkim Changes between 1.0.1a and 1.0.1b [26 Apr 2012] 438237657Sjkim 439238405Sjkim *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and 440238405Sjkim 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately 441238405Sjkim mean any application compiled against OpenSSL 1.0.0 headers setting 442238405Sjkim SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng 443238405Sjkim TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to 444238405Sjkim 0x10000000L Any application which was previously compiled against 445238405Sjkim OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1 446238405Sjkim will need to be recompiled as a result. Letting be results in 447238405Sjkim inability to disable specifically TLS 1.1 and in client context, 448238405Sjkim in unlike event, limit maximum offered version to TLS 1.0 [see below]. 449238405Sjkim [Steve Henson] 450237657Sjkim 451238405Sjkim *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not 452238405Sjkim disable just protocol X, but all protocols above X *if* there are 453238405Sjkim protocols *below* X still enabled. In more practical terms it means 454238405Sjkim that if application wants to disable TLS1.0 in favor of TLS1.1 and 455238405Sjkim above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass 456238405Sjkim SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to 457238405Sjkim client side. 458238405Sjkim [Andy Polyakov] 459238405Sjkim 460238405Sjkim Changes between 1.0.1 and 1.0.1a [19 Apr 2012] 461238405Sjkim 462237657Sjkim *) Check for potentially exploitable overflows in asn1_d2i_read_bio 463237657Sjkim BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 464237657Sjkim in CRYPTO_realloc_clean. 465237657Sjkim 466237657Sjkim Thanks to Tavis Ormandy, Google Security Team, for discovering this 467237657Sjkim issue and to Adam Langley <agl@chromium.org> for fixing it. 468237657Sjkim (CVE-2012-2110) 469237657Sjkim [Adam Langley (Google), Tavis Ormandy, Google Security Team] 470237657Sjkim 471238405Sjkim *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections. 472238405Sjkim [Adam Langley] 473237657Sjkim 474238405Sjkim *) Workarounds for some broken servers that "hang" if a client hello 475238405Sjkim record length exceeds 255 bytes. 476238405Sjkim 477238405Sjkim 1. Do not use record version number > TLS 1.0 in initial client 478238405Sjkim hello: some (but not all) hanging servers will now work. 479238405Sjkim 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate 480238405Sjkim the number of ciphers sent in the client hello. This should be 481238405Sjkim set to an even number, such as 50, for example by passing: 482238405Sjkim -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure. 483238405Sjkim Most broken servers should now work. 484238405Sjkim 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable 485238405Sjkim TLS 1.2 client support entirely. 486238405Sjkim [Steve Henson] 487238405Sjkim 488238405Sjkim *) Fix SEGV in Vector Permutation AES module observed in OpenSSH. 489238405Sjkim [Andy Polyakov] 490238405Sjkim 491238405Sjkim Changes between 1.0.0h and 1.0.1 [14 Mar 2012] 492238405Sjkim 493238405Sjkim *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET 494238405Sjkim STRING form instead of a DigestInfo. 495238405Sjkim [Steve Henson] 496238405Sjkim 497238405Sjkim *) The format used for MDC2 RSA signatures is inconsistent between EVP 498238405Sjkim and the RSA_sign/RSA_verify functions. This was made more apparent when 499238405Sjkim OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular 500238405Sjkim those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect 501238405Sjkim the correct format in RSA_verify so both forms transparently work. 502238405Sjkim [Steve Henson] 503238405Sjkim 504238405Sjkim *) Some servers which support TLS 1.0 can choke if we initially indicate 505238405Sjkim support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA 506238405Sjkim encrypted premaster secret. As a workaround use the maximum pemitted 507238405Sjkim client version in client hello, this should keep such servers happy 508238405Sjkim and still work with previous versions of OpenSSL. 509238405Sjkim [Steve Henson] 510238405Sjkim 511238405Sjkim *) Add support for TLS/DTLS heartbeats. 512238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 513238405Sjkim 514238405Sjkim *) Add support for SCTP. 515238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 516238405Sjkim 517238405Sjkim *) Improved PRNG seeding for VOS. 518238405Sjkim [Paul Green <Paul.Green@stratus.com>] 519238405Sjkim 520238405Sjkim *) Extensive assembler packs updates, most notably: 521238405Sjkim 522238405Sjkim - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support; 523238405Sjkim - x86[_64]: SSSE3 support (SHA1, vector-permutation AES); 524238405Sjkim - x86_64: bit-sliced AES implementation; 525238405Sjkim - ARM: NEON support, contemporary platforms optimizations; 526238405Sjkim - s390x: z196 support; 527238405Sjkim - *: GHASH and GF(2^m) multiplication implementations; 528238405Sjkim 529238405Sjkim [Andy Polyakov] 530238405Sjkim 531238405Sjkim *) Make TLS-SRP code conformant with RFC 5054 API cleanup 532238405Sjkim (removal of unnecessary code) 533238405Sjkim [Peter Sylvester <peter.sylvester@edelweb.fr>] 534238405Sjkim 535238405Sjkim *) Add TLS key material exporter from RFC 5705. 536238405Sjkim [Eric Rescorla] 537238405Sjkim 538238405Sjkim *) Add DTLS-SRTP negotiation from RFC 5764. 539238405Sjkim [Eric Rescorla] 540238405Sjkim 541238405Sjkim *) Add Next Protocol Negotiation, 542238405Sjkim http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be 543238405Sjkim disabled with a no-npn flag to config or Configure. Code donated 544238405Sjkim by Google. 545238405Sjkim [Adam Langley <agl@google.com> and Ben Laurie] 546238405Sjkim 547238405Sjkim *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224, 548238405Sjkim NIST-P256, NIST-P521, with constant-time single point multiplication on 549238405Sjkim typical inputs. Compiler support for the nonstandard type __uint128_t is 550238405Sjkim required to use this (present in gcc 4.4 and later, for 64-bit builds). 551238405Sjkim Code made available under Apache License version 2.0. 552238405Sjkim 553238405Sjkim Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command 554238405Sjkim line to include this in your build of OpenSSL, and run "make depend" (or 555238405Sjkim "make update"). This enables the following EC_METHODs: 556238405Sjkim 557238405Sjkim EC_GFp_nistp224_method() 558238405Sjkim EC_GFp_nistp256_method() 559238405Sjkim EC_GFp_nistp521_method() 560238405Sjkim 561238405Sjkim EC_GROUP_new_by_curve_name() will automatically use these (while 562238405Sjkim EC_GROUP_new_curve_GFp() currently prefers the more flexible 563238405Sjkim implementations). 564238405Sjkim [Emilia K�sper, Adam Langley, Bodo Moeller (Google)] 565238405Sjkim 566238405Sjkim *) Use type ossl_ssize_t instad of ssize_t which isn't available on 567238405Sjkim all platforms. Move ssize_t definition from e_os.h to the public 568238405Sjkim header file e_os2.h as it now appears in public header file cms.h 569238405Sjkim [Steve Henson] 570238405Sjkim 571238405Sjkim *) New -sigopt option to the ca, req and x509 utilities. Additional 572238405Sjkim signature parameters can be passed using this option and in 573238405Sjkim particular PSS. 574238405Sjkim [Steve Henson] 575238405Sjkim 576238405Sjkim *) Add RSA PSS signing function. This will generate and set the 577238405Sjkim appropriate AlgorithmIdentifiers for PSS based on those in the 578238405Sjkim corresponding EVP_MD_CTX structure. No application support yet. 579238405Sjkim [Steve Henson] 580238405Sjkim 581238405Sjkim *) Support for companion algorithm specific ASN1 signing routines. 582238405Sjkim New function ASN1_item_sign_ctx() signs a pre-initialised 583238405Sjkim EVP_MD_CTX structure and sets AlgorithmIdentifiers based on 584238405Sjkim the appropriate parameters. 585238405Sjkim [Steve Henson] 586238405Sjkim 587238405Sjkim *) Add new algorithm specific ASN1 verification initialisation function 588238405Sjkim to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1 589238405Sjkim handling will be the same no matter what EVP_PKEY_METHOD is used. 590238405Sjkim Add a PSS handler to support verification of PSS signatures: checked 591238405Sjkim against a number of sample certificates. 592238405Sjkim [Steve Henson] 593238405Sjkim 594238405Sjkim *) Add signature printing for PSS. Add PSS OIDs. 595238405Sjkim [Steve Henson, Martin Kaiser <lists@kaiser.cx>] 596238405Sjkim 597238405Sjkim *) Add algorithm specific signature printing. An individual ASN1 method 598238405Sjkim can now print out signatures instead of the standard hex dump. 599238405Sjkim 600238405Sjkim More complex signatures (e.g. PSS) can print out more meaningful 601238405Sjkim information. Include DSA version that prints out the signature 602238405Sjkim parameters r, s. 603238405Sjkim [Steve Henson] 604238405Sjkim 605238405Sjkim *) Password based recipient info support for CMS library: implementing 606238405Sjkim RFC3211. 607238405Sjkim [Steve Henson] 608238405Sjkim 609238405Sjkim *) Split password based encryption into PBES2 and PBKDF2 functions. This 610238405Sjkim neatly separates the code into cipher and PBE sections and is required 611238405Sjkim for some algorithms that split PBES2 into separate pieces (such as 612238405Sjkim password based CMS). 613238405Sjkim [Steve Henson] 614238405Sjkim 615238405Sjkim *) Session-handling fixes: 616238405Sjkim - Fix handling of connections that are resuming with a session ID, 617238405Sjkim but also support Session Tickets. 618238405Sjkim - Fix a bug that suppressed issuing of a new ticket if the client 619238405Sjkim presented a ticket with an expired session. 620238405Sjkim - Try to set the ticket lifetime hint to something reasonable. 621238405Sjkim - Make tickets shorter by excluding irrelevant information. 622238405Sjkim - On the client side, don't ignore renewed tickets. 623238405Sjkim [Adam Langley, Bodo Moeller (Google)] 624238405Sjkim 625238405Sjkim *) Fix PSK session representation. 626238405Sjkim [Bodo Moeller] 627238405Sjkim 628238405Sjkim *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations. 629238405Sjkim 630238405Sjkim This work was sponsored by Intel. 631238405Sjkim [Andy Polyakov] 632238405Sjkim 633238405Sjkim *) Add GCM support to TLS library. Some custom code is needed to split 634238405Sjkim the IV between the fixed (from PRF) and explicit (from TLS record) 635238405Sjkim portions. This adds all GCM ciphersuites supported by RFC5288 and 636238405Sjkim RFC5289. Generalise some AES* cipherstrings to inlclude GCM and 637238405Sjkim add a special AESGCM string for GCM only. 638238405Sjkim [Steve Henson] 639238405Sjkim 640238405Sjkim *) Expand range of ctrls for AES GCM. Permit setting invocation 641238405Sjkim field on decrypt and retrieval of invocation field only on encrypt. 642238405Sjkim [Steve Henson] 643238405Sjkim 644238405Sjkim *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support. 645238405Sjkim As required by RFC5289 these ciphersuites cannot be used if for 646238405Sjkim versions of TLS earlier than 1.2. 647238405Sjkim [Steve Henson] 648238405Sjkim 649238405Sjkim *) For FIPS capable OpenSSL interpret a NULL default public key method 650238405Sjkim as unset and return the appopriate default but do *not* set the default. 651238405Sjkim This means we can return the appopriate method in applications that 652238405Sjkim swicth between FIPS and non-FIPS modes. 653238405Sjkim [Steve Henson] 654238405Sjkim 655238405Sjkim *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an 656238405Sjkim ENGINE is used then we cannot handle that in the FIPS module so we 657238405Sjkim keep original code iff non-FIPS operations are allowed. 658238405Sjkim [Steve Henson] 659238405Sjkim 660238405Sjkim *) Add -attime option to openssl utilities. 661238405Sjkim [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson] 662238405Sjkim 663238405Sjkim *) Redirect DSA and DH operations to FIPS module in FIPS mode. 664238405Sjkim [Steve Henson] 665238405Sjkim 666238405Sjkim *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use 667238405Sjkim FIPS EC methods unconditionally for now. 668238405Sjkim [Steve Henson] 669238405Sjkim 670238405Sjkim *) New build option no-ec2m to disable characteristic 2 code. 671238405Sjkim [Steve Henson] 672238405Sjkim 673238405Sjkim *) Backport libcrypto audit of return value checking from 1.1.0-dev; not 674238405Sjkim all cases can be covered as some introduce binary incompatibilities. 675238405Sjkim [Steve Henson] 676238405Sjkim 677238405Sjkim *) Redirect RSA operations to FIPS module including keygen, 678238405Sjkim encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods. 679238405Sjkim [Steve Henson] 680238405Sjkim 681238405Sjkim *) Add similar low level API blocking to ciphers. 682238405Sjkim [Steve Henson] 683238405Sjkim 684238405Sjkim *) Low level digest APIs are not approved in FIPS mode: any attempt 685238405Sjkim to use these will cause a fatal error. Applications that *really* want 686238405Sjkim to use them can use the private_* version instead. 687238405Sjkim [Steve Henson] 688238405Sjkim 689238405Sjkim *) Redirect cipher operations to FIPS module for FIPS builds. 690238405Sjkim [Steve Henson] 691238405Sjkim 692238405Sjkim *) Redirect digest operations to FIPS module for FIPS builds. 693238405Sjkim [Steve Henson] 694238405Sjkim 695238405Sjkim *) Update build system to add "fips" flag which will link in fipscanister.o 696238405Sjkim for static and shared library builds embedding a signature if needed. 697238405Sjkim [Steve Henson] 698238405Sjkim 699238405Sjkim *) Output TLS supported curves in preference order instead of numerical 700238405Sjkim order. This is currently hardcoded for the highest order curves first. 701238405Sjkim This should be configurable so applications can judge speed vs strength. 702238405Sjkim [Steve Henson] 703238405Sjkim 704238405Sjkim *) Add TLS v1.2 server support for client authentication. 705238405Sjkim [Steve Henson] 706238405Sjkim 707238405Sjkim *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers 708238405Sjkim and enable MD5. 709238405Sjkim [Steve Henson] 710238405Sjkim 711238405Sjkim *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying 712238405Sjkim FIPS modules versions. 713238405Sjkim [Steve Henson] 714238405Sjkim 715238405Sjkim *) Add TLS v1.2 client side support for client authentication. Keep cache 716238405Sjkim of handshake records longer as we don't know the hash algorithm to use 717238405Sjkim until after the certificate request message is received. 718238405Sjkim [Steve Henson] 719238405Sjkim 720238405Sjkim *) Initial TLS v1.2 client support. Add a default signature algorithms 721238405Sjkim extension including all the algorithms we support. Parse new signature 722238405Sjkim format in client key exchange. Relax some ECC signing restrictions for 723238405Sjkim TLS v1.2 as indicated in RFC5246. 724238405Sjkim [Steve Henson] 725238405Sjkim 726238405Sjkim *) Add server support for TLS v1.2 signature algorithms extension. Switch 727238405Sjkim to new signature format when needed using client digest preference. 728238405Sjkim All server ciphersuites should now work correctly in TLS v1.2. No client 729238405Sjkim support yet and no support for client certificates. 730238405Sjkim [Steve Henson] 731238405Sjkim 732238405Sjkim *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch 733238405Sjkim to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based 734238405Sjkim ciphersuites. At present only RSA key exchange ciphersuites work with 735238405Sjkim TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete 736238405Sjkim SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods 737238405Sjkim and version checking. 738238405Sjkim [Steve Henson] 739238405Sjkim 740238405Sjkim *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled 741238405Sjkim with this defined it will not be affected by any changes to ssl internal 742238405Sjkim structures. Add several utility functions to allow openssl application 743238405Sjkim to work with OPENSSL_NO_SSL_INTERN defined. 744238405Sjkim [Steve Henson] 745238405Sjkim 746238405Sjkim *) Add SRP support. 747238405Sjkim [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie] 748238405Sjkim 749238405Sjkim *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id. 750238405Sjkim [Steve Henson] 751238405Sjkim 752238405Sjkim *) Permit abbreviated handshakes when renegotiating using the function 753238405Sjkim SSL_renegotiate_abbreviated(). 754238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 755238405Sjkim 756238405Sjkim *) Add call to ENGINE_register_all_complete() to 757238405Sjkim ENGINE_load_builtin_engines(), so some implementations get used 758238405Sjkim automatically instead of needing explicit application support. 759238405Sjkim [Steve Henson] 760238405Sjkim 761238405Sjkim *) Add support for TLS key exporter as described in RFC5705. 762238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson] 763238405Sjkim 764238405Sjkim *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only 765238405Sjkim a few changes are required: 766238405Sjkim 767238405Sjkim Add SSL_OP_NO_TLSv1_1 flag. 768238405Sjkim Add TLSv1_1 methods. 769238405Sjkim Update version checking logic to handle version 1.1. 770238405Sjkim Add explicit IV handling (ported from DTLS code). 771238405Sjkim Add command line options to s_client/s_server. 772238405Sjkim [Steve Henson] 773238405Sjkim 774264331Sjkim Changes between 1.0.0j and 1.0.0k [5 Feb 2013] 775264331Sjkim 776264331Sjkim *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 777264331Sjkim 778264331Sjkim This addresses the flaw in CBC record processing discovered by 779264331Sjkim Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 780264331Sjkim at: http://www.isg.rhul.ac.uk/tls/ 781264331Sjkim 782264331Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 783264331Sjkim Security Group at Royal Holloway, University of London 784264331Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 785264331Sjkim Emilia K�sper for the initial patch. 786264331Sjkim (CVE-2013-0169) 787264331Sjkim [Emilia K�sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 788264331Sjkim 789264331Sjkim *) Return an error when checking OCSP signatures when key is NULL. 790264331Sjkim This fixes a DoS attack. (CVE-2013-0166) 791264331Sjkim [Steve Henson] 792264331Sjkim 793264331Sjkim *) Call OCSP Stapling callback after ciphersuite has been chosen, so 794264331Sjkim the right response is stapled. Also change SSL_get_certificate() 795264331Sjkim so it returns the certificate actually sent. 796264331Sjkim See http://rt.openssl.org/Ticket/Display.html?id=2836. 797264331Sjkim (This is a backport) 798264331Sjkim [Rob Stradling <rob.stradling@comodo.com>] 799264331Sjkim 800264331Sjkim *) Fix possible deadlock when decoding public keys. 801264331Sjkim [Steve Henson] 802264331Sjkim 803264331Sjkim Changes between 1.0.0i and 1.0.0j [10 May 2012] 804264331Sjkim 805264331Sjkim [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after 806264331Sjkim OpenSSL 1.0.1.] 807264331Sjkim 808264331Sjkim *) Sanity check record length before skipping explicit IV in DTLS 809264331Sjkim to fix DoS attack. 810264331Sjkim 811264331Sjkim Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 812264331Sjkim fuzzing as a service testing platform. 813264331Sjkim (CVE-2012-2333) 814264331Sjkim [Steve Henson] 815264331Sjkim 816264331Sjkim *) Initialise tkeylen properly when encrypting CMS messages. 817264331Sjkim Thanks to Solar Designer of Openwall for reporting this issue. 818264331Sjkim [Steve Henson] 819264331Sjkim 820264331Sjkim Changes between 1.0.0h and 1.0.0i [19 Apr 2012] 821264331Sjkim 822264331Sjkim *) Check for potentially exploitable overflows in asn1_d2i_read_bio 823264331Sjkim BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 824264331Sjkim in CRYPTO_realloc_clean. 825264331Sjkim 826264331Sjkim Thanks to Tavis Ormandy, Google Security Team, for discovering this 827264331Sjkim issue and to Adam Langley <agl@chromium.org> for fixing it. 828264331Sjkim (CVE-2012-2110) 829264331Sjkim [Adam Langley (Google), Tavis Ormandy, Google Security Team] 830264331Sjkim 831238405Sjkim Changes between 1.0.0g and 1.0.0h [12 Mar 2012] 832238405Sjkim 833237657Sjkim *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 834237657Sjkim in CMS and PKCS7 code. When RSA decryption fails use a random key for 835237657Sjkim content decryption and always return the same error. Note: this attack 836237657Sjkim needs on average 2^20 messages so it only affects automated senders. The 837237657Sjkim old behaviour can be reenabled in the CMS code by setting the 838237657Sjkim CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 839237657Sjkim an MMA defence is not necessary. 840237657Sjkim Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 841237657Sjkim this issue. (CVE-2012-0884) 842237657Sjkim [Steve Henson] 843237657Sjkim 844237657Sjkim *) Fix CVE-2011-4619: make sure we really are receiving a 845237657Sjkim client hello before rejecting multiple SGC restarts. Thanks to 846237657Sjkim Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 847237657Sjkim [Steve Henson] 848237657Sjkim 849238405Sjkim Changes between 1.0.0f and 1.0.0g [18 Jan 2012] 850238405Sjkim 851238405Sjkim *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 852238405Sjkim Thanks to Antonio Martin, Enterprise Secure Access Research and 853238405Sjkim Development, Cisco Systems, Inc. for discovering this bug and 854238405Sjkim preparing a fix. (CVE-2012-0050) 855238405Sjkim [Antonio Martin] 856238405Sjkim 857238405Sjkim Changes between 1.0.0e and 1.0.0f [4 Jan 2012] 858238405Sjkim 859238405Sjkim *) Nadhem Alfardan and Kenny Paterson have discovered an extension 860238405Sjkim of the Vaudenay padding oracle attack on CBC mode encryption 861238405Sjkim which enables an efficient plaintext recovery attack against 862238405Sjkim the OpenSSL implementation of DTLS. Their attack exploits timing 863238405Sjkim differences arising during decryption processing. A research 864238405Sjkim paper describing this attack can be found at: 865238405Sjkim http://www.isg.rhul.ac.uk/~kp/dtls.pdf 866238405Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 867238405Sjkim Security Group at Royal Holloway, University of London 868238405Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 869238405Sjkim <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 870238405Sjkim for preparing the fix. (CVE-2011-4108) 871238405Sjkim [Robin Seggelmann, Michael Tuexen] 872238405Sjkim 873238405Sjkim *) Clear bytes used for block padding of SSL 3.0 records. 874238405Sjkim (CVE-2011-4576) 875238405Sjkim [Adam Langley (Google)] 876238405Sjkim 877238405Sjkim *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 878238405Sjkim Kadianakis <desnacked@gmail.com> for discovering this issue and 879238405Sjkim Adam Langley for preparing the fix. (CVE-2011-4619) 880238405Sjkim [Adam Langley (Google)] 881238405Sjkim 882238405Sjkim *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027) 883238405Sjkim [Andrey Kulikov <amdeich@gmail.com>] 884238405Sjkim 885238405Sjkim *) Prevent malformed RFC3779 data triggering an assertion failure. 886238405Sjkim Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 887238405Sjkim and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 888238405Sjkim [Rob Austein <sra@hactrn.net>] 889238405Sjkim 890238405Sjkim *) Improved PRNG seeding for VOS. 891238405Sjkim [Paul Green <Paul.Green@stratus.com>] 892238405Sjkim 893238405Sjkim *) Fix ssl_ciph.c set-up race. 894238405Sjkim [Adam Langley (Google)] 895238405Sjkim 896238405Sjkim *) Fix spurious failures in ecdsatest.c. 897238405Sjkim [Emilia K�sper (Google)] 898238405Sjkim 899238405Sjkim *) Fix the BIO_f_buffer() implementation (which was mixing different 900238405Sjkim interpretations of the '..._len' fields). 901238405Sjkim [Adam Langley (Google)] 902238405Sjkim 903238405Sjkim *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 904238405Sjkim BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 905238405Sjkim threads won't reuse the same blinding coefficients. 906238405Sjkim 907238405Sjkim This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 908238405Sjkim lock to call BN_BLINDING_invert_ex, and avoids one use of 909238405Sjkim BN_BLINDING_update for each BN_BLINDING structure (previously, 910238405Sjkim the last update always remained unused). 911238405Sjkim [Emilia K�sper (Google)] 912238405Sjkim 913238405Sjkim *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf. 914238405Sjkim [Bob Buckholz (Google)] 915238405Sjkim 916238405Sjkim Changes between 1.0.0d and 1.0.0e [6 Sep 2011] 917238405Sjkim 918238405Sjkim *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted 919238405Sjkim by initialising X509_STORE_CTX properly. (CVE-2011-3207) 920238405Sjkim [Kaspar Brand <ossl@velox.ch>] 921238405Sjkim 922238405Sjkim *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 923238405Sjkim for multi-threaded use of ECDH. (CVE-2011-3210) 924238405Sjkim [Adam Langley (Google)] 925238405Sjkim 926238405Sjkim *) Fix x509_name_ex_d2i memory leak on bad inputs. 927238405Sjkim [Bodo Moeller] 928238405Sjkim 929238405Sjkim *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check 930238405Sjkim signature public key algorithm by using OID xref utilities instead. 931238405Sjkim Before this you could only use some ECC ciphersuites with SHA1 only. 932238405Sjkim [Steve Henson] 933238405Sjkim 934238405Sjkim *) Add protection against ECDSA timing attacks as mentioned in the paper 935238405Sjkim by Billy Bob Brumley and Nicola Tuveri, see: 936238405Sjkim 937238405Sjkim http://eprint.iacr.org/2011/232.pdf 938238405Sjkim 939238405Sjkim [Billy Bob Brumley and Nicola Tuveri] 940238405Sjkim 941238405Sjkim Changes between 1.0.0c and 1.0.0d [8 Feb 2011] 942238405Sjkim 943238405Sjkim *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 944238405Sjkim [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 945238405Sjkim 946238405Sjkim *) Fix bug in string printing code: if *any* escaping is enabled we must 947238405Sjkim escape the escape character (backslash) or the resulting string is 948238405Sjkim ambiguous. 949238405Sjkim [Steve Henson] 950238405Sjkim 951238405Sjkim Changes between 1.0.0b and 1.0.0c [2 Dec 2010] 952238405Sjkim 953238405Sjkim *) Disable code workaround for ancient and obsolete Netscape browsers 954238405Sjkim and servers: an attacker can use it in a ciphersuite downgrade attack. 955238405Sjkim Thanks to Martin Rex for discovering this bug. CVE-2010-4180 956238405Sjkim [Steve Henson] 957238405Sjkim 958238405Sjkim *) Fixed J-PAKE implementation error, originally discovered by 959238405Sjkim Sebastien Martini, further info and confirmation from Stefan 960238405Sjkim Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 961238405Sjkim [Ben Laurie] 962238405Sjkim 963238405Sjkim Changes between 1.0.0a and 1.0.0b [16 Nov 2010] 964238405Sjkim 965238405Sjkim *) Fix extension code to avoid race conditions which can result in a buffer 966238405Sjkim overrun vulnerability: resumed sessions must not be modified as they can 967238405Sjkim be shared by multiple threads. CVE-2010-3864 968238405Sjkim [Steve Henson] 969238405Sjkim 970238405Sjkim *) Fix WIN32 build system to correctly link an ENGINE directory into 971238405Sjkim a DLL. 972238405Sjkim [Steve Henson] 973238405Sjkim 974238405Sjkim Changes between 1.0.0 and 1.0.0a [01 Jun 2010] 975238405Sjkim 976238405Sjkim *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover 977238405Sjkim (CVE-2010-1633) 978238405Sjkim [Steve Henson, Peter-Michael Hager <hager@dortmund.net>] 979238405Sjkim 980238405Sjkim Changes between 0.9.8n and 1.0.0 [29 Mar 2010] 981238405Sjkim 982238405Sjkim *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher 983238405Sjkim context. The operation can be customised via the ctrl mechanism in 984238405Sjkim case ENGINEs want to include additional functionality. 985238405Sjkim [Steve Henson] 986238405Sjkim 987238405Sjkim *) Tolerate yet another broken PKCS#8 key format: private key value negative. 988238405Sjkim [Steve Henson] 989238405Sjkim 990238405Sjkim *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to 991238405Sjkim output hashes compatible with older versions of OpenSSL. 992238405Sjkim [Willy Weisz <weisz@vcpc.univie.ac.at>] 993238405Sjkim 994238405Sjkim *) Fix compression algorithm handling: if resuming a session use the 995238405Sjkim compression algorithm of the resumed session instead of determining 996238405Sjkim it from client hello again. Don't allow server to change algorithm. 997238405Sjkim [Steve Henson] 998238405Sjkim 999238405Sjkim *) Add load_crls() function to apps tidying load_certs() too. Add option 1000238405Sjkim to verify utility to allow additional CRLs to be included. 1001238405Sjkim [Steve Henson] 1002238405Sjkim 1003238405Sjkim *) Update OCSP request code to permit adding custom headers to the request: 1004238405Sjkim some responders need this. 1005238405Sjkim [Steve Henson] 1006238405Sjkim 1007238405Sjkim *) The function EVP_PKEY_sign() returns <=0 on error: check return code 1008238405Sjkim correctly. 1009238405Sjkim [Julia Lawall <julia@diku.dk>] 1010238405Sjkim 1011238405Sjkim *) Update verify callback code in apps/s_cb.c and apps/verify.c, it 1012238405Sjkim needlessly dereferenced structures, used obsolete functions and 1013238405Sjkim didn't handle all updated verify codes correctly. 1014238405Sjkim [Steve Henson] 1015238405Sjkim 1016238405Sjkim *) Disable MD2 in the default configuration. 1017238405Sjkim [Steve Henson] 1018238405Sjkim 1019238405Sjkim *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to 1020238405Sjkim indicate the initial BIO being pushed or popped. This makes it possible 1021238405Sjkim to determine whether the BIO is the one explicitly called or as a result 1022238405Sjkim of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so 1023238405Sjkim it handles reference counts correctly and doesn't zero out the I/O bio 1024238405Sjkim when it is not being explicitly popped. WARNING: applications which 1025238405Sjkim included workarounds for the old buggy behaviour will need to be modified 1026238405Sjkim or they could free up already freed BIOs. 1027238405Sjkim [Steve Henson] 1028238405Sjkim 1029238405Sjkim *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni 1030238405Sjkim renaming to all platforms (within the 0.9.8 branch, this was 1031238405Sjkim done conditionally on Netware platforms to avoid a name clash). 1032238405Sjkim [Guenter <lists@gknw.net>] 1033238405Sjkim 1034238405Sjkim *) Add ECDHE and PSK support to DTLS. 1035238405Sjkim [Michael Tuexen <tuexen@fh-muenster.de>] 1036238405Sjkim 1037238405Sjkim *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't 1038238405Sjkim be used on C++. 1039238405Sjkim [Steve Henson] 1040238405Sjkim 1041238405Sjkim *) Add "missing" function EVP_MD_flags() (without this the only way to 1042238405Sjkim retrieve a digest flags is by accessing the structure directly. Update 1043238405Sjkim EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest 1044238405Sjkim or cipher is registered as in the "from" argument. Print out all 1045238405Sjkim registered digests in the dgst usage message instead of manually 1046238405Sjkim attempting to work them out. 1047238405Sjkim [Steve Henson] 1048238405Sjkim 1049238405Sjkim *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello: 1050238405Sjkim this allows the use of compression and extensions. Change default cipher 1051238405Sjkim string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2 1052238405Sjkim by default unless an application cipher string requests it. 1053238405Sjkim [Steve Henson] 1054238405Sjkim 1055238405Sjkim *) Alter match criteria in PKCS12_parse(). It used to try to use local 1056238405Sjkim key ids to find matching certificates and keys but some PKCS#12 files 1057238405Sjkim don't follow the (somewhat unwritten) rules and this strategy fails. 1058238405Sjkim Now just gather all certificates together and the first private key 1059238405Sjkim then look for the first certificate that matches the key. 1060238405Sjkim [Steve Henson] 1061238405Sjkim 1062238405Sjkim *) Support use of registered digest and cipher names for dgst and cipher 1063238405Sjkim commands instead of having to add each one as a special case. So now 1064238405Sjkim you can do: 1065238405Sjkim 1066238405Sjkim openssl sha256 foo 1067238405Sjkim 1068238405Sjkim as well as: 1069238405Sjkim 1070238405Sjkim openssl dgst -sha256 foo 1071238405Sjkim 1072238405Sjkim and this works for ENGINE based algorithms too. 1073238405Sjkim 1074238405Sjkim [Steve Henson] 1075238405Sjkim 1076238405Sjkim *) Update Gost ENGINE to support parameter files. 1077238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 1078238405Sjkim 1079238405Sjkim *) Support GeneralizedTime in ca utility. 1080238405Sjkim [Oliver Martin <oliver@volatilevoid.net>, Steve Henson] 1081238405Sjkim 1082238405Sjkim *) Enhance the hash format used for certificate directory links. The new 1083238405Sjkim form uses the canonical encoding (meaning equivalent names will work 1084238405Sjkim even if they aren't identical) and uses SHA1 instead of MD5. This form 1085238405Sjkim is incompatible with the older format and as a result c_rehash should 1086238405Sjkim be used to rebuild symbolic links. 1087238405Sjkim [Steve Henson] 1088238405Sjkim 1089238405Sjkim *) Make PKCS#8 the default write format for private keys, replacing the 1090238405Sjkim traditional format. This form is standardised, more secure and doesn't 1091238405Sjkim include an implicit MD5 dependency. 1092238405Sjkim [Steve Henson] 1093238405Sjkim 1094238405Sjkim *) Add a $gcc_devteam_warn option to Configure. The idea is that any code 1095238405Sjkim committed to OpenSSL should pass this lot as a minimum. 1096238405Sjkim [Steve Henson] 1097238405Sjkim 1098238405Sjkim *) Add session ticket override functionality for use by EAP-FAST. 1099238405Sjkim [Jouni Malinen <j@w1.fi>] 1100238405Sjkim 1101238405Sjkim *) Modify HMAC functions to return a value. Since these can be implemented 1102238405Sjkim in an ENGINE errors can occur. 1103238405Sjkim [Steve Henson] 1104238405Sjkim 1105238405Sjkim *) Type-checked OBJ_bsearch_ex. 1106238405Sjkim [Ben Laurie] 1107238405Sjkim 1108238405Sjkim *) Type-checked OBJ_bsearch. Also some constification necessitated 1109238405Sjkim by type-checking. Still to come: TXT_DB, bsearch(?), 1110238405Sjkim OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING, 1111238405Sjkim CONF_VALUE. 1112238405Sjkim [Ben Laurie] 1113238405Sjkim 1114238405Sjkim *) New function OPENSSL_gmtime_adj() to add a specific number of days and 1115238405Sjkim seconds to a tm structure directly, instead of going through OS 1116238405Sjkim specific date routines. This avoids any issues with OS routines such 1117238405Sjkim as the year 2038 bug. New *_adj() functions for ASN1 time structures 1118238405Sjkim and X509_time_adj_ex() to cover the extended range. The existing 1119238405Sjkim X509_time_adj() is still usable and will no longer have any date issues. 1120238405Sjkim [Steve Henson] 1121238405Sjkim 1122238405Sjkim *) Delta CRL support. New use deltas option which will attempt to locate 1123238405Sjkim and search any appropriate delta CRLs available. 1124238405Sjkim 1125238405Sjkim This work was sponsored by Google. 1126238405Sjkim [Steve Henson] 1127238405Sjkim 1128238405Sjkim *) Support for CRLs partitioned by reason code. Reorganise CRL processing 1129238405Sjkim code and add additional score elements. Validate alternate CRL paths 1130238405Sjkim as part of the CRL checking and indicate a new error "CRL path validation 1131238405Sjkim error" in this case. Applications wanting additional details can use 1132238405Sjkim the verify callback and check the new "parent" field. If this is not 1133238405Sjkim NULL CRL path validation is taking place. Existing applications wont 1134238405Sjkim see this because it requires extended CRL support which is off by 1135238405Sjkim default. 1136238405Sjkim 1137238405Sjkim This work was sponsored by Google. 1138238405Sjkim [Steve Henson] 1139238405Sjkim 1140238405Sjkim *) Support for freshest CRL extension. 1141238405Sjkim 1142238405Sjkim This work was sponsored by Google. 1143238405Sjkim [Steve Henson] 1144238405Sjkim 1145238405Sjkim *) Initial indirect CRL support. Currently only supported in the CRLs 1146238405Sjkim passed directly and not via lookup. Process certificate issuer 1147238405Sjkim CRL entry extension and lookup CRL entries by bother issuer name 1148238405Sjkim and serial number. Check and process CRL issuer entry in IDP extension. 1149238405Sjkim 1150238405Sjkim This work was sponsored by Google. 1151238405Sjkim [Steve Henson] 1152238405Sjkim 1153238405Sjkim *) Add support for distinct certificate and CRL paths. The CRL issuer 1154238405Sjkim certificate is validated separately in this case. Only enabled if 1155238405Sjkim an extended CRL support flag is set: this flag will enable additional 1156238405Sjkim CRL functionality in future. 1157238405Sjkim 1158238405Sjkim This work was sponsored by Google. 1159238405Sjkim [Steve Henson] 1160238405Sjkim 1161238405Sjkim *) Add support for policy mappings extension. 1162238405Sjkim 1163238405Sjkim This work was sponsored by Google. 1164238405Sjkim [Steve Henson] 1165238405Sjkim 1166238405Sjkim *) Fixes to pathlength constraint, self issued certificate handling, 1167238405Sjkim policy processing to align with RFC3280 and PKITS tests. 1168238405Sjkim 1169238405Sjkim This work was sponsored by Google. 1170238405Sjkim [Steve Henson] 1171238405Sjkim 1172238405Sjkim *) Support for name constraints certificate extension. DN, email, DNS 1173238405Sjkim and URI types are currently supported. 1174238405Sjkim 1175238405Sjkim This work was sponsored by Google. 1176238405Sjkim [Steve Henson] 1177238405Sjkim 1178238405Sjkim *) To cater for systems that provide a pointer-based thread ID rather 1179238405Sjkim than numeric, deprecate the current numeric thread ID mechanism and 1180238405Sjkim replace it with a structure and associated callback type. This 1181238405Sjkim mechanism allows a numeric "hash" to be extracted from a thread ID in 1182238405Sjkim either case, and on platforms where pointers are larger than 'long', 1183238405Sjkim mixing is done to help ensure the numeric 'hash' is usable even if it 1184238405Sjkim can't be guaranteed unique. The default mechanism is to use "&errno" 1185238405Sjkim as a pointer-based thread ID to distinguish between threads. 1186238405Sjkim 1187238405Sjkim Applications that want to provide their own thread IDs should now use 1188238405Sjkim CRYPTO_THREADID_set_callback() to register a callback that will call 1189238405Sjkim either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer(). 1190238405Sjkim 1191238405Sjkim Note that ERR_remove_state() is now deprecated, because it is tied 1192238405Sjkim to the assumption that thread IDs are numeric. ERR_remove_state(0) 1193238405Sjkim to free the current thread's error state should be replaced by 1194238405Sjkim ERR_remove_thread_state(NULL). 1195238405Sjkim 1196238405Sjkim (This new approach replaces the functions CRYPTO_set_idptr_callback(), 1197238405Sjkim CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in 1198238405Sjkim OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an 1199238405Sjkim application was previously providing a numeric thread callback that 1200238405Sjkim was inappropriate for distinguishing threads, then uniqueness might 1201238405Sjkim have been obtained with &errno that happened immediately in the 1202238405Sjkim intermediate development versions of OpenSSL; this is no longer the 1203238405Sjkim case, the numeric thread callback will now override the automatic use 1204238405Sjkim of &errno.) 1205238405Sjkim [Geoff Thorpe, with help from Bodo Moeller] 1206238405Sjkim 1207238405Sjkim *) Initial support for different CRL issuing certificates. This covers a 1208238405Sjkim simple case where the self issued certificates in the chain exist and 1209238405Sjkim the real CRL issuer is higher in the existing chain. 1210238405Sjkim 1211238405Sjkim This work was sponsored by Google. 1212238405Sjkim [Steve Henson] 1213238405Sjkim 1214238405Sjkim *) Removed effectively defunct crypto/store from the build. 1215238405Sjkim [Ben Laurie] 1216238405Sjkim 1217238405Sjkim *) Revamp of STACK to provide stronger type-checking. Still to come: 1218238405Sjkim TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE, 1219238405Sjkim ASN1_STRING, CONF_VALUE. 1220238405Sjkim [Ben Laurie] 1221238405Sjkim 1222238405Sjkim *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer 1223238405Sjkim RAM on SSL connections. This option can save about 34k per idle SSL. 1224238405Sjkim [Nick Mathewson] 1225238405Sjkim 1226238405Sjkim *) Revamp of LHASH to provide stronger type-checking. Still to come: 1227238405Sjkim STACK, TXT_DB, bsearch, qsort. 1228238405Sjkim [Ben Laurie] 1229238405Sjkim 1230238405Sjkim *) Initial support for Cryptographic Message Syntax (aka CMS) based 1231238405Sjkim on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility, 1232238405Sjkim support for data, signedData, compressedData, digestedData and 1233238405Sjkim encryptedData, envelopedData types included. Scripts to check against 1234238405Sjkim RFC4134 examples draft and interop and consistency checks of many 1235238405Sjkim content types and variants. 1236238405Sjkim [Steve Henson] 1237238405Sjkim 1238238405Sjkim *) Add options to enc utility to support use of zlib compression BIO. 1239238405Sjkim [Steve Henson] 1240238405Sjkim 1241238405Sjkim *) Extend mk1mf to support importing of options and assembly language 1242238405Sjkim files from Configure script, currently only included in VC-WIN32. 1243238405Sjkim The assembly language rules can now optionally generate the source 1244238405Sjkim files from the associated perl scripts. 1245238405Sjkim [Steve Henson] 1246238405Sjkim 1247238405Sjkim *) Implement remaining functionality needed to support GOST ciphersuites. 1248238405Sjkim Interop testing has been performed using CryptoPro implementations. 1249238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 1250238405Sjkim 1251238405Sjkim *) s390x assembler pack. 1252238405Sjkim [Andy Polyakov] 1253238405Sjkim 1254238405Sjkim *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU 1255238405Sjkim "family." 1256238405Sjkim [Andy Polyakov] 1257238405Sjkim 1258238405Sjkim *) Implement Opaque PRF Input TLS extension as specified in 1259238405Sjkim draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an 1260238405Sjkim official specification yet and no extension type assignment by 1261238405Sjkim IANA exists, this extension (for now) will have to be explicitly 1262238405Sjkim enabled when building OpenSSL by providing the extension number 1263238405Sjkim to use. For example, specify an option 1264238405Sjkim 1265238405Sjkim -DTLSEXT_TYPE_opaque_prf_input=0x9527 1266238405Sjkim 1267238405Sjkim to the "config" or "Configure" script to enable the extension, 1268238405Sjkim assuming extension number 0x9527 (which is a completely arbitrary 1269238405Sjkim and unofficial assignment based on the MD5 hash of the Internet 1270238405Sjkim Draft). Note that by doing so, you potentially lose 1271238405Sjkim interoperability with other TLS implementations since these might 1272238405Sjkim be using the same extension number for other purposes. 1273238405Sjkim 1274238405Sjkim SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the 1275238405Sjkim opaque PRF input value to use in the handshake. This will create 1276238405Sjkim an interal copy of the length-'len' string at 'src', and will 1277238405Sjkim return non-zero for success. 1278238405Sjkim 1279238405Sjkim To get more control and flexibility, provide a callback function 1280238405Sjkim by using 1281238405Sjkim 1282238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb) 1283238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg) 1284238405Sjkim 1285238405Sjkim where 1286238405Sjkim 1287238405Sjkim int (*cb)(SSL *, void *peerinput, size_t len, void *arg); 1288238405Sjkim void *arg; 1289238405Sjkim 1290238405Sjkim Callback function 'cb' will be called in handshakes, and is 1291238405Sjkim expected to use SSL_set_tlsext_opaque_prf_input() as appropriate. 1292238405Sjkim Argument 'arg' is for application purposes (the value as given to 1293238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly 1294238405Sjkim be provided to the callback function). The callback function 1295238405Sjkim has to return non-zero to report success: usually 1 to use opaque 1296238405Sjkim PRF input just if possible, or 2 to enforce use of the opaque PRF 1297238405Sjkim input. In the latter case, the library will abort the handshake 1298238405Sjkim if opaque PRF input is not successfully negotiated. 1299238405Sjkim 1300238405Sjkim Arguments 'peerinput' and 'len' given to the callback function 1301238405Sjkim will always be NULL and 0 in the case of a client. A server will 1302238405Sjkim see the client's opaque PRF input through these variables if 1303238405Sjkim available (NULL and 0 otherwise). Note that if the server 1304238405Sjkim provides an opaque PRF input, the length must be the same as the 1305238405Sjkim length of the client's opaque PRF input. 1306238405Sjkim 1307238405Sjkim Note that the callback function will only be called when creating 1308238405Sjkim a new session (session resumption can resume whatever was 1309238405Sjkim previously negotiated), and will not be called in SSL 2.0 1310238405Sjkim handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or 1311238405Sjkim SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended 1312238405Sjkim for applications that need to enforce opaque PRF input. 1313238405Sjkim 1314238405Sjkim [Bodo Moeller] 1315238405Sjkim 1316238405Sjkim *) Update ssl code to support digests other than SHA1+MD5 for handshake 1317238405Sjkim MAC. 1318238405Sjkim 1319238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 1320238405Sjkim 1321238405Sjkim *) Add RFC4507 support to OpenSSL. This includes the corrections in 1322238405Sjkim RFC4507bis. The encrypted ticket format is an encrypted encoded 1323238405Sjkim SSL_SESSION structure, that way new session features are automatically 1324238405Sjkim supported. 1325238405Sjkim 1326238405Sjkim If a client application caches session in an SSL_SESSION structure 1327238405Sjkim support is transparent because tickets are now stored in the encoded 1328238405Sjkim SSL_SESSION. 1329238405Sjkim 1330238405Sjkim The SSL_CTX structure automatically generates keys for ticket 1331238405Sjkim protection in servers so again support should be possible 1332238405Sjkim with no application modification. 1333238405Sjkim 1334238405Sjkim If a client or server wishes to disable RFC4507 support then the option 1335238405Sjkim SSL_OP_NO_TICKET can be set. 1336238405Sjkim 1337238405Sjkim Add a TLS extension debugging callback to allow the contents of any client 1338238405Sjkim or server extensions to be examined. 1339238405Sjkim 1340238405Sjkim This work was sponsored by Google. 1341238405Sjkim [Steve Henson] 1342238405Sjkim 1343238405Sjkim *) Final changes to avoid use of pointer pointer casts in OpenSSL. 1344238405Sjkim OpenSSL should now compile cleanly on gcc 4.2 1345238405Sjkim [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson] 1346238405Sjkim 1347238405Sjkim *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC 1348238405Sjkim support including streaming MAC support: this is required for GOST 1349238405Sjkim ciphersuite support. 1350238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson] 1351238405Sjkim 1352238405Sjkim *) Add option -stream to use PKCS#7 streaming in smime utility. New 1353238405Sjkim function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream() 1354238405Sjkim to output in BER and PEM format. 1355238405Sjkim [Steve Henson] 1356238405Sjkim 1357238405Sjkim *) Experimental support for use of HMAC via EVP_PKEY interface. This 1358238405Sjkim allows HMAC to be handled via the EVP_DigestSign*() interface. The 1359238405Sjkim EVP_PKEY "key" in this case is the HMAC key, potentially allowing 1360238405Sjkim ENGINE support for HMAC keys which are unextractable. New -mac and 1361238405Sjkim -macopt options to dgst utility. 1362238405Sjkim [Steve Henson] 1363238405Sjkim 1364238405Sjkim *) New option -sigopt to dgst utility. Update dgst to use 1365238405Sjkim EVP_Digest{Sign,Verify}*. These two changes make it possible to use 1366238405Sjkim alternative signing paramaters such as X9.31 or PSS in the dgst 1367238405Sjkim utility. 1368238405Sjkim [Steve Henson] 1369238405Sjkim 1370238405Sjkim *) Change ssl_cipher_apply_rule(), the internal function that does 1371238405Sjkim the work each time a ciphersuite string requests enabling 1372238405Sjkim ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or 1373238405Sjkim removing ("!foo+bar") a class of ciphersuites: Now it maintains 1374238405Sjkim the order of disabled ciphersuites such that those ciphersuites 1375238405Sjkim that most recently went from enabled to disabled not only stay 1376238405Sjkim in order with respect to each other, but also have higher priority 1377238405Sjkim than other disabled ciphersuites the next time ciphersuites are 1378238405Sjkim enabled again. 1379238405Sjkim 1380238405Sjkim This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable 1381238405Sjkim the same ciphersuites as with "HIGH" alone, but in a specific 1382238405Sjkim order where the PSK ciphersuites come first (since they are the 1383238405Sjkim most recently disabled ciphersuites when "HIGH" is parsed). 1384238405Sjkim 1385238405Sjkim Also, change ssl_create_cipher_list() (using this new 1386238405Sjkim funcionality) such that between otherwise identical 1387238405Sjkim cihpersuites, ephemeral ECDH is preferred over ephemeral DH in 1388238405Sjkim the default order. 1389238405Sjkim [Bodo Moeller] 1390238405Sjkim 1391238405Sjkim *) Change ssl_create_cipher_list() so that it automatically 1392238405Sjkim arranges the ciphersuites in reasonable order before starting 1393238405Sjkim to process the rule string. Thus, the definition for "DEFAULT" 1394238405Sjkim (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but 1395238405Sjkim remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH". 1396238405Sjkim This makes it much easier to arrive at a reasonable default order 1397238405Sjkim in applications for which anonymous ciphers are OK (meaning 1398238405Sjkim that you can't actually use DEFAULT). 1399238405Sjkim [Bodo Moeller; suggested by Victor Duchovni] 1400238405Sjkim 1401238405Sjkim *) Split the SSL/TLS algorithm mask (as used for ciphersuite string 1402238405Sjkim processing) into multiple integers instead of setting 1403238405Sjkim "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK", 1404238405Sjkim "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer. 1405238405Sjkim (These masks as well as the individual bit definitions are hidden 1406238405Sjkim away into the non-exported interface ssl/ssl_locl.h, so this 1407238405Sjkim change to the definition of the SSL_CIPHER structure shouldn't 1408238405Sjkim affect applications.) This give us more bits for each of these 1409238405Sjkim categories, so there is no longer a need to coagulate AES128 and 1410238405Sjkim AES256 into a single algorithm bit, and to coagulate Camellia128 1411238405Sjkim and Camellia256 into a single algorithm bit, which has led to all 1412238405Sjkim kinds of kludges. 1413238405Sjkim 1414238405Sjkim Thus, among other things, the kludge introduced in 0.9.7m and 1415238405Sjkim 0.9.8e for masking out AES256 independently of AES128 or masking 1416238405Sjkim out Camellia256 independently of AES256 is not needed here in 0.9.9. 1417238405Sjkim 1418238405Sjkim With the change, we also introduce new ciphersuite aliases that 1419238405Sjkim so far were missing: "AES128", "AES256", "CAMELLIA128", and 1420238405Sjkim "CAMELLIA256". 1421238405Sjkim [Bodo Moeller] 1422238405Sjkim 1423238405Sjkim *) Add support for dsa-with-SHA224 and dsa-with-SHA256. 1424238405Sjkim Use the leftmost N bytes of the signature input if the input is 1425238405Sjkim larger than the prime q (with N being the size in bytes of q). 1426238405Sjkim [Nils Larsch] 1427238405Sjkim 1428238405Sjkim *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses 1429238405Sjkim it yet and it is largely untested. 1430238405Sjkim [Steve Henson] 1431238405Sjkim 1432238405Sjkim *) Add support for the ecdsa-with-SHA224/256/384/512 signature types. 1433238405Sjkim [Nils Larsch] 1434238405Sjkim 1435238405Sjkim *) Initial incomplete changes to avoid need for function casts in OpenSSL 1436238405Sjkim some compilers (gcc 4.2 and later) reject their use. Safestack is 1437238405Sjkim reimplemented. Update ASN1 to avoid use of legacy functions. 1438238405Sjkim [Steve Henson] 1439238405Sjkim 1440238405Sjkim *) Win32/64 targets are linked with Winsock2. 1441238405Sjkim [Andy Polyakov] 1442238405Sjkim 1443238405Sjkim *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected 1444238405Sjkim to external functions. This can be used to increase CRL handling 1445238405Sjkim efficiency especially when CRLs are very large by (for example) storing 1446238405Sjkim the CRL revoked certificates in a database. 1447238405Sjkim [Steve Henson] 1448238405Sjkim 1449238405Sjkim *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so 1450238405Sjkim new CRLs added to a directory can be used. New command line option 1451238405Sjkim -verify_return_error to s_client and s_server. This causes real errors 1452238405Sjkim to be returned by the verify callback instead of carrying on no matter 1453238405Sjkim what. This reflects the way a "real world" verify callback would behave. 1454238405Sjkim [Steve Henson] 1455238405Sjkim 1456238405Sjkim *) GOST engine, supporting several GOST algorithms and public key formats. 1457238405Sjkim Kindly donated by Cryptocom. 1458238405Sjkim [Cryptocom] 1459238405Sjkim 1460238405Sjkim *) Partial support for Issuing Distribution Point CRL extension. CRLs 1461238405Sjkim partitioned by DP are handled but no indirect CRL or reason partitioning 1462238405Sjkim (yet). Complete overhaul of CRL handling: now the most suitable CRL is 1463238405Sjkim selected via a scoring technique which handles IDP and AKID in CRLs. 1464238405Sjkim [Steve Henson] 1465238405Sjkim 1466238405Sjkim *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which 1467238405Sjkim will ultimately be used for all verify operations: this will remove the 1468238405Sjkim X509_STORE dependency on certificate verification and allow alternative 1469238405Sjkim lookup methods. X509_STORE based implementations of these two callbacks. 1470238405Sjkim [Steve Henson] 1471238405Sjkim 1472238405Sjkim *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names. 1473238405Sjkim Modify get_crl() to find a valid (unexpired) CRL if possible. 1474238405Sjkim [Steve Henson] 1475238405Sjkim 1476238405Sjkim *) New function X509_CRL_match() to check if two CRLs are identical. Normally 1477238405Sjkim this would be called X509_CRL_cmp() but that name is already used by 1478238405Sjkim a function that just compares CRL issuer names. Cache several CRL 1479238405Sjkim extensions in X509_CRL structure and cache CRLDP in X509. 1480238405Sjkim [Steve Henson] 1481238405Sjkim 1482238405Sjkim *) Store a "canonical" representation of X509_NAME structure (ASN1 Name) 1483238405Sjkim this maps equivalent X509_NAME structures into a consistent structure. 1484238405Sjkim Name comparison can then be performed rapidly using memcmp(). 1485238405Sjkim [Steve Henson] 1486238405Sjkim 1487238405Sjkim *) Non-blocking OCSP request processing. Add -timeout option to ocsp 1488238405Sjkim utility. 1489238405Sjkim [Steve Henson] 1490238405Sjkim 1491238405Sjkim *) Allow digests to supply their own micalg string for S/MIME type using 1492238405Sjkim the ctrl EVP_MD_CTRL_MICALG. 1493238405Sjkim [Steve Henson] 1494238405Sjkim 1495238405Sjkim *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the 1496238405Sjkim EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN 1497238405Sjkim ctrl. It can then customise the structure before and/or after signing 1498238405Sjkim if necessary. 1499238405Sjkim [Steve Henson] 1500238405Sjkim 1501238405Sjkim *) New function OBJ_add_sigid() to allow application defined signature OIDs 1502238405Sjkim to be added to OpenSSLs internal tables. New function OBJ_sigid_free() 1503238405Sjkim to free up any added signature OIDs. 1504238405Sjkim [Steve Henson] 1505238405Sjkim 1506238405Sjkim *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(), 1507238405Sjkim EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal 1508238405Sjkim digest and cipher tables. New options added to openssl utility: 1509238405Sjkim list-message-digest-algorithms and list-cipher-algorithms. 1510238405Sjkim [Steve Henson] 1511238405Sjkim 1512238405Sjkim *) Change the array representation of binary polynomials: the list 1513238405Sjkim of degrees of non-zero coefficients is now terminated with -1. 1514238405Sjkim Previously it was terminated with 0, which was also part of the 1515238405Sjkim value; thus, the array representation was not applicable to 1516238405Sjkim polynomials where t^0 has coefficient zero. This change makes 1517238405Sjkim the array representation useful in a more general context. 1518238405Sjkim [Douglas Stebila] 1519238405Sjkim 1520238405Sjkim *) Various modifications and fixes to SSL/TLS cipher string 1521238405Sjkim handling. For ECC, the code now distinguishes between fixed ECDH 1522238405Sjkim with RSA certificates on the one hand and with ECDSA certificates 1523238405Sjkim on the other hand, since these are separate ciphersuites. The 1524238405Sjkim unused code for Fortezza ciphersuites has been removed. 1525238405Sjkim 1526238405Sjkim For consistency with EDH, ephemeral ECDH is now called "EECDH" 1527238405Sjkim (not "ECDHE"). For consistency with the code for DH 1528238405Sjkim certificates, use of ECDH certificates is now considered ECDH 1529238405Sjkim authentication, not RSA or ECDSA authentication (the latter is 1530238405Sjkim merely the CA's signing algorithm and not actively used in the 1531238405Sjkim protocol). 1532238405Sjkim 1533238405Sjkim The temporary ciphersuite alias "ECCdraft" is no longer 1534238405Sjkim available, and ECC ciphersuites are no longer excluded from "ALL" 1535238405Sjkim and "DEFAULT". The following aliases now exist for RFC 4492 1536238405Sjkim ciphersuites, most of these by analogy with the DH case: 1537238405Sjkim 1538238405Sjkim kECDHr - ECDH cert, signed with RSA 1539238405Sjkim kECDHe - ECDH cert, signed with ECDSA 1540238405Sjkim kECDH - ECDH cert (signed with either RSA or ECDSA) 1541238405Sjkim kEECDH - ephemeral ECDH 1542238405Sjkim ECDH - ECDH cert or ephemeral ECDH 1543238405Sjkim 1544238405Sjkim aECDH - ECDH cert 1545238405Sjkim aECDSA - ECDSA cert 1546238405Sjkim ECDSA - ECDSA cert 1547238405Sjkim 1548238405Sjkim AECDH - anonymous ECDH 1549238405Sjkim EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH") 1550238405Sjkim 1551238405Sjkim [Bodo Moeller] 1552238405Sjkim 1553238405Sjkim *) Add additional S/MIME capabilities for AES and GOST ciphers if supported. 1554238405Sjkim Use correct micalg parameters depending on digest(s) in signed message. 1555238405Sjkim [Steve Henson] 1556238405Sjkim 1557238405Sjkim *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process 1558238405Sjkim an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code. 1559238405Sjkim [Steve Henson] 1560238405Sjkim 1561238405Sjkim *) Initial engine support for EVP_PKEY_METHOD. New functions to permit 1562238405Sjkim an engine to register a method. Add ENGINE lookups for methods and 1563238405Sjkim functional reference processing. 1564238405Sjkim [Steve Henson] 1565238405Sjkim 1566238405Sjkim *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of 1567238405Sjkim EVP_{Sign,Verify}* which allow an application to customise the signature 1568238405Sjkim process. 1569238405Sjkim [Steve Henson] 1570238405Sjkim 1571238405Sjkim *) New -resign option to smime utility. This adds one or more signers 1572238405Sjkim to an existing PKCS#7 signedData structure. Also -md option to use an 1573238405Sjkim alternative message digest algorithm for signing. 1574238405Sjkim [Steve Henson] 1575238405Sjkim 1576238405Sjkim *) Tidy up PKCS#7 routines and add new functions to make it easier to 1577238405Sjkim create PKCS7 structures containing multiple signers. Update smime 1578238405Sjkim application to support multiple signers. 1579238405Sjkim [Steve Henson] 1580238405Sjkim 1581238405Sjkim *) New -macalg option to pkcs12 utility to allow setting of an alternative 1582238405Sjkim digest MAC. 1583238405Sjkim [Steve Henson] 1584238405Sjkim 1585238405Sjkim *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC. 1586238405Sjkim Reorganize PBE internals to lookup from a static table using NIDs, 1587238405Sjkim add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl: 1588238405Sjkim EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative 1589238405Sjkim PRF which will be automatically used with PBES2. 1590238405Sjkim [Steve Henson] 1591238405Sjkim 1592238405Sjkim *) Replace the algorithm specific calls to generate keys in "req" with the 1593238405Sjkim new API. 1594238405Sjkim [Steve Henson] 1595238405Sjkim 1596238405Sjkim *) Update PKCS#7 enveloped data routines to use new API. This is now 1597238405Sjkim supported by any public key method supporting the encrypt operation. A 1598238405Sjkim ctrl is added to allow the public key algorithm to examine or modify 1599238405Sjkim the PKCS#7 RecipientInfo structure if it needs to: for RSA this is 1600238405Sjkim a no op. 1601238405Sjkim [Steve Henson] 1602238405Sjkim 1603238405Sjkim *) Add a ctrl to asn1 method to allow a public key algorithm to express 1604238405Sjkim a default digest type to use. In most cases this will be SHA1 but some 1605238405Sjkim algorithms (such as GOST) need to specify an alternative digest. The 1606238405Sjkim return value indicates how strong the prefernce is 1 means optional and 1607238405Sjkim 2 is mandatory (that is it is the only supported type). Modify 1608238405Sjkim ASN1_item_sign() to accept a NULL digest argument to indicate it should 1609238405Sjkim use the default md. Update openssl utilities to use the default digest 1610238405Sjkim type for signing if it is not explicitly indicated. 1611238405Sjkim [Steve Henson] 1612238405Sjkim 1613238405Sjkim *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New 1614238405Sjkim EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant 1615238405Sjkim signing method from the key type. This effectively removes the link 1616238405Sjkim between digests and public key types. 1617238405Sjkim [Steve Henson] 1618238405Sjkim 1619238405Sjkim *) Add an OID cross reference table and utility functions. Its purpose is to 1620238405Sjkim translate between signature OIDs such as SHA1WithrsaEncryption and SHA1, 1621238405Sjkim rsaEncryption. This will allow some of the algorithm specific hackery 1622238405Sjkim needed to use the correct OID to be removed. 1623238405Sjkim [Steve Henson] 1624238405Sjkim 1625238405Sjkim *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO 1626238405Sjkim structures for PKCS7_sign(). They are now set up by the relevant public 1627238405Sjkim key ASN1 method. 1628238405Sjkim [Steve Henson] 1629238405Sjkim 1630238405Sjkim *) Add provisional EC pkey method with support for ECDSA and ECDH. 1631238405Sjkim [Steve Henson] 1632238405Sjkim 1633238405Sjkim *) Add support for key derivation (agreement) in the API, DH method and 1634238405Sjkim pkeyutl. 1635238405Sjkim [Steve Henson] 1636238405Sjkim 1637238405Sjkim *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support 1638238405Sjkim public and private key formats. As a side effect these add additional 1639238405Sjkim command line functionality not previously available: DSA signatures can be 1640238405Sjkim generated and verified using pkeyutl and DH key support and generation in 1641238405Sjkim pkey, genpkey. 1642238405Sjkim [Steve Henson] 1643238405Sjkim 1644238405Sjkim *) BeOS support. 1645238405Sjkim [Oliver Tappe <zooey@hirschkaefer.de>] 1646238405Sjkim 1647238405Sjkim *) New make target "install_html_docs" installs HTML renditions of the 1648238405Sjkim manual pages. 1649238405Sjkim [Oliver Tappe <zooey@hirschkaefer.de>] 1650238405Sjkim 1651238405Sjkim *) New utility "genpkey" this is analagous to "genrsa" etc except it can 1652238405Sjkim generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to 1653238405Sjkim support key and parameter generation and add initial key generation 1654238405Sjkim functionality for RSA. 1655238405Sjkim [Steve Henson] 1656238405Sjkim 1657238405Sjkim *) Add functions for main EVP_PKEY_method operations. The undocumented 1658238405Sjkim functions EVP_PKEY_{encrypt,decrypt} have been renamed to 1659238405Sjkim EVP_PKEY_{encrypt,decrypt}_old. 1660238405Sjkim [Steve Henson] 1661238405Sjkim 1662238405Sjkim *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public 1663238405Sjkim key API, doesn't do much yet. 1664238405Sjkim [Steve Henson] 1665238405Sjkim 1666238405Sjkim *) New function EVP_PKEY_asn1_get0_info() to retrieve information about 1667238405Sjkim public key algorithms. New option to openssl utility: 1668238405Sjkim "list-public-key-algorithms" to print out info. 1669238405Sjkim [Steve Henson] 1670238405Sjkim 1671238405Sjkim *) Implement the Supported Elliptic Curves Extension for 1672238405Sjkim ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1673238405Sjkim [Douglas Stebila] 1674238405Sjkim 1675238405Sjkim *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or 1676238405Sjkim EVP_CIPHER structures to avoid later problems in EVP_cleanup(). 1677238405Sjkim [Steve Henson] 1678238405Sjkim 1679238405Sjkim *) New utilities pkey and pkeyparam. These are similar to algorithm specific 1680238405Sjkim utilities such as rsa, dsa, dsaparam etc except they process any key 1681238405Sjkim type. 1682238405Sjkim [Steve Henson] 1683238405Sjkim 1684238405Sjkim *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New 1685238405Sjkim functions EVP_PKEY_print_public(), EVP_PKEY_print_private(), 1686238405Sjkim EVP_PKEY_print_param() to print public key data from an EVP_PKEY 1687238405Sjkim structure. 1688238405Sjkim [Steve Henson] 1689238405Sjkim 1690238405Sjkim *) Initial support for pluggable public key ASN1. 1691238405Sjkim De-spaghettify the public key ASN1 handling. Move public and private 1692238405Sjkim key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate 1693238405Sjkim algorithm specific handling to a single module within the relevant 1694238405Sjkim algorithm directory. Add functions to allow (near) opaque processing 1695238405Sjkim of public and private key structures. 1696238405Sjkim [Steve Henson] 1697238405Sjkim 1698238405Sjkim *) Implement the Supported Point Formats Extension for 1699238405Sjkim ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1700238405Sjkim [Douglas Stebila] 1701238405Sjkim 1702238405Sjkim *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members 1703238405Sjkim for the psk identity [hint] and the psk callback functions to the 1704238405Sjkim SSL_SESSION, SSL and SSL_CTX structure. 1705238405Sjkim 1706238405Sjkim New ciphersuites: 1707238405Sjkim PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA, 1708238405Sjkim PSK-AES256-CBC-SHA 1709238405Sjkim 1710238405Sjkim New functions: 1711238405Sjkim SSL_CTX_use_psk_identity_hint 1712238405Sjkim SSL_get_psk_identity_hint 1713238405Sjkim SSL_get_psk_identity 1714238405Sjkim SSL_use_psk_identity_hint 1715238405Sjkim 1716238405Sjkim [Mika Kousa and Pasi Eronen of Nokia Corporation] 1717238405Sjkim 1718238405Sjkim *) Add RFC 3161 compliant time stamp request creation, response generation 1719238405Sjkim and response verification functionality. 1720238405Sjkim [Zolt�n Gl�zik <zglozik@opentsa.org>, The OpenTSA Project] 1721238405Sjkim 1722238405Sjkim *) Add initial support for TLS extensions, specifically for the server_name 1723238405Sjkim extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 1724238405Sjkim have new members for a host name. The SSL data structure has an 1725238405Sjkim additional member SSL_CTX *initial_ctx so that new sessions can be 1726238405Sjkim stored in that context to allow for session resumption, even after the 1727238405Sjkim SSL has been switched to a new SSL_CTX in reaction to a client's 1728238405Sjkim server_name extension. 1729238405Sjkim 1730238405Sjkim New functions (subject to change): 1731238405Sjkim 1732238405Sjkim SSL_get_servername() 1733238405Sjkim SSL_get_servername_type() 1734238405Sjkim SSL_set_SSL_CTX() 1735238405Sjkim 1736238405Sjkim New CTRL codes and macros (subject to change): 1737238405Sjkim 1738238405Sjkim SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 1739238405Sjkim - SSL_CTX_set_tlsext_servername_callback() 1740238405Sjkim SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 1741238405Sjkim - SSL_CTX_set_tlsext_servername_arg() 1742238405Sjkim SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 1743238405Sjkim 1744238405Sjkim openssl s_client has a new '-servername ...' option. 1745238405Sjkim 1746238405Sjkim openssl s_server has new options '-servername_host ...', '-cert2 ...', 1747238405Sjkim '-key2 ...', '-servername_fatal' (subject to change). This allows 1748238405Sjkim testing the HostName extension for a specific single host name ('-cert' 1749238405Sjkim and '-key' remain fallbacks for handshakes without HostName 1750238405Sjkim negotiation). If the unrecogninzed_name alert has to be sent, this by 1751238405Sjkim default is a warning; it becomes fatal with the '-servername_fatal' 1752238405Sjkim option. 1753238405Sjkim 1754238405Sjkim [Peter Sylvester, Remy Allais, Christophe Renou] 1755238405Sjkim 1756238405Sjkim *) Whirlpool hash implementation is added. 1757238405Sjkim [Andy Polyakov] 1758238405Sjkim 1759238405Sjkim *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to 1760238405Sjkim bn(64,32). Because of instruction set limitations it doesn't have 1761238405Sjkim any negative impact on performance. This was done mostly in order 1762238405Sjkim to make it possible to share assembler modules, such as bn_mul_mont 1763238405Sjkim implementations, between 32- and 64-bit builds without hassle. 1764238405Sjkim [Andy Polyakov] 1765238405Sjkim 1766238405Sjkim *) Move code previously exiled into file crypto/ec/ec2_smpt.c 1767238405Sjkim to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP 1768238405Sjkim macro. 1769238405Sjkim [Bodo Moeller] 1770238405Sjkim 1771238405Sjkim *) New candidate for BIGNUM assembler implementation, bn_mul_mont, 1772238405Sjkim dedicated Montgomery multiplication procedure, is introduced. 1773238405Sjkim BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher 1774238405Sjkim "64-bit" performance on certain 32-bit targets. 1775238405Sjkim [Andy Polyakov] 1776238405Sjkim 1777238405Sjkim *) New option SSL_OP_NO_COMP to disable use of compression selectively 1778238405Sjkim in SSL structures. New SSL ctrl to set maximum send fragment size. 1779238405Sjkim Save memory by seeting the I/O buffer sizes dynamically instead of 1780238405Sjkim using the maximum available value. 1781238405Sjkim [Steve Henson] 1782238405Sjkim 1783238405Sjkim *) New option -V for 'openssl ciphers'. This prints the ciphersuite code 1784238405Sjkim in addition to the text details. 1785238405Sjkim [Bodo Moeller] 1786238405Sjkim 1787238405Sjkim *) Very, very preliminary EXPERIMENTAL support for printing of general 1788238405Sjkim ASN1 structures. This currently produces rather ugly output and doesn't 1789238405Sjkim handle several customised structures at all. 1790238405Sjkim [Steve Henson] 1791238405Sjkim 1792238405Sjkim *) Integrated support for PVK file format and some related formats such 1793238405Sjkim as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support 1794238405Sjkim these in the 'rsa' and 'dsa' utilities. 1795238405Sjkim [Steve Henson] 1796238405Sjkim 1797238405Sjkim *) Support for PKCS#1 RSAPublicKey format on rsa utility command line. 1798238405Sjkim [Steve Henson] 1799238405Sjkim 1800238405Sjkim *) Remove the ancient ASN1_METHOD code. This was only ever used in one 1801238405Sjkim place for the (very old) "NETSCAPE" format certificates which are now 1802238405Sjkim handled using new ASN1 code equivalents. 1803238405Sjkim [Steve Henson] 1804238405Sjkim 1805238405Sjkim *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD 1806238405Sjkim pointer and make the SSL_METHOD parameter in SSL_CTX_new, 1807238405Sjkim SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'. 1808238405Sjkim [Nils Larsch] 1809238405Sjkim 1810238405Sjkim *) Modify CRL distribution points extension code to print out previously 1811238405Sjkim unsupported fields. Enhance extension setting code to allow setting of 1812238405Sjkim all fields. 1813238405Sjkim [Steve Henson] 1814238405Sjkim 1815238405Sjkim *) Add print and set support for Issuing Distribution Point CRL extension. 1816238405Sjkim [Steve Henson] 1817238405Sjkim 1818238405Sjkim *) Change 'Configure' script to enable Camellia by default. 1819238405Sjkim [NTT] 1820238405Sjkim 1821264331Sjkim Changes between 0.9.8x and 0.9.8y [5 Feb 2013] 1822264331Sjkim 1823264331Sjkim *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 1824264331Sjkim 1825264331Sjkim This addresses the flaw in CBC record processing discovered by 1826264331Sjkim Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 1827264331Sjkim at: http://www.isg.rhul.ac.uk/tls/ 1828264331Sjkim 1829264331Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 1830264331Sjkim Security Group at Royal Holloway, University of London 1831264331Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 1832264331Sjkim Emilia K�sper for the initial patch. 1833264331Sjkim (CVE-2013-0169) 1834264331Sjkim [Emilia K�sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 1835264331Sjkim 1836264331Sjkim *) Return an error when checking OCSP signatures when key is NULL. 1837264331Sjkim This fixes a DoS attack. (CVE-2013-0166) 1838264331Sjkim [Steve Henson] 1839264331Sjkim 1840264331Sjkim *) Call OCSP Stapling callback after ciphersuite has been chosen, so 1841264331Sjkim the right response is stapled. Also change SSL_get_certificate() 1842264331Sjkim so it returns the certificate actually sent. 1843264331Sjkim See http://rt.openssl.org/Ticket/Display.html?id=2836. 1844264331Sjkim (This is a backport) 1845264331Sjkim [Rob Stradling <rob.stradling@comodo.com>] 1846264331Sjkim 1847264331Sjkim *) Fix possible deadlock when decoding public keys. 1848264331Sjkim [Steve Henson] 1849264331Sjkim 1850264331Sjkim Changes between 0.9.8w and 0.9.8x [10 May 2012] 1851264331Sjkim 1852264331Sjkim *) Sanity check record length before skipping explicit IV in DTLS 1853264331Sjkim to fix DoS attack. 1854264331Sjkim 1855264331Sjkim Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 1856264331Sjkim fuzzing as a service testing platform. 1857264331Sjkim (CVE-2012-2333) 1858264331Sjkim [Steve Henson] 1859264331Sjkim 1860264331Sjkim *) Initialise tkeylen properly when encrypting CMS messages. 1861264331Sjkim Thanks to Solar Designer of Openwall for reporting this issue. 1862264331Sjkim [Steve Henson] 1863264331Sjkim 1864264331Sjkim Changes between 0.9.8v and 0.9.8w [23 Apr 2012] 1865264331Sjkim 1866264331Sjkim *) The fix for CVE-2012-2110 did not take into account that the 1867264331Sjkim 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an 1868264331Sjkim int in OpenSSL 0.9.8, making it still vulnerable. Fix by 1869264331Sjkim rejecting negative len parameter. (CVE-2012-2131) 1870264331Sjkim [Tomas Hoger <thoger@redhat.com>] 1871264331Sjkim 1872264331Sjkim Changes between 0.9.8u and 0.9.8v [19 Apr 2012] 1873264331Sjkim 1874264331Sjkim *) Check for potentially exploitable overflows in asn1_d2i_read_bio 1875264331Sjkim BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 1876264331Sjkim in CRYPTO_realloc_clean. 1877264331Sjkim 1878264331Sjkim Thanks to Tavis Ormandy, Google Security Team, for discovering this 1879264331Sjkim issue and to Adam Langley <agl@chromium.org> for fixing it. 1880264331Sjkim (CVE-2012-2110) 1881264331Sjkim [Adam Langley (Google), Tavis Ormandy, Google Security Team] 1882264331Sjkim 1883264331Sjkim Changes between 0.9.8t and 0.9.8u [12 Mar 2012] 1884264331Sjkim 1885264331Sjkim *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 1886264331Sjkim in CMS and PKCS7 code. When RSA decryption fails use a random key for 1887264331Sjkim content decryption and always return the same error. Note: this attack 1888264331Sjkim needs on average 2^20 messages so it only affects automated senders. The 1889264331Sjkim old behaviour can be reenabled in the CMS code by setting the 1890264331Sjkim CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 1891264331Sjkim an MMA defence is not necessary. 1892264331Sjkim Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 1893264331Sjkim this issue. (CVE-2012-0884) 1894264331Sjkim [Steve Henson] 1895264331Sjkim 1896264331Sjkim *) Fix CVE-2011-4619: make sure we really are receiving a 1897264331Sjkim client hello before rejecting multiple SGC restarts. Thanks to 1898264331Sjkim Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 1899264331Sjkim [Steve Henson] 1900264331Sjkim 1901237657Sjkim Changes between 0.9.8s and 0.9.8t [18 Jan 2012] 1902237657Sjkim 1903237657Sjkim *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 1904237657Sjkim Thanks to Antonio Martin, Enterprise Secure Access Research and 1905237657Sjkim Development, Cisco Systems, Inc. for discovering this bug and 1906237657Sjkim preparing a fix. (CVE-2012-0050) 1907237657Sjkim [Antonio Martin] 1908264331Sjkim 1909237657Sjkim Changes between 0.9.8r and 0.9.8s [4 Jan 2012] 1910237657Sjkim 1911237657Sjkim *) Nadhem Alfardan and Kenny Paterson have discovered an extension 1912237657Sjkim of the Vaudenay padding oracle attack on CBC mode encryption 1913237657Sjkim which enables an efficient plaintext recovery attack against 1914237657Sjkim the OpenSSL implementation of DTLS. Their attack exploits timing 1915237657Sjkim differences arising during decryption processing. A research 1916237657Sjkim paper describing this attack can be found at: 1917237657Sjkim http://www.isg.rhul.ac.uk/~kp/dtls.pdf 1918237657Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 1919237657Sjkim Security Group at Royal Holloway, University of London 1920237657Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 1921237657Sjkim <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 1922237657Sjkim for preparing the fix. (CVE-2011-4108) 1923237657Sjkim [Robin Seggelmann, Michael Tuexen] 1924237657Sjkim 1925237657Sjkim *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109) 1926237657Sjkim [Ben Laurie, Kasper <ekasper@google.com>] 1927237657Sjkim 1928237657Sjkim *) Clear bytes used for block padding of SSL 3.0 records. 1929237657Sjkim (CVE-2011-4576) 1930237657Sjkim [Adam Langley (Google)] 1931237657Sjkim 1932237657Sjkim *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 1933237657Sjkim Kadianakis <desnacked@gmail.com> for discovering this issue and 1934237657Sjkim Adam Langley for preparing the fix. (CVE-2011-4619) 1935237657Sjkim [Adam Langley (Google)] 1936237657Sjkim 1937237657Sjkim *) Prevent malformed RFC3779 data triggering an assertion failure. 1938237657Sjkim Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 1939237657Sjkim and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 1940237657Sjkim [Rob Austein <sra@hactrn.net>] 1941237657Sjkim 1942237657Sjkim *) Fix ssl_ciph.c set-up race. 1943237657Sjkim [Adam Langley (Google)] 1944237657Sjkim 1945237657Sjkim *) Fix spurious failures in ecdsatest.c. 1946237657Sjkim [Emilia K�sper (Google)] 1947237657Sjkim 1948237657Sjkim *) Fix the BIO_f_buffer() implementation (which was mixing different 1949237657Sjkim interpretations of the '..._len' fields). 1950237657Sjkim [Adam Langley (Google)] 1951237657Sjkim 1952237657Sjkim *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 1953237657Sjkim BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 1954237657Sjkim threads won't reuse the same blinding coefficients. 1955237657Sjkim 1956237657Sjkim This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 1957237657Sjkim lock to call BN_BLINDING_invert_ex, and avoids one use of 1958237657Sjkim BN_BLINDING_update for each BN_BLINDING structure (previously, 1959237657Sjkim the last update always remained unused). 1960237657Sjkim [Emilia K�sper (Google)] 1961237657Sjkim 1962237657Sjkim *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 1963237657Sjkim for multi-threaded use of ECDH. 1964237657Sjkim [Adam Langley (Google)] 1965237657Sjkim 1966237657Sjkim *) Fix x509_name_ex_d2i memory leak on bad inputs. 1967237657Sjkim [Bodo Moeller] 1968237657Sjkim 1969237657Sjkim *) Add protection against ECDSA timing attacks as mentioned in the paper 1970237657Sjkim by Billy Bob Brumley and Nicola Tuveri, see: 1971237657Sjkim 1972237657Sjkim http://eprint.iacr.org/2011/232.pdf 1973237657Sjkim 1974237657Sjkim [Billy Bob Brumley and Nicola Tuveri] 1975237657Sjkim 1976237657Sjkim Changes between 0.9.8q and 0.9.8r [8 Feb 2011] 1977237657Sjkim 1978237657Sjkim *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 1979237657Sjkim [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 1980237657Sjkim 1981237657Sjkim *) Fix bug in string printing code: if *any* escaping is enabled we must 1982237657Sjkim escape the escape character (backslash) or the resulting string is 1983237657Sjkim ambiguous. 1984237657Sjkim [Steve Henson] 1985237657Sjkim 1986216166Ssimon Changes between 0.9.8p and 0.9.8q [2 Dec 2010] 1987216166Ssimon 1988216166Ssimon *) Disable code workaround for ancient and obsolete Netscape browsers 1989216166Ssimon and servers: an attacker can use it in a ciphersuite downgrade attack. 1990216166Ssimon Thanks to Martin Rex for discovering this bug. CVE-2010-4180 1991216166Ssimon [Steve Henson] 1992216166Ssimon 1993216166Ssimon *) Fixed J-PAKE implementation error, originally discovered by 1994216166Ssimon Sebastien Martini, further info and confirmation from Stefan 1995216166Ssimon Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 1996216166Ssimon [Ben Laurie] 1997216166Ssimon 1998215697Ssimon Changes between 0.9.8o and 0.9.8p [16 Nov 2010] 1999215697Ssimon 2000215697Ssimon *) Fix extension code to avoid race conditions which can result in a buffer 2001215697Ssimon overrun vulnerability: resumed sessions must not be modified as they can 2002215697Ssimon be shared by multiple threads. CVE-2010-3864 2003215697Ssimon [Steve Henson] 2004215697Ssimon 2005215697Ssimon *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939 2006215697Ssimon [Steve Henson] 2007215697Ssimon 2008215697Ssimon *) Don't reencode certificate when calculating signature: cache and use 2009215697Ssimon the original encoding instead. This makes signature verification of 2010215697Ssimon some broken encodings work correctly. 2011215697Ssimon [Steve Henson] 2012215697Ssimon 2013215697Ssimon *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT 2014215697Ssimon is also one of the inputs. 2015215697Ssimon [Emilia K�sper <emilia.kasper@esat.kuleuven.be> (Google)] 2016215697Ssimon 2017215697Ssimon *) Don't repeatedly append PBE algorithms to table if they already exist. 2018215697Ssimon Sort table on each new add. This effectively makes the table read only 2019215697Ssimon after all algorithms are added and subsequent calls to PKCS12_pbe_add 2020215697Ssimon etc are non-op. 2021215697Ssimon [Steve Henson] 2022215697Ssimon 2023215697Ssimon Changes between 0.9.8n and 0.9.8o [01 Jun 2010] 2024215697Ssimon 2025215697Ssimon [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after 2026215697Ssimon OpenSSL 1.0.0.] 2027215697Ssimon 2028215697Ssimon *) Correct a typo in the CMS ASN1 module which can result in invalid memory 2029215697Ssimon access or freeing data twice (CVE-2010-0742) 2030215697Ssimon [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>] 2031215697Ssimon 2032215697Ssimon *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more 2033215697Ssimon common in certificates and some applications which only call 2034215697Ssimon SSL_library_init and not OpenSSL_add_all_algorithms() will fail. 2035215697Ssimon [Steve Henson] 2036215697Ssimon 2037215697Ssimon *) VMS fixes: 2038215697Ssimon Reduce copying into .apps and .test in makevms.com 2039215697Ssimon Don't try to use blank CA certificate in CA.com 2040215697Ssimon Allow use of C files from original directories in maketests.com 2041215697Ssimon [Steven M. Schweda" <sms@antinode.info>] 2042215697Ssimon 2043206046Ssimon Changes between 0.9.8m and 0.9.8n [24 Mar 2010] 2044206046Ssimon 2045206046Ssimon *) When rejecting SSL/TLS records due to an incorrect version number, never 2046206046Ssimon update s->server with a new major version number. As of 2047206046Ssimon - OpenSSL 0.9.8m if 'short' is a 16-bit type, 2048206046Ssimon - OpenSSL 0.9.8f if 'short' is longer than 16 bits, 2049206046Ssimon the previous behavior could result in a read attempt at NULL when 2050206046Ssimon receiving specific incorrect SSL/TLS records once record payload 2051206046Ssimon protection is active. (CVE-2010-0740) 2052206046Ssimon [Bodo Moeller, Adam Langley <agl@chromium.org>] 2053206046Ssimon 2054206046Ssimon *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL 2055206046Ssimon could be crashed if the relevant tables were not present (e.g. chrooted). 2056206046Ssimon [Tomas Hoger <thoger@redhat.com>] 2057206046Ssimon 2058205128Ssimon Changes between 0.9.8l and 0.9.8m [25 Feb 2010] 2059205128Ssimon 2060205128Ssimon *) Always check bn_wexpend() return values for failure. (CVE-2009-3245) 2061205128Ssimon [Martin Olsson, Neel Mehta] 2062205128Ssimon 2063205128Ssimon *) Fix X509_STORE locking: Every 'objs' access requires a lock (to 2064205128Ssimon accommodate for stack sorting, always a write lock!). 2065205128Ssimon [Bodo Moeller] 2066205128Ssimon 2067205128Ssimon *) On some versions of WIN32 Heap32Next is very slow. This can cause 2068205128Ssimon excessive delays in the RAND_poll(): over a minute. As a workaround 2069205128Ssimon include a time check in the inner Heap32Next loop too. 2070205128Ssimon [Steve Henson] 2071205128Ssimon 2072205128Ssimon *) The code that handled flushing of data in SSL/TLS originally used the 2073205128Ssimon BIO_CTRL_INFO ctrl to see if any data was pending first. This caused 2074205128Ssimon the problem outlined in PR#1949. The fix suggested there however can 2075205128Ssimon trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions 2076205128Ssimon of Apache). So instead simplify the code to flush unconditionally. 2077205128Ssimon This should be fine since flushing with no data to flush is a no op. 2078205128Ssimon [Steve Henson] 2079205128Ssimon 2080205128Ssimon *) Handle TLS versions 2.0 and later properly and correctly use the 2081205128Ssimon highest version of TLS/SSL supported. Although TLS >= 2.0 is some way 2082205128Ssimon off ancient servers have a habit of sticking around for a while... 2083205128Ssimon [Steve Henson] 2084205128Ssimon 2085205128Ssimon *) Modify compression code so it frees up structures without using the 2086205128Ssimon ex_data callbacks. This works around a problem where some applications 2087205128Ssimon call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when 2088205128Ssimon restarting) then use compression (e.g. SSL with compression) later. 2089205128Ssimon This results in significant per-connection memory leaks and 2090205128Ssimon has caused some security issues including CVE-2008-1678 and 2091205128Ssimon CVE-2009-4355. 2092205128Ssimon [Steve Henson] 2093205128Ssimon 2094205128Ssimon *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't 2095205128Ssimon change when encrypting or decrypting. 2096205128Ssimon [Bodo Moeller] 2097205128Ssimon 2098205128Ssimon *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to 2099205128Ssimon connect and renegotiate with servers which do not support RI. 2100205128Ssimon Until RI is more widely deployed this option is enabled by default. 2101205128Ssimon [Steve Henson] 2102205128Ssimon 2103205128Ssimon *) Add "missing" ssl ctrls to clear options and mode. 2104205128Ssimon [Steve Henson] 2105205128Ssimon 2106205128Ssimon *) If client attempts to renegotiate and doesn't support RI respond with 2107205128Ssimon a no_renegotiation alert as required by RFC5746. Some renegotiating 2108205128Ssimon TLS clients will continue a connection gracefully when they receive 2109205128Ssimon the alert. Unfortunately OpenSSL mishandled this alert and would hang 2110205128Ssimon waiting for a server hello which it will never receive. Now we treat a 2111205128Ssimon received no_renegotiation alert as a fatal error. This is because 2112205128Ssimon applications requesting a renegotiation might well expect it to succeed 2113205128Ssimon and would have no code in place to handle the server denying it so the 2114205128Ssimon only safe thing to do is to terminate the connection. 2115205128Ssimon [Steve Henson] 2116205128Ssimon 2117205128Ssimon *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if 2118205128Ssimon peer supports secure renegotiation and 0 otherwise. Print out peer 2119205128Ssimon renegotiation support in s_client/s_server. 2120205128Ssimon [Steve Henson] 2121205128Ssimon 2122205128Ssimon *) Replace the highly broken and deprecated SPKAC certification method with 2123205128Ssimon the updated NID creation version. This should correctly handle UTF8. 2124205128Ssimon [Steve Henson] 2125205128Ssimon 2126205128Ssimon *) Implement RFC5746. Re-enable renegotiation but require the extension 2127205128Ssimon as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION 2128205128Ssimon turns out to be a bad idea. It has been replaced by 2129205128Ssimon SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with 2130205128Ssimon SSL_CTX_set_options(). This is really not recommended unless you 2131205128Ssimon know what you are doing. 2132205128Ssimon [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson] 2133205128Ssimon 2134205128Ssimon *) Fixes to stateless session resumption handling. Use initial_ctx when 2135205128Ssimon issuing and attempting to decrypt tickets in case it has changed during 2136205128Ssimon servername handling. Use a non-zero length session ID when attempting 2137205128Ssimon stateless session resumption: this makes it possible to determine if 2138205128Ssimon a resumption has occurred immediately after receiving server hello 2139205128Ssimon (several places in OpenSSL subtly assume this) instead of later in 2140205128Ssimon the handshake. 2141205128Ssimon [Steve Henson] 2142205128Ssimon 2143205128Ssimon *) The functions ENGINE_ctrl(), OPENSSL_isservice(), 2144205128Ssimon CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error 2145205128Ssimon fixes for a few places where the return code is not checked 2146205128Ssimon correctly. 2147205128Ssimon [Julia Lawall <julia@diku.dk>] 2148205128Ssimon 2149205128Ssimon *) Add --strict-warnings option to Configure script to include devteam 2150205128Ssimon warnings in other configurations. 2151205128Ssimon [Steve Henson] 2152205128Ssimon 2153205128Ssimon *) Add support for --libdir option and LIBDIR variable in makefiles. This 2154205128Ssimon makes it possible to install openssl libraries in locations which 2155205128Ssimon have names other than "lib", for example "/usr/lib64" which some 2156205128Ssimon systems need. 2157205128Ssimon [Steve Henson, based on patch from Jeremy Utley] 2158205128Ssimon 2159205128Ssimon *) Don't allow the use of leading 0x80 in OIDs. This is a violation of 2160205128Ssimon X690 8.9.12 and can produce some misleading textual output of OIDs. 2161205128Ssimon [Steve Henson, reported by Dan Kaminsky] 2162205128Ssimon 2163205128Ssimon *) Delete MD2 from algorithm tables. This follows the recommendation in 2164205128Ssimon several standards that it is not used in new applications due to 2165205128Ssimon several cryptographic weaknesses. For binary compatibility reasons 2166205128Ssimon the MD2 API is still compiled in by default. 2167205128Ssimon [Steve Henson] 2168205128Ssimon 2169205128Ssimon *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved 2170205128Ssimon and restored. 2171205128Ssimon [Steve Henson] 2172205128Ssimon 2173205128Ssimon *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and 2174205128Ssimon OPENSSL_asc2uni conditionally on Netware platforms to avoid a name 2175205128Ssimon clash. 2176205128Ssimon [Guenter <lists@gknw.net>] 2177205128Ssimon 2178205128Ssimon *) Fix the server certificate chain building code to use X509_verify_cert(), 2179205128Ssimon it used to have an ad-hoc builder which was unable to cope with anything 2180205128Ssimon other than a simple chain. 2181205128Ssimon [David Woodhouse <dwmw2@infradead.org>, Steve Henson] 2182205128Ssimon 2183205128Ssimon *) Don't check self signed certificate signatures in X509_verify_cert() 2184205128Ssimon by default (a flag can override this): it just wastes time without 2185205128Ssimon adding any security. As a useful side effect self signed root CAs 2186205128Ssimon with non-FIPS digests are now usable in FIPS mode. 2187205128Ssimon [Steve Henson] 2188205128Ssimon 2189205128Ssimon *) In dtls1_process_out_of_seq_message() the check if the current message 2190205128Ssimon is already buffered was missing. For every new message was memory 2191205128Ssimon allocated, allowing an attacker to perform an denial of service attack 2192205128Ssimon with sending out of seq handshake messages until there is no memory 2193205128Ssimon left. Additionally every future messege was buffered, even if the 2194205128Ssimon sequence number made no sense and would be part of another handshake. 2195205128Ssimon So only messages with sequence numbers less than 10 in advance will be 2196205128Ssimon buffered. (CVE-2009-1378) 2197205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 2198205128Ssimon 2199205128Ssimon *) Records are buffered if they arrive with a future epoch to be 2200205128Ssimon processed after finishing the corresponding handshake. There is 2201205128Ssimon currently no limitation to this buffer allowing an attacker to perform 2202205128Ssimon a DOS attack with sending records with future epochs until there is no 2203205128Ssimon memory left. This patch adds the pqueue_size() function to detemine 2204205128Ssimon the size of a buffer and limits the record buffer to 100 entries. 2205205128Ssimon (CVE-2009-1377) 2206205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 2207205128Ssimon 2208205128Ssimon *) Keep a copy of frag->msg_header.frag_len so it can be used after the 2209205128Ssimon parent structure is freed. (CVE-2009-1379) 2210205128Ssimon [Daniel Mentz] 2211205128Ssimon 2212205128Ssimon *) Handle non-blocking I/O properly in SSL_shutdown() call. 2213205128Ssimon [Darryl Miles <darryl-mailinglists@netbauds.net>] 2214205128Ssimon 2215205128Ssimon *) Add 2.5.4.* OIDs 2216205128Ssimon [Ilya O. <vrghost@gmail.com>] 2217205128Ssimon 2218205128Ssimon Changes between 0.9.8k and 0.9.8l [5 Nov 2009] 2219205128Ssimon 2220205128Ssimon *) Disable renegotiation completely - this fixes a severe security 2221205128Ssimon problem (CVE-2009-3555) at the cost of breaking all 2222205128Ssimon renegotiation. Renegotiation can be re-enabled by setting 2223205128Ssimon SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at 2224205128Ssimon run-time. This is really not recommended unless you know what 2225205128Ssimon you're doing. 2226205128Ssimon [Ben Laurie] 2227205128Ssimon 2228194206Ssimon Changes between 0.9.8j and 0.9.8k [25 Mar 2009] 2229167612Ssimon 2230194206Ssimon *) Don't set val to NULL when freeing up structures, it is freed up by 2231194206Ssimon underlying code. If sizeof(void *) > sizeof(long) this can result in 2232194206Ssimon zeroing past the valid field. (CVE-2009-0789) 2233194206Ssimon [Paolo Ganci <Paolo.Ganci@AdNovum.CH>] 2234194206Ssimon 2235194206Ssimon *) Fix bug where return value of CMS_SignerInfo_verify_content() was not 2236194206Ssimon checked correctly. This would allow some invalid signed attributes to 2237194206Ssimon appear to verify correctly. (CVE-2009-0591) 2238194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 2239194206Ssimon 2240194206Ssimon *) Reject UniversalString and BMPString types with invalid lengths. This 2241194206Ssimon prevents a crash in ASN1_STRING_print_ex() which assumes the strings have 2242194206Ssimon a legal length. (CVE-2009-0590) 2243194206Ssimon [Steve Henson] 2244194206Ssimon 2245194206Ssimon *) Set S/MIME signing as the default purpose rather than setting it 2246194206Ssimon unconditionally. This allows applications to override it at the store 2247194206Ssimon level. 2248194206Ssimon [Steve Henson] 2249194206Ssimon 2250194206Ssimon *) Permit restricted recursion of ASN1 strings. This is needed in practice 2251194206Ssimon to handle some structures. 2252194206Ssimon [Steve Henson] 2253194206Ssimon 2254194206Ssimon *) Improve efficiency of mem_gets: don't search whole buffer each time 2255194206Ssimon for a '\n' 2256194206Ssimon [Jeremy Shapiro <jnshapir@us.ibm.com>] 2257194206Ssimon 2258194206Ssimon *) New -hex option for openssl rand. 2259194206Ssimon [Matthieu Herrb] 2260194206Ssimon 2261194206Ssimon *) Print out UTF8String and NumericString when parsing ASN1. 2262194206Ssimon [Steve Henson] 2263194206Ssimon 2264194206Ssimon *) Support NumericString type for name components. 2265194206Ssimon [Steve Henson] 2266194206Ssimon 2267194206Ssimon *) Allow CC in the environment to override the automatically chosen 2268194206Ssimon compiler. Note that nothing is done to ensure flags work with the 2269194206Ssimon chosen compiler. 2270194206Ssimon [Ben Laurie] 2271194206Ssimon 2272194206Ssimon Changes between 0.9.8i and 0.9.8j [07 Jan 2009] 2273194206Ssimon 2274194206Ssimon *) Properly check EVP_VerifyFinal() and similar return values 2275194206Ssimon (CVE-2008-5077). 2276194206Ssimon [Ben Laurie, Bodo Moeller, Google Security Team] 2277194206Ssimon 2278194206Ssimon *) Enable TLS extensions by default. 2279194206Ssimon [Ben Laurie] 2280194206Ssimon 2281194206Ssimon *) Allow the CHIL engine to be loaded, whether the application is 2282194206Ssimon multithreaded or not. (This does not release the developer from the 2283194206Ssimon obligation to set up the dynamic locking callbacks.) 2284194206Ssimon [Sander Temme <sander@temme.net>] 2285194206Ssimon 2286194206Ssimon *) Use correct exit code if there is an error in dgst command. 2287194206Ssimon [Steve Henson; problem pointed out by Roland Dirlewanger] 2288194206Ssimon 2289194206Ssimon *) Tweak Configure so that you need to say "experimental-jpake" to enable 2290194206Ssimon JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. 2291194206Ssimon [Bodo Moeller] 2292194206Ssimon 2293194206Ssimon *) Add experimental JPAKE support, including demo authentication in 2294194206Ssimon s_client and s_server. 2295194206Ssimon [Ben Laurie] 2296194206Ssimon 2297194206Ssimon *) Set the comparison function in v3_addr_canonize(). 2298194206Ssimon [Rob Austein <sra@hactrn.net>] 2299194206Ssimon 2300194206Ssimon *) Add support for XMPP STARTTLS in s_client. 2301194206Ssimon [Philip Paeps <philip@freebsd.org>] 2302194206Ssimon 2303194206Ssimon *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior 2304194206Ssimon to ensure that even with this option, only ciphersuites in the 2305194206Ssimon server's preference list will be accepted. (Note that the option 2306194206Ssimon applies only when resuming a session, so the earlier behavior was 2307194206Ssimon just about the algorithm choice for symmetric cryptography.) 2308194206Ssimon [Bodo Moeller] 2309194206Ssimon 2310194206Ssimon Changes between 0.9.8h and 0.9.8i [15 Sep 2008] 2311194206Ssimon 2312205128Ssimon *) Fix NULL pointer dereference if a DTLS server received 2313205128Ssimon ChangeCipherSpec as first record (CVE-2009-1386). 2314205128Ssimon [PR #1679] 2315205128Ssimon 2316194206Ssimon *) Fix a state transitition in s3_srvr.c and d1_srvr.c 2317194206Ssimon (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...). 2318194206Ssimon [Nagendra Modadugu] 2319194206Ssimon 2320194206Ssimon *) The fix in 0.9.8c that supposedly got rid of unsafe 2321194206Ssimon double-checked locking was incomplete for RSA blinding, 2322194206Ssimon addressing just one layer of what turns out to have been 2323194206Ssimon doubly unsafe triple-checked locking. 2324194206Ssimon 2325194206Ssimon So now fix this for real by retiring the MONT_HELPER macro 2326194206Ssimon in crypto/rsa/rsa_eay.c. 2327194206Ssimon 2328194206Ssimon [Bodo Moeller; problem pointed out by Marius Schilder] 2329194206Ssimon 2330194206Ssimon *) Various precautionary measures: 2331194206Ssimon 2332194206Ssimon - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). 2333194206Ssimon 2334194206Ssimon - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). 2335194206Ssimon (NB: This would require knowledge of the secret session ticket key 2336194206Ssimon to exploit, in which case you'd be SOL either way.) 2337194206Ssimon 2338194206Ssimon - Change bn_nist.c so that it will properly handle input BIGNUMs 2339194206Ssimon outside the expected range. 2340194206Ssimon 2341194206Ssimon - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG 2342194206Ssimon builds. 2343194206Ssimon 2344194206Ssimon [Neel Mehta, Bodo Moeller] 2345194206Ssimon 2346194206Ssimon *) Allow engines to be "soft loaded" - i.e. optionally don't die if 2347194206Ssimon the load fails. Useful for distros. 2348194206Ssimon [Ben Laurie and the FreeBSD team] 2349194206Ssimon 2350194206Ssimon *) Add support for Local Machine Keyset attribute in PKCS#12 files. 2351194206Ssimon [Steve Henson] 2352194206Ssimon 2353194206Ssimon *) Fix BN_GF2m_mod_arr() top-bit cleanup code. 2354194206Ssimon [Huang Ying] 2355194206Ssimon 2356194206Ssimon *) Expand ENGINE to support engine supplied SSL client certificate functions. 2357194206Ssimon 2358194206Ssimon This work was sponsored by Logica. 2359194206Ssimon [Steve Henson] 2360194206Ssimon 2361194206Ssimon *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows 2362194206Ssimon keystores. Support for SSL/TLS client authentication too. 2363194206Ssimon Not compiled unless enable-capieng specified to Configure. 2364194206Ssimon 2365194206Ssimon This work was sponsored by Logica. 2366194206Ssimon [Steve Henson] 2367194206Ssimon 2368194206Ssimon *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using 2369194206Ssimon ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain 2370194206Ssimon attribute creation routines such as certifcate requests and PKCS#12 2371194206Ssimon files. 2372194206Ssimon [Steve Henson] 2373194206Ssimon 2374194206Ssimon Changes between 0.9.8g and 0.9.8h [28 May 2008] 2375194206Ssimon 2376194206Ssimon *) Fix flaw if 'Server Key exchange message' is omitted from a TLS 2377194206Ssimon handshake which could lead to a cilent crash as found using the 2378194206Ssimon Codenomicon TLS test suite (CVE-2008-1672) 2379194206Ssimon [Steve Henson, Mark Cox] 2380194206Ssimon 2381194206Ssimon *) Fix double free in TLS server name extensions which could lead to 2382194206Ssimon a remote crash found by Codenomicon TLS test suite (CVE-2008-0891) 2383194206Ssimon [Joe Orton] 2384194206Ssimon 2385194206Ssimon *) Clear error queue in SSL_CTX_use_certificate_chain_file() 2386194206Ssimon 2387194206Ssimon Clear the error queue to ensure that error entries left from 2388194206Ssimon older function calls do not interfere with the correct operation. 2389194206Ssimon [Lutz Jaenicke, Erik de Castro Lopo] 2390194206Ssimon 2391194206Ssimon *) Remove root CA certificates of commercial CAs: 2392194206Ssimon 2393194206Ssimon The OpenSSL project does not recommend any specific CA and does not 2394194206Ssimon have any policy with respect to including or excluding any CA. 2395194206Ssimon Therefore it does not make any sense to ship an arbitrary selection 2396194206Ssimon of root CA certificates with the OpenSSL software. 2397194206Ssimon [Lutz Jaenicke] 2398194206Ssimon 2399194206Ssimon *) RSA OAEP patches to fix two separate invalid memory reads. 2400194206Ssimon The first one involves inputs when 'lzero' is greater than 2401194206Ssimon 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes 2402194206Ssimon before the beginning of from). The second one involves inputs where 2403194206Ssimon the 'db' section contains nothing but zeroes (there is a one-byte 2404194206Ssimon invalid read after the end of 'db'). 2405194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 2406194206Ssimon 2407194206Ssimon *) Partial backport from 0.9.9-dev: 2408194206Ssimon 2409194206Ssimon Introduce bn_mul_mont (dedicated Montgomery multiplication 2410194206Ssimon procedure) as a candidate for BIGNUM assembler implementation. 2411194206Ssimon While 0.9.9-dev uses assembler for various architectures, only 2412194206Ssimon x86_64 is available by default here in the 0.9.8 branch, and 2413194206Ssimon 32-bit x86 is available through a compile-time setting. 2414194206Ssimon 2415194206Ssimon To try the 32-bit x86 assembler implementation, use Configure 2416194206Ssimon option "enable-montasm" (which exists only for this backport). 2417194206Ssimon 2418194206Ssimon As "enable-montasm" for 32-bit x86 disclaims code stability 2419194206Ssimon anyway, in this constellation we activate additional code 2420194206Ssimon backported from 0.9.9-dev for further performance improvements, 2421194206Ssimon namely BN_from_montgomery_word. (To enable this otherwise, 2422194206Ssimon e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".) 2423194206Ssimon 2424194206Ssimon [Andy Polyakov (backport partially by Bodo Moeller)] 2425194206Ssimon 2426194206Ssimon *) Add TLS session ticket callback. This allows an application to set 2427194206Ssimon TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed 2428194206Ssimon values. This is useful for key rollover for example where several key 2429194206Ssimon sets may exist with different names. 2430194206Ssimon [Steve Henson] 2431194206Ssimon 2432194206Ssimon *) Reverse ENGINE-internal logic for caching default ENGINE handles. 2433194206Ssimon This was broken until now in 0.9.8 releases, such that the only way 2434194206Ssimon a registered ENGINE could be used (assuming it initialises 2435194206Ssimon successfully on the host) was to explicitly set it as the default 2436194206Ssimon for the relevant algorithms. This is in contradiction with 0.9.7 2437194206Ssimon behaviour and the documentation. With this fix, when an ENGINE is 2438194206Ssimon registered into a given algorithm's table of implementations, the 2439194206Ssimon 'uptodate' flag is reset so that auto-discovery will be used next 2440194206Ssimon time a new context for that algorithm attempts to select an 2441194206Ssimon implementation. 2442194206Ssimon [Ian Lister (tweaked by Geoff Thorpe)] 2443194206Ssimon 2444194206Ssimon *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 2445194206Ssimon implemention in the following ways: 2446194206Ssimon 2447194206Ssimon Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be 2448194206Ssimon hard coded. 2449194206Ssimon 2450194206Ssimon Lack of BER streaming support means one pass streaming processing is 2451194206Ssimon only supported if data is detached: setting the streaming flag is 2452194206Ssimon ignored for embedded content. 2453194206Ssimon 2454194206Ssimon CMS support is disabled by default and must be explicitly enabled 2455194206Ssimon with the enable-cms configuration option. 2456194206Ssimon [Steve Henson] 2457194206Ssimon 2458194206Ssimon *) Update the GMP engine glue to do direct copies between BIGNUM and 2459194206Ssimon mpz_t when openssl and GMP use the same limb size. Otherwise the 2460194206Ssimon existing "conversion via a text string export" trick is still used. 2461194206Ssimon [Paul Sheer <paulsheer@gmail.com>] 2462194206Ssimon 2463194206Ssimon *) Zlib compression BIO. This is a filter BIO which compressed and 2464194206Ssimon uncompresses any data passed through it. 2465194206Ssimon [Steve Henson] 2466194206Ssimon 2467194206Ssimon *) Add AES_wrap_key() and AES_unwrap_key() functions to implement 2468194206Ssimon RFC3394 compatible AES key wrapping. 2469194206Ssimon [Steve Henson] 2470194206Ssimon 2471194206Ssimon *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): 2472194206Ssimon sets string data without copying. X509_ALGOR_set0() and 2473194206Ssimon X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) 2474194206Ssimon data. Attribute function X509at_get0_data_by_OBJ(): retrieves data 2475194206Ssimon from an X509_ATTRIBUTE structure optionally checking it occurs only 2476194206Ssimon once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied 2477194206Ssimon data. 2478194206Ssimon [Steve Henson] 2479194206Ssimon 2480194206Ssimon *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() 2481194206Ssimon to get the expected BN_FLG_CONSTTIME behavior. 2482194206Ssimon [Bodo Moeller (Google)] 2483194206Ssimon 2484194206Ssimon *) Netware support: 2485194206Ssimon 2486194206Ssimon - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets 2487194206Ssimon - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) 2488194206Ssimon - added some more tests to do_tests.pl 2489194206Ssimon - fixed RunningProcess usage so that it works with newer LIBC NDKs too 2490194206Ssimon - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency 2491194206Ssimon - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, 2492194206Ssimon netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc 2493194206Ssimon - various changes to netware.pl to enable gcc-cross builds on Win32 2494194206Ssimon platform 2495194206Ssimon - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) 2496194206Ssimon - various changes to fix missing prototype warnings 2497194206Ssimon - fixed x86nasm.pl to create correct asm files for NASM COFF output 2498194206Ssimon - added AES, WHIRLPOOL and CPUID assembler code to build files 2499194206Ssimon - added missing AES assembler make rules to mk1mf.pl 2500194206Ssimon - fixed order of includes in apps/ocsp.c so that e_os.h settings apply 2501194206Ssimon [Guenter Knauf <eflash@gmx.net>] 2502194206Ssimon 2503194206Ssimon *) Implement certificate status request TLS extension defined in RFC3546. 2504194206Ssimon A client can set the appropriate parameters and receive the encoded 2505194206Ssimon OCSP response via a callback. A server can query the supplied parameters 2506194206Ssimon and set the encoded OCSP response in the callback. Add simplified examples 2507194206Ssimon to s_client and s_server. 2508194206Ssimon [Steve Henson] 2509194206Ssimon 2510194206Ssimon Changes between 0.9.8f and 0.9.8g [19 Oct 2007] 2511194206Ssimon 2512194206Ssimon *) Fix various bugs: 2513194206Ssimon + Binary incompatibility of ssl_ctx_st structure 2514194206Ssimon + DTLS interoperation with non-compliant servers 2515194206Ssimon + Don't call get_session_cb() without proposed session 2516194206Ssimon + Fix ia64 assembler code 2517194206Ssimon [Andy Polyakov, Steve Henson] 2518194206Ssimon 2519194206Ssimon Changes between 0.9.8e and 0.9.8f [11 Oct 2007] 2520194206Ssimon 2521194206Ssimon *) DTLS Handshake overhaul. There were longstanding issues with 2522194206Ssimon OpenSSL DTLS implementation, which were making it impossible for 2523194206Ssimon RFC 4347 compliant client to communicate with OpenSSL server. 2524194206Ssimon Unfortunately just fixing these incompatibilities would "cut off" 2525194206Ssimon pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e 2526194206Ssimon server keeps tolerating non RFC compliant syntax. The opposite is 2527194206Ssimon not true, 0.9.8f client can not communicate with earlier server. 2528194206Ssimon This update even addresses CVE-2007-4995. 2529194206Ssimon [Andy Polyakov] 2530194206Ssimon 2531194206Ssimon *) Changes to avoid need for function casts in OpenSSL: some compilers 2532194206Ssimon (gcc 4.2 and later) reject their use. 2533194206Ssimon [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, 2534194206Ssimon Steve Henson] 2535194206Ssimon 2536194206Ssimon *) Add RFC4507 support to OpenSSL. This includes the corrections in 2537194206Ssimon RFC4507bis. The encrypted ticket format is an encrypted encoded 2538194206Ssimon SSL_SESSION structure, that way new session features are automatically 2539194206Ssimon supported. 2540194206Ssimon 2541194206Ssimon If a client application caches session in an SSL_SESSION structure 2542194206Ssimon support is transparent because tickets are now stored in the encoded 2543194206Ssimon SSL_SESSION. 2544194206Ssimon 2545194206Ssimon The SSL_CTX structure automatically generates keys for ticket 2546194206Ssimon protection in servers so again support should be possible 2547194206Ssimon with no application modification. 2548194206Ssimon 2549194206Ssimon If a client or server wishes to disable RFC4507 support then the option 2550194206Ssimon SSL_OP_NO_TICKET can be set. 2551194206Ssimon 2552194206Ssimon Add a TLS extension debugging callback to allow the contents of any client 2553194206Ssimon or server extensions to be examined. 2554194206Ssimon 2555194206Ssimon This work was sponsored by Google. 2556194206Ssimon [Steve Henson] 2557194206Ssimon 2558194206Ssimon *) Add initial support for TLS extensions, specifically for the server_name 2559194206Ssimon extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 2560194206Ssimon have new members for a host name. The SSL data structure has an 2561194206Ssimon additional member SSL_CTX *initial_ctx so that new sessions can be 2562194206Ssimon stored in that context to allow for session resumption, even after the 2563194206Ssimon SSL has been switched to a new SSL_CTX in reaction to a client's 2564194206Ssimon server_name extension. 2565194206Ssimon 2566194206Ssimon New functions (subject to change): 2567194206Ssimon 2568194206Ssimon SSL_get_servername() 2569194206Ssimon SSL_get_servername_type() 2570194206Ssimon SSL_set_SSL_CTX() 2571194206Ssimon 2572194206Ssimon New CTRL codes and macros (subject to change): 2573194206Ssimon 2574194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 2575194206Ssimon - SSL_CTX_set_tlsext_servername_callback() 2576194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 2577194206Ssimon - SSL_CTX_set_tlsext_servername_arg() 2578194206Ssimon SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 2579194206Ssimon 2580194206Ssimon openssl s_client has a new '-servername ...' option. 2581194206Ssimon 2582194206Ssimon openssl s_server has new options '-servername_host ...', '-cert2 ...', 2583194206Ssimon '-key2 ...', '-servername_fatal' (subject to change). This allows 2584194206Ssimon testing the HostName extension for a specific single host name ('-cert' 2585194206Ssimon and '-key' remain fallbacks for handshakes without HostName 2586194206Ssimon negotiation). If the unrecogninzed_name alert has to be sent, this by 2587194206Ssimon default is a warning; it becomes fatal with the '-servername_fatal' 2588194206Ssimon option. 2589194206Ssimon 2590194206Ssimon [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson] 2591194206Ssimon 2592194206Ssimon *) Add AES and SSE2 assembly language support to VC++ build. 2593194206Ssimon [Steve Henson] 2594194206Ssimon 2595194206Ssimon *) Mitigate attack on final subtraction in Montgomery reduction. 2596194206Ssimon [Andy Polyakov] 2597194206Ssimon 2598194206Ssimon *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 2599194206Ssimon (which previously caused an internal error). 2600194206Ssimon [Bodo Moeller] 2601194206Ssimon 2602194206Ssimon *) Squeeze another 10% out of IGE mode when in != out. 2603194206Ssimon [Ben Laurie] 2604194206Ssimon 2605194206Ssimon *) AES IGE mode speedup. 2606194206Ssimon [Dean Gaudet (Google)] 2607194206Ssimon 2608194206Ssimon *) Add the Korean symmetric 128-bit cipher SEED (see 2609194206Ssimon http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and 2610194206Ssimon add SEED ciphersuites from RFC 4162: 2611194206Ssimon 2612194206Ssimon TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" 2613194206Ssimon TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" 2614194206Ssimon TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" 2615194206Ssimon TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" 2616194206Ssimon 2617194206Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 2618194206Ssimon series, SEED remains excluded from compilation unless OpenSSL 2619194206Ssimon is configured with 'enable-seed'. 2620194206Ssimon [KISA, Bodo Moeller] 2621194206Ssimon 2622194206Ssimon *) Mitigate branch prediction attacks, which can be practical if a 2623194206Ssimon single processor is shared, allowing a spy process to extract 2624194206Ssimon information. For detailed background information, see 2625194206Ssimon http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron, 2626194206Ssimon J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL 2627194206Ssimon and Necessary Software Countermeasures"). The core of the change 2628194206Ssimon are new versions BN_div_no_branch() and 2629194206Ssimon BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), 2630194206Ssimon respectively, which are slower, but avoid the security-relevant 2631194206Ssimon conditional branches. These are automatically called by BN_div() 2632194206Ssimon and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one 2633194206Ssimon of the input BIGNUMs. Also, BN_is_bit_set() has been changed to 2634194206Ssimon remove a conditional branch. 2635194206Ssimon 2636194206Ssimon BN_FLG_CONSTTIME is the new name for the previous 2637194206Ssimon BN_FLG_EXP_CONSTTIME flag, since it now affects more than just 2638194206Ssimon modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag 2639194206Ssimon in the exponent causes BN_mod_exp_mont() to use the alternative 2640194206Ssimon implementation in BN_mod_exp_mont_consttime().) The old name 2641194206Ssimon remains as a deprecated alias. 2642194206Ssimon 2643194206Ssimon Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general 2644194206Ssimon RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses 2645194206Ssimon constant-time implementations for more than just exponentiation. 2646194206Ssimon Here too the old name is kept as a deprecated alias. 2647194206Ssimon 2648194206Ssimon BN_BLINDING_new() will now use BN_dup() for the modulus so that 2649194206Ssimon the BN_BLINDING structure gets an independent copy of the 2650194206Ssimon modulus. This means that the previous "BIGNUM *m" argument to 2651194206Ssimon BN_BLINDING_new() and to BN_BLINDING_create_param() now 2652194206Ssimon essentially becomes "const BIGNUM *m", although we can't actually 2653194206Ssimon change this in the header file before 0.9.9. It allows 2654194206Ssimon RSA_setup_blinding() to use BN_with_flags() on the modulus to 2655194206Ssimon enable BN_FLG_CONSTTIME. 2656194206Ssimon 2657194206Ssimon [Matthew D Wood (Intel Corp)] 2658194206Ssimon 2659194206Ssimon *) In the SSL/TLS server implementation, be strict about session ID 2660194206Ssimon context matching (which matters if an application uses a single 2661194206Ssimon external cache for different purposes). Previously, 2662194206Ssimon out-of-context reuse was forbidden only if SSL_VERIFY_PEER was 2663194206Ssimon set. This did ensure strict client verification, but meant that, 2664194206Ssimon with applications using a single external cache for quite 2665194206Ssimon different requirements, clients could circumvent ciphersuite 2666194206Ssimon restrictions for a given session ID context by starting a session 2667194206Ssimon in a different context. 2668194206Ssimon [Bodo Moeller] 2669194206Ssimon 2670167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 2671167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 2672167612Ssimon authentication-only ciphersuites. 2673167612Ssimon [Bodo Moeller] 2674167612Ssimon 2675194206Ssimon *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was 2676194206Ssimon not complete and could lead to a possible single byte overflow 2677194206Ssimon (CVE-2007-5135) [Ben Laurie] 2678194206Ssimon 2679194206Ssimon Changes between 0.9.8d and 0.9.8e [23 Feb 2007] 2680194206Ssimon 2681167612Ssimon *) Since AES128 and AES256 (and similarly Camellia128 and 2682167612Ssimon Camellia256) share a single mask bit in the logic of 2683167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 2684167612Ssimon kludge to work properly if AES128 is available and AES256 isn't 2685167612Ssimon (or if Camellia128 is available and Camellia256 isn't). 2686167612Ssimon [Victor Duchovni] 2687167612Ssimon 2688167612Ssimon *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c 2689167612Ssimon (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): 2690167612Ssimon When a point or a seed is encoded in a BIT STRING, we need to 2691167612Ssimon prevent the removal of trailing zero bits to get the proper DER 2692167612Ssimon encoding. (By default, crypto/asn1/a_bitstr.c assumes the case 2693167612Ssimon of a NamedBitList, for which trailing 0 bits need to be removed.) 2694167612Ssimon [Bodo Moeller] 2695167612Ssimon 2696167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 2697167612Ssimon protocol version while receiving ClientHello even if the 2698167612Ssimon ClientHello is fragmented. (The server can't insist on the 2699167612Ssimon particular protocol version it has chosen before the ServerHello 2700167612Ssimon message has informed the client about his choice.) 2701167612Ssimon [Bodo Moeller] 2702167612Ssimon 2703167612Ssimon *) Add RFC 3779 support. 2704167612Ssimon [Rob Austein for ARIN, Ben Laurie] 2705167612Ssimon 2706167612Ssimon *) Load error codes if they are not already present instead of using a 2707167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 2708167612Ssimon Improve header file function name parsing. 2709167612Ssimon [Steve Henson] 2710167612Ssimon 2711167612Ssimon *) extend SMTP and IMAP protocol emulation in s_client to use EHLO 2712167612Ssimon or CAPABILITY handshake as required by RFCs. 2713167612Ssimon [Goetz Babin-Ebell] 2714167612Ssimon 2715162911Ssimon Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 2716162911Ssimon 2717162911Ssimon *) Introduce limits to prevent malicious keys being able to 2718162911Ssimon cause a denial of service. (CVE-2006-2940) 2719162911Ssimon [Steve Henson, Bodo Moeller] 2720162911Ssimon 2721162911Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 2722162911Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 2723162911Ssimon 2724162911Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 2725162911Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 2726162911Ssimon 2727162911Ssimon *) Fix SSL client code which could crash if connecting to a 2728162911Ssimon malicious SSLv2 server. (CVE-2006-4343) 2729162911Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 2730162911Ssimon 2731162911Ssimon *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 2732162911Ssimon match only those. Before that, "AES256-SHA" would be interpreted 2733162911Ssimon as a pattern and match "AES128-SHA" too (since AES128-SHA got 2734162911Ssimon the same strength classification in 0.9.7h) as we currently only 2735162911Ssimon have a single AES bit in the ciphersuite description bitmap. 2736162911Ssimon That change, however, also applied to ciphersuite strings such as 2737162911Ssimon "RC4-MD5" that intentionally matched multiple ciphersuites -- 2738162911Ssimon namely, SSL 2.0 ciphersuites in addition to the more common ones 2739162911Ssimon from SSL 3.0/TLS 1.0. 2740162911Ssimon 2741162911Ssimon So we change the selection algorithm again: Naming an explicit 2742162911Ssimon ciphersuite selects this one ciphersuite, and any other similar 2743162911Ssimon ciphersuite (same bitmap) from *other* protocol versions. 2744162911Ssimon Thus, "RC4-MD5" again will properly select both the SSL 2.0 2745162911Ssimon ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 2746162911Ssimon 2747162911Ssimon Since SSL 2.0 does not have any ciphersuites for which the 2748162911Ssimon 128/256 bit distinction would be relevant, this works for now. 2749162911Ssimon The proper fix will be to use different bits for AES128 and 2750162911Ssimon AES256, which would have avoided the problems from the beginning; 2751162911Ssimon however, bits are scarce, so we can only do this in a new release 2752162911Ssimon (not just a patchlevel) when we can change the SSL_CIPHER 2753162911Ssimon definition to split the single 'unsigned long mask' bitmap into 2754162911Ssimon multiple values to extend the available space. 2755162911Ssimon 2756162911Ssimon [Bodo Moeller] 2757162911Ssimon 2758162911Ssimon Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 2759162911Ssimon 2760162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 2761162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 2762162911Ssimon 2763162911Ssimon *) Add AES IGE and biIGE modes. 2764162911Ssimon [Ben Laurie] 2765162911Ssimon 2766162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 2767162911Ssimon possible instead of select(), since the latter has some 2768162911Ssimon undesirable limitations. 2769162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 2770162911Ssimon 2771162911Ssimon *) Disable "ECCdraft" ciphersuites more thoroughly. Now special 2772162911Ssimon treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 2773162911Ssimon cannot be implicitly activated as part of, e.g., the "AES" alias. 2774162911Ssimon However, please upgrade to OpenSSL 0.9.9[-dev] for 2775162911Ssimon non-experimental use of the ECC ciphersuites to get TLS extension 2776162911Ssimon support, which is required for curve and point format negotiation 2777162911Ssimon to avoid potential handshake problems. 2778162911Ssimon [Bodo Moeller] 2779162911Ssimon 2780162911Ssimon *) Disable rogue ciphersuites: 2781162911Ssimon 2782162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 2783162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 2784162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 2785162911Ssimon 2786162911Ssimon The latter two were purportedly from 2787162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 2788162911Ssimon appear there. 2789162911Ssimon 2790167612Ssimon Also deactivate the remaining ciphersuites from 2791162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 2792162911Ssimon unofficial, and the ID has long expired. 2793162911Ssimon [Bodo Moeller] 2794162911Ssimon 2795162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 2796162911Ssimon dual-core machines) and other potential thread-safety issues. 2797162911Ssimon [Bodo Moeller] 2798162911Ssimon 2799162911Ssimon *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 2800162911Ssimon versions), which is now available for royalty-free use 2801162911Ssimon (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). 2802162911Ssimon Also, add Camellia TLS ciphersuites from RFC 4132. 2803162911Ssimon 2804162911Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 2805162911Ssimon series, Camellia remains excluded from compilation unless OpenSSL 2806162911Ssimon is configured with 'enable-camellia'. 2807162911Ssimon [NTT] 2808162911Ssimon 2809162911Ssimon *) Disable the padding bug check when compression is in use. The padding 2810162911Ssimon bug check assumes the first packet is of even length, this is not 2811162911Ssimon necessarily true if compresssion is enabled and can result in false 2812162911Ssimon positives causing handshake failure. The actual bug test is ancient 2813162911Ssimon code so it is hoped that implementations will either have fixed it by 2814162911Ssimon now or any which still have the bug do not support compression. 2815162911Ssimon [Steve Henson] 2816162911Ssimon 2817160814Ssimon Changes between 0.9.8a and 0.9.8b [04 May 2006] 2818160814Ssimon 2819160814Ssimon *) When applying a cipher rule check to see if string match is an explicit 2820160814Ssimon cipher suite and only match that one cipher suite if it is. 2821160814Ssimon [Steve Henson] 2822160814Ssimon 2823160814Ssimon *) Link in manifests for VC++ if needed. 2824160814Ssimon [Austin Ziegler <halostatue@gmail.com>] 2825160814Ssimon 2826160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 2827160814Ssimon draft-ietf-tls-ecc-12.txt with proposed changes (but without 2828160814Ssimon TLS extensions, which are supported starting with the 0.9.9 2829160814Ssimon branch, not in the OpenSSL 0.9.8 branch). 2830160814Ssimon [Douglas Stebila] 2831160814Ssimon 2832160814Ssimon *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 2833160814Ssimon opaque EVP_CIPHER_CTX handling. 2834160814Ssimon [Steve Henson] 2835160814Ssimon 2836160814Ssimon *) Fixes and enhancements to zlib compression code. We now only use 2837160814Ssimon "zlib1.dll" and use the default __cdecl calling convention on Win32 2838160814Ssimon to conform with the standards mentioned here: 2839160814Ssimon http://www.zlib.net/DLL_FAQ.txt 2840160814Ssimon Static zlib linking now works on Windows and the new --with-zlib-include 2841160814Ssimon --with-zlib-lib options to Configure can be used to supply the location 2842160814Ssimon of the headers and library. Gracefully handle case where zlib library 2843160814Ssimon can't be loaded. 2844160814Ssimon [Steve Henson] 2845160814Ssimon 2846160814Ssimon *) Several fixes and enhancements to the OID generation code. The old code 2847160814Ssimon sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 2848160814Ssimon handle numbers larger than ULONG_MAX, truncated printing and had a 2849160814Ssimon non standard OBJ_obj2txt() behaviour. 2850160814Ssimon [Steve Henson] 2851160814Ssimon 2852160814Ssimon *) Add support for building of engines under engine/ as shared libraries 2853160814Ssimon under VC++ build system. 2854160814Ssimon [Steve Henson] 2855160814Ssimon 2856160814Ssimon *) Corrected the numerous bugs in the Win32 path splitter in DSO. 2857160814Ssimon Hopefully, we will not see any false combination of paths any more. 2858160814Ssimon [Richard Levitte] 2859160814Ssimon 2860160814Ssimon Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 2861160814Ssimon 2862160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 2863160814Ssimon (part of SSL_OP_ALL). This option used to disable the 2864160814Ssimon countermeasure against man-in-the-middle protocol-version 2865160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 2866160814Ssimon idea. (CVE-2005-2969) 2867160814Ssimon 2868160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 2869160814Ssimon for Information Security, National Institute of Advanced Industrial 2870160814Ssimon Science and Technology [AIST], Japan)] 2871160814Ssimon 2872160814Ssimon *) Add two function to clear and return the verify parameter flags. 2873160814Ssimon [Steve Henson] 2874160814Ssimon 2875160814Ssimon *) Keep cipherlists sorted in the source instead of sorting them at 2876160814Ssimon runtime, thus removing the need for a lock. 2877160814Ssimon [Nils Larsch] 2878160814Ssimon 2879160814Ssimon *) Avoid some small subgroup attacks in Diffie-Hellman. 2880160814Ssimon [Nick Mathewson and Ben Laurie] 2881160814Ssimon 2882160814Ssimon *) Add functions for well-known primes. 2883160814Ssimon [Nick Mathewson] 2884160814Ssimon 2885160814Ssimon *) Extended Windows CE support. 2886160814Ssimon [Satoshi Nakamura and Andy Polyakov] 2887160814Ssimon 2888160814Ssimon *) Initialize SSL_METHOD structures at compile time instead of during 2889160814Ssimon runtime, thus removing the need for a lock. 2890160814Ssimon [Steve Henson] 2891160814Ssimon 2892160814Ssimon *) Make PKCS7_decrypt() work even if no certificate is supplied by 2893160814Ssimon attempting to decrypt each encrypted key in turn. Add support to 2894160814Ssimon smime utility. 2895160814Ssimon [Steve Henson] 2896160814Ssimon 2897160814Ssimon Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 2898160814Ssimon 2899162911Ssimon [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 2900162911Ssimon OpenSSL 0.9.8.] 2901162911Ssimon 2902160814Ssimon *) Add libcrypto.pc and libssl.pc for those who feel they need them. 2903160814Ssimon [Richard Levitte] 2904160814Ssimon 2905160814Ssimon *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 2906160814Ssimon key into the same file any more. 2907160814Ssimon [Richard Levitte] 2908160814Ssimon 2909160814Ssimon *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 2910160814Ssimon [Andy Polyakov] 2911160814Ssimon 2912160814Ssimon *) Add -utf8 command line and config file option to 'ca'. 2913160814Ssimon [Stefan <stf@udoma.org] 2914160814Ssimon 2915160814Ssimon *) Removed the macro des_crypt(), as it seems to conflict with some 2916160814Ssimon libraries. Use DES_crypt(). 2917160814Ssimon [Richard Levitte] 2918160814Ssimon 2919160814Ssimon *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 2920160814Ssimon involves renaming the source and generated shared-libs for 2921160814Ssimon both. The engines will accept the corrected or legacy ids 2922160814Ssimon ('ncipher' and '4758_cca' respectively) when binding. NB, 2923160814Ssimon this only applies when building 'shared'. 2924160814Ssimon [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 2925160814Ssimon 2926160814Ssimon *) Add attribute functions to EVP_PKEY structure. Modify 2927160814Ssimon PKCS12_create() to recognize a CSP name attribute and 2928160814Ssimon use it. Make -CSP option work again in pkcs12 utility. 2929160814Ssimon [Steve Henson] 2930160814Ssimon 2931160814Ssimon *) Add new functionality to the bn blinding code: 2932160814Ssimon - automatic re-creation of the BN_BLINDING parameters after 2933160814Ssimon a fixed number of uses (currently 32) 2934160814Ssimon - add new function for parameter creation 2935160814Ssimon - introduce flags to control the update behaviour of the 2936160814Ssimon BN_BLINDING parameters 2937160814Ssimon - hide BN_BLINDING structure 2938160814Ssimon Add a second BN_BLINDING slot to the RSA structure to improve 2939160814Ssimon performance when a single RSA object is shared among several 2940160814Ssimon threads. 2941160814Ssimon [Nils Larsch] 2942160814Ssimon 2943160814Ssimon *) Add support for DTLS. 2944160814Ssimon [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 2945160814Ssimon 2946160814Ssimon *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 2947160814Ssimon to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 2948160814Ssimon [Walter Goulet] 2949160814Ssimon 2950160814Ssimon *) Remove buggy and incompletet DH cert support from 2951160814Ssimon ssl/ssl_rsa.c and ssl/s3_both.c 2952160814Ssimon [Nils Larsch] 2953160814Ssimon 2954160814Ssimon *) Use SHA-1 instead of MD5 as the default digest algorithm for 2955160814Ssimon the apps/openssl applications. 2956160814Ssimon [Nils Larsch] 2957160814Ssimon 2958160814Ssimon *) Compile clean with "-Wall -Wmissing-prototypes 2959160814Ssimon -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 2960160814Ssimon DEBUG_SAFESTACK must also be set. 2961160814Ssimon [Ben Laurie] 2962160814Ssimon 2963160814Ssimon *) Change ./Configure so that certain algorithms can be disabled by default. 2964160814Ssimon The new counterpiece to "no-xxx" is "enable-xxx". 2965160814Ssimon 2966160814Ssimon The patented RC5 and MDC2 algorithms will now be disabled unless 2967160814Ssimon "enable-rc5" and "enable-mdc2", respectively, are specified. 2968160814Ssimon 2969160814Ssimon (IDEA remains enabled despite being patented. This is because IDEA 2970160814Ssimon is frequently required for interoperability, and there is no license 2971160814Ssimon fee for non-commercial use. As before, "no-idea" can be used to 2972160814Ssimon avoid this algorithm.) 2973160814Ssimon 2974160814Ssimon [Bodo Moeller] 2975160814Ssimon 2976160814Ssimon *) Add processing of proxy certificates (see RFC 3820). This work was 2977160814Ssimon sponsored by KTH (The Royal Institute of Technology in Stockholm) and 2978160814Ssimon EGEE (Enabling Grids for E-science in Europe). 2979160814Ssimon [Richard Levitte] 2980160814Ssimon 2981160814Ssimon *) RC4 performance overhaul on modern architectures/implementations, such 2982160814Ssimon as Intel P4, IA-64 and AMD64. 2983160814Ssimon [Andy Polyakov] 2984160814Ssimon 2985160814Ssimon *) New utility extract-section.pl. This can be used specify an alternative 2986160814Ssimon section number in a pod file instead of having to treat each file as 2987160814Ssimon a separate case in Makefile. This can be done by adding two lines to the 2988160814Ssimon pod file: 2989160814Ssimon 2990160814Ssimon =for comment openssl_section:XXX 2991160814Ssimon 2992160814Ssimon The blank line is mandatory. 2993160814Ssimon 2994160814Ssimon [Steve Henson] 2995160814Ssimon 2996160814Ssimon *) New arguments -certform, -keyform and -pass for s_client and s_server 2997160814Ssimon to allow alternative format key and certificate files and passphrase 2998160814Ssimon sources. 2999160814Ssimon [Steve Henson] 3000160814Ssimon 3001160814Ssimon *) New structure X509_VERIFY_PARAM which combines current verify parameters, 3002160814Ssimon update associated structures and add various utility functions. 3003160814Ssimon 3004160814Ssimon Add new policy related verify parameters, include policy checking in 3005160814Ssimon standard verify code. Enhance 'smime' application with extra parameters 3006160814Ssimon to support policy checking and print out. 3007160814Ssimon [Steve Henson] 3008160814Ssimon 3009160814Ssimon *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 3010160814Ssimon Nehemiah processors. These extensions support AES encryption in hardware 3011160814Ssimon as well as RNG (though RNG support is currently disabled). 3012160814Ssimon [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 3013160814Ssimon 3014160814Ssimon *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 3015160814Ssimon [Geoff Thorpe] 3016160814Ssimon 3017160814Ssimon *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 3018160814Ssimon [Andy Polyakov and a number of other people] 3019160814Ssimon 3020160814Ssimon *) Improved PowerPC platform support. Most notably BIGNUM assembler 3021160814Ssimon implementation contributed by IBM. 3022160814Ssimon [Suresh Chari, Peter Waltenberg, Andy Polyakov] 3023160814Ssimon 3024160814Ssimon *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 3025160814Ssimon exponent rather than 'unsigned long'. There is a corresponding change to 3026160814Ssimon the new 'rsa_keygen' element of the RSA_METHOD structure. 3027160814Ssimon [Jelte Jansen, Geoff Thorpe] 3028160814Ssimon 3029160814Ssimon *) Functionality for creating the initial serial number file is now 3030160814Ssimon moved from CA.pl to the 'ca' utility with a new option -create_serial. 3031160814Ssimon 3032160814Ssimon (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 3033160814Ssimon number file to 1, which is bound to cause problems. To avoid 3034160814Ssimon the problems while respecting compatibility between different 0.9.7 3035160814Ssimon patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 3036160814Ssimon CA.pl for serial number initialization. With the new release 0.9.8, 3037160814Ssimon we can fix the problem directly in the 'ca' utility.) 3038160814Ssimon [Steve Henson] 3039160814Ssimon 3040160814Ssimon *) Reduced header interdepencies by declaring more opaque objects in 3041160814Ssimon ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 3042160814Ssimon give fewer recursive includes, which could break lazy source code - so 3043160814Ssimon this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 3044160814Ssimon developers should define this symbol when building and using openssl to 3045160814Ssimon ensure they track the recommended behaviour, interfaces, [etc], but 3046160814Ssimon backwards-compatible behaviour prevails when this isn't defined. 3047160814Ssimon [Geoff Thorpe] 3048160814Ssimon 3049160814Ssimon *) New function X509_POLICY_NODE_print() which prints out policy nodes. 3050160814Ssimon [Steve Henson] 3051160814Ssimon 3052160814Ssimon *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 3053160814Ssimon This will generate a random key of the appropriate length based on the 3054160814Ssimon cipher context. The EVP_CIPHER can provide its own random key generation 3055160814Ssimon routine to support keys of a specific form. This is used in the des and 3056160814Ssimon 3des routines to generate a key of the correct parity. Update S/MIME 3057160814Ssimon code to use new functions and hence generate correct parity DES keys. 3058160814Ssimon Add EVP_CHECK_DES_KEY #define to return an error if the key is not 3059160814Ssimon valid (weak or incorrect parity). 3060160814Ssimon [Steve Henson] 3061160814Ssimon 3062160814Ssimon *) Add a local set of CRLs that can be used by X509_verify_cert() as well 3063160814Ssimon as looking them up. This is useful when the verified structure may contain 3064160814Ssimon CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 3065160814Ssimon present unless the new PKCS7_NO_CRL flag is asserted. 3066160814Ssimon [Steve Henson] 3067160814Ssimon 3068160814Ssimon *) Extend ASN1 oid configuration module. It now additionally accepts the 3069160814Ssimon syntax: 3070160814Ssimon 3071160814Ssimon shortName = some long name, 1.2.3.4 3072160814Ssimon [Steve Henson] 3073160814Ssimon 3074160814Ssimon *) Reimplemented the BN_CTX implementation. There is now no more static 3075160814Ssimon limitation on the number of variables it can handle nor the depth of the 3076160814Ssimon "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 3077160814Ssimon information can now expand as required, and rather than having a single 3078160814Ssimon static array of bignums, BN_CTX now uses a linked-list of such arrays 3079160814Ssimon allowing it to expand on demand whilst maintaining the usefulness of 3080160814Ssimon BN_CTX's "bundling". 3081160814Ssimon [Geoff Thorpe] 3082160814Ssimon 3083160814Ssimon *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 3084160814Ssimon to allow all RSA operations to function using a single BN_CTX. 3085160814Ssimon [Geoff Thorpe] 3086160814Ssimon 3087160814Ssimon *) Preliminary support for certificate policy evaluation and checking. This 3088160814Ssimon is initially intended to pass the tests outlined in "Conformance Testing 3089160814Ssimon of Relying Party Client Certificate Path Processing Logic" v1.07. 3090160814Ssimon [Steve Henson] 3091160814Ssimon 3092160814Ssimon *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 3093160814Ssimon remained unused and not that useful. A variety of other little bignum 3094160814Ssimon tweaks and fixes have also been made continuing on from the audit (see 3095160814Ssimon below). 3096160814Ssimon [Geoff Thorpe] 3097160814Ssimon 3098160814Ssimon *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 3099160814Ssimon associated ASN1, EVP and SSL functions and old ASN1 macros. 3100160814Ssimon [Richard Levitte] 3101160814Ssimon 3102160814Ssimon *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 3103160814Ssimon and this should never fail. So the return value from the use of 3104160814Ssimon BN_set_word() (which can fail due to needless expansion) is now deprecated; 3105160814Ssimon if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 3106160814Ssimon [Geoff Thorpe] 3107160814Ssimon 3108160814Ssimon *) BN_CTX_get() should return zero-valued bignums, providing the same 3109160814Ssimon initialised value as BN_new(). 3110160814Ssimon [Geoff Thorpe, suggested by Ulf M�ller] 3111160814Ssimon 3112160814Ssimon *) Support for inhibitAnyPolicy certificate extension. 3113160814Ssimon [Steve Henson] 3114160814Ssimon 3115160814Ssimon *) An audit of the BIGNUM code is underway, for which debugging code is 3116160814Ssimon enabled when BN_DEBUG is defined. This makes stricter enforcements on what 3117160814Ssimon is considered valid when processing BIGNUMs, and causes execution to 3118160814Ssimon assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 3119160814Ssimon further steps are taken to deliberately pollute unused data in BIGNUM 3120160814Ssimon structures to try and expose faulty code further on. For now, openssl will 3121160814Ssimon (in its default mode of operation) continue to tolerate the inconsistent 3122160814Ssimon forms that it has tolerated in the past, but authors and packagers should 3123160814Ssimon consider trying openssl and their own applications when compiled with 3124160814Ssimon these debugging symbols defined. It will help highlight potential bugs in 3125160814Ssimon their own code, and will improve the test coverage for OpenSSL itself. At 3126160814Ssimon some point, these tighter rules will become openssl's default to improve 3127160814Ssimon maintainability, though the assert()s and other overheads will remain only 3128160814Ssimon in debugging configurations. See bn.h for more details. 3129160814Ssimon [Geoff Thorpe, Nils Larsch, Ulf M�ller] 3130160814Ssimon 3131160814Ssimon *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 3132160814Ssimon that can only be obtained through BN_CTX_new() (which implicitly 3133160814Ssimon initialises it). The presence of this function only made it possible 3134160814Ssimon to overwrite an existing structure (and cause memory leaks). 3135160814Ssimon [Geoff Thorpe] 3136160814Ssimon 3137160814Ssimon *) Because of the callback-based approach for implementing LHASH as a 3138160814Ssimon template type, lh_insert() adds opaque objects to hash-tables and 3139160814Ssimon lh_doall() or lh_doall_arg() are typically used with a destructor callback 3140160814Ssimon to clean up those corresponding objects before destroying the hash table 3141160814Ssimon (and losing the object pointers). So some over-zealous constifications in 3142160814Ssimon LHASH have been relaxed so that lh_insert() does not take (nor store) the 3143160814Ssimon objects as "const" and the lh_doall[_arg] callback wrappers are not 3144160814Ssimon prototyped to have "const" restrictions on the object pointers they are 3145160814Ssimon given (and so aren't required to cast them away any more). 3146160814Ssimon [Geoff Thorpe] 3147160814Ssimon 3148160814Ssimon *) The tmdiff.h API was so ugly and minimal that our own timing utility 3149160814Ssimon (speed) prefers to use its own implementation. The two implementations 3150160814Ssimon haven't been consolidated as yet (volunteers?) but the tmdiff API has had 3151160814Ssimon its object type properly exposed (MS_TM) instead of casting to/from "char 3152160814Ssimon *". This may still change yet if someone realises MS_TM and "ms_time_***" 3153160814Ssimon aren't necessarily the greatest nomenclatures - but this is what was used 3154160814Ssimon internally to the implementation so I've used that for now. 3155160814Ssimon [Geoff Thorpe] 3156160814Ssimon 3157160814Ssimon *) Ensure that deprecated functions do not get compiled when 3158160814Ssimon OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 3159160814Ssimon the self-tests were still using deprecated key-generation functions so 3160160814Ssimon these have been updated also. 3161160814Ssimon [Geoff Thorpe] 3162160814Ssimon 3163160814Ssimon *) Reorganise PKCS#7 code to separate the digest location functionality 3164160814Ssimon into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 3165160814Ssimon New function PKCS7_set_digest() to set the digest type for PKCS#7 3166160814Ssimon digestedData type. Add additional code to correctly generate the 3167160814Ssimon digestedData type and add support for this type in PKCS7 initialization 3168160814Ssimon functions. 3169160814Ssimon [Steve Henson] 3170160814Ssimon 3171160814Ssimon *) New function PKCS7_set0_type_other() this initializes a PKCS7 3172160814Ssimon structure of type "other". 3173160814Ssimon [Steve Henson] 3174160814Ssimon 3175160814Ssimon *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 3176160814Ssimon sure the loop does correctly stop and breaking ("division by zero") 3177160814Ssimon modulus operations are not performed. The (pre-generated) prime 3178160814Ssimon table crypto/bn/bn_prime.h was already correct, but it could not be 3179160814Ssimon re-generated on some platforms because of the "division by zero" 3180160814Ssimon situation in the script. 3181160814Ssimon [Ralf S. Engelschall] 3182160814Ssimon 3183160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 3184160814Ssimon draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 3185160814Ssimon SHA-1 now is only used for "small" curves (where the 3186160814Ssimon representation of a field element takes up to 24 bytes); for 3187160814Ssimon larger curves, the field element resulting from ECDH is directly 3188160814Ssimon used as premaster secret. 3189160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 3190160814Ssimon 3191160814Ssimon *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 3192160814Ssimon curve secp160r1 to the tests. 3193160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 3194160814Ssimon 3195160814Ssimon *) Add the possibility to load symbols globally with DSO. 3196160814Ssimon [G�tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 3197160814Ssimon 3198160814Ssimon *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 3199160814Ssimon control of the error stack. 3200160814Ssimon [Richard Levitte] 3201160814Ssimon 3202160814Ssimon *) Add support for STORE in ENGINE. 3203160814Ssimon [Richard Levitte] 3204160814Ssimon 3205160814Ssimon *) Add the STORE type. The intention is to provide a common interface 3206160814Ssimon to certificate and key stores, be they simple file-based stores, or 3207160814Ssimon HSM-type store, or LDAP stores, or... 3208160814Ssimon NOTE: The code is currently UNTESTED and isn't really used anywhere. 3209160814Ssimon [Richard Levitte] 3210160814Ssimon 3211160814Ssimon *) Add a generic structure called OPENSSL_ITEM. This can be used to 3212160814Ssimon pass a list of arguments to any function as well as provide a way 3213160814Ssimon for a function to pass data back to the caller. 3214160814Ssimon [Richard Levitte] 3215160814Ssimon 3216160814Ssimon *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 3217160814Ssimon works like BUF_strdup() but can be used to duplicate a portion of 3218160814Ssimon a string. The copy gets NUL-terminated. BUF_memdup() duplicates 3219160814Ssimon a memory area. 3220160814Ssimon [Richard Levitte] 3221160814Ssimon 3222160814Ssimon *) Add the function sk_find_ex() which works like sk_find(), but will 3223160814Ssimon return an index to an element even if an exact match couldn't be 3224160814Ssimon found. The index is guaranteed to point at the element where the 3225160814Ssimon searched-for key would be inserted to preserve sorting order. 3226160814Ssimon [Richard Levitte] 3227160814Ssimon 3228160814Ssimon *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 3229160814Ssimon takes an extra flags argument for optional functionality. Currently, 3230160814Ssimon the following flags are defined: 3231160814Ssimon 3232160814Ssimon OBJ_BSEARCH_VALUE_ON_NOMATCH 3233160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 3234160814Ssimon element where the comparing function returns a negative or zero 3235160814Ssimon number. 3236160814Ssimon 3237160814Ssimon OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 3238160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 3239160814Ssimon element where the comparing function returns zero. This is useful 3240160814Ssimon if there are more than one element where the comparing function 3241160814Ssimon returns zero. 3242160814Ssimon [Richard Levitte] 3243160814Ssimon 3244160814Ssimon *) Make it possible to create self-signed certificates with 'openssl ca' 3245160814Ssimon in such a way that the self-signed certificate becomes part of the 3246160814Ssimon CA database and uses the same mechanisms for serial number generation 3247160814Ssimon as all other certificate signing. The new flag '-selfsign' enables 3248160814Ssimon this functionality. Adapt CA.sh and CA.pl.in. 3249160814Ssimon [Richard Levitte] 3250160814Ssimon 3251160814Ssimon *) Add functionality to check the public key of a certificate request 3252160814Ssimon against a given private. This is useful to check that a certificate 3253160814Ssimon request can be signed by that key (self-signing). 3254160814Ssimon [Richard Levitte] 3255160814Ssimon 3256160814Ssimon *) Make it possible to have multiple active certificates with the same 3257160814Ssimon subject in the CA index file. This is done only if the keyword 3258160814Ssimon 'unique_subject' is set to 'no' in the main CA section (default 3259160814Ssimon if 'CA_default') of the configuration file. The value is saved 3260160814Ssimon with the database itself in a separate index attribute file, 3261160814Ssimon named like the index file with '.attr' appended to the name. 3262160814Ssimon [Richard Levitte] 3263160814Ssimon 3264160814Ssimon *) Generate muti valued AVAs using '+' notation in config files for 3265160814Ssimon req and dirName. 3266160814Ssimon [Steve Henson] 3267160814Ssimon 3268160814Ssimon *) Support for nameConstraints certificate extension. 3269160814Ssimon [Steve Henson] 3270160814Ssimon 3271160814Ssimon *) Support for policyConstraints certificate extension. 3272160814Ssimon [Steve Henson] 3273160814Ssimon 3274160814Ssimon *) Support for policyMappings certificate extension. 3275160814Ssimon [Steve Henson] 3276160814Ssimon 3277160814Ssimon *) Make sure the default DSA_METHOD implementation only uses its 3278160814Ssimon dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 3279160814Ssimon and change its own handlers to be NULL so as to remove unnecessary 3280160814Ssimon indirection. This lets alternative implementations fallback to the 3281160814Ssimon default implementation more easily. 3282160814Ssimon [Geoff Thorpe] 3283160814Ssimon 3284160814Ssimon *) Support for directoryName in GeneralName related extensions 3285160814Ssimon in config files. 3286160814Ssimon [Steve Henson] 3287160814Ssimon 3288160814Ssimon *) Make it possible to link applications using Makefile.shared. 3289160814Ssimon Make that possible even when linking against static libraries! 3290160814Ssimon [Richard Levitte] 3291160814Ssimon 3292160814Ssimon *) Support for single pass processing for S/MIME signing. This now 3293160814Ssimon means that S/MIME signing can be done from a pipe, in addition 3294160814Ssimon cleartext signing (multipart/signed type) is effectively streaming 3295160814Ssimon and the signed data does not need to be all held in memory. 3296160814Ssimon 3297160814Ssimon This is done with a new flag PKCS7_STREAM. When this flag is set 3298160814Ssimon PKCS7_sign() only initializes the PKCS7 structure and the actual signing 3299160814Ssimon is done after the data is output (and digests calculated) in 3300160814Ssimon SMIME_write_PKCS7(). 3301160814Ssimon [Steve Henson] 3302160814Ssimon 3303160814Ssimon *) Add full support for -rpath/-R, both in shared libraries and 3304160814Ssimon applications, at least on the platforms where it's known how 3305160814Ssimon to do it. 3306160814Ssimon [Richard Levitte] 3307160814Ssimon 3308160814Ssimon *) In crypto/ec/ec_mult.c, implement fast point multiplication with 3309160814Ssimon precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 3310160814Ssimon will now compute a table of multiples of the generator that 3311160814Ssimon makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 3312160814Ssimon faster (notably in the case of a single point multiplication, 3313160814Ssimon scalar * generator). 3314160814Ssimon [Nils Larsch, Bodo Moeller] 3315160814Ssimon 3316160814Ssimon *) IPv6 support for certificate extensions. The various extensions 3317160814Ssimon which use the IP:a.b.c.d can now take IPv6 addresses using the 3318160814Ssimon formats of RFC1884 2.2 . IPv6 addresses are now also displayed 3319160814Ssimon correctly. 3320160814Ssimon [Steve Henson] 3321160814Ssimon 3322160814Ssimon *) Added an ENGINE that implements RSA by performing private key 3323160814Ssimon exponentiations with the GMP library. The conversions to and from 3324160814Ssimon GMP's mpz_t format aren't optimised nor are any montgomery forms 3325160814Ssimon cached, and on x86 it appears OpenSSL's own performance has caught up. 3326160814Ssimon However there are likely to be other architectures where GMP could 3327160814Ssimon provide a boost. This ENGINE is not built in by default, but it can be 3328160814Ssimon specified at Configure time and should be accompanied by the necessary 3329160814Ssimon linker additions, eg; 3330160814Ssimon ./config -DOPENSSL_USE_GMP -lgmp 3331160814Ssimon [Geoff Thorpe] 3332160814Ssimon 3333160814Ssimon *) "openssl engine" will not display ENGINE/DSO load failure errors when 3334160814Ssimon testing availability of engines with "-t" - the old behaviour is 3335160814Ssimon produced by increasing the feature's verbosity with "-tt". 3336160814Ssimon [Geoff Thorpe] 3337160814Ssimon 3338160814Ssimon *) ECDSA routines: under certain error conditions uninitialized BN objects 3339160814Ssimon could be freed. Solution: make sure initialization is performed early 3340160814Ssimon enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 3341160814Ssimon via PR#459) 3342160814Ssimon [Lutz Jaenicke] 3343160814Ssimon 3344160814Ssimon *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 3345160814Ssimon and DH_METHOD (eg. by ENGINE implementations) to override the normal 3346160814Ssimon software implementations. For DSA and DH, parameter generation can 3347160814Ssimon also be overriden by providing the appropriate method callbacks. 3348160814Ssimon [Geoff Thorpe] 3349160814Ssimon 3350160814Ssimon *) Change the "progress" mechanism used in key-generation and 3351160814Ssimon primality testing to functions that take a new BN_GENCB pointer in 3352160814Ssimon place of callback/argument pairs. The new API functions have "_ex" 3353160814Ssimon postfixes and the older functions are reimplemented as wrappers for 3354160814Ssimon the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 3355160814Ssimon declarations of the old functions to help (graceful) attempts to 3356160814Ssimon migrate to the new functions. Also, the new key-generation API 3357160814Ssimon functions operate on a caller-supplied key-structure and return 3358160814Ssimon success/failure rather than returning a key or NULL - this is to 3359160814Ssimon help make "keygen" another member function of RSA_METHOD etc. 3360160814Ssimon 3361160814Ssimon Example for using the new callback interface: 3362160814Ssimon 3363160814Ssimon int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 3364160814Ssimon void *my_arg = ...; 3365160814Ssimon BN_GENCB my_cb; 3366160814Ssimon 3367160814Ssimon BN_GENCB_set(&my_cb, my_callback, my_arg); 3368160814Ssimon 3369160814Ssimon return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 3370160814Ssimon /* For the meaning of a, b in calls to my_callback(), see the 3371160814Ssimon * documentation of the function that calls the callback. 3372160814Ssimon * cb will point to my_cb; my_arg can be retrieved as cb->arg. 3373160814Ssimon * my_callback should return 1 if it wants BN_is_prime_ex() 3374160814Ssimon * to continue, or 0 to stop. 3375160814Ssimon */ 3376160814Ssimon 3377160814Ssimon [Geoff Thorpe] 3378160814Ssimon 3379160814Ssimon *) Change the ZLIB compression method to be stateful, and make it 3380160814Ssimon available to TLS with the number defined in 3381160814Ssimon draft-ietf-tls-compression-04.txt. 3382160814Ssimon [Richard Levitte] 3383160814Ssimon 3384160814Ssimon *) Add the ASN.1 structures and functions for CertificatePair, which 3385160814Ssimon is defined as follows (according to X.509_4thEditionDraftV6.pdf): 3386160814Ssimon 3387160814Ssimon CertificatePair ::= SEQUENCE { 3388160814Ssimon forward [0] Certificate OPTIONAL, 3389160814Ssimon reverse [1] Certificate OPTIONAL, 3390160814Ssimon -- at least one of the pair shall be present -- } 3391160814Ssimon 3392160814Ssimon Also implement the PEM functions to read and write certificate 3393160814Ssimon pairs, and defined the PEM tag as "CERTIFICATE PAIR". 3394160814Ssimon 3395160814Ssimon This needed to be defined, mostly for the sake of the LDAP 3396160814Ssimon attribute crossCertificatePair, but may prove useful elsewhere as 3397160814Ssimon well. 3398160814Ssimon [Richard Levitte] 3399160814Ssimon 3400160814Ssimon *) Make it possible to inhibit symlinking of shared libraries in 3401160814Ssimon Makefile.shared, for Cygwin's sake. 3402160814Ssimon [Richard Levitte] 3403160814Ssimon 3404160814Ssimon *) Extend the BIGNUM API by creating a function 3405160814Ssimon void BN_set_negative(BIGNUM *a, int neg); 3406160814Ssimon and a macro that behave like 3407160814Ssimon int BN_is_negative(const BIGNUM *a); 3408160814Ssimon 3409160814Ssimon to avoid the need to access 'a->neg' directly in applications. 3410160814Ssimon [Nils Larsch] 3411160814Ssimon 3412160814Ssimon *) Implement fast modular reduction for pseudo-Mersenne primes 3413160814Ssimon used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 3414160814Ssimon EC_GROUP_new_curve_GFp() will now automatically use this 3415160814Ssimon if applicable. 3416160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3417160814Ssimon 3418160814Ssimon *) Add new lock type (CRYPTO_LOCK_BN). 3419160814Ssimon [Bodo Moeller] 3420160814Ssimon 3421160814Ssimon *) Change the ENGINE framework to automatically load engines 3422160814Ssimon dynamically from specific directories unless they could be 3423160814Ssimon found to already be built in or loaded. Move all the 3424160814Ssimon current engines except for the cryptodev one to a new 3425160814Ssimon directory engines/. 3426160814Ssimon The engines in engines/ are built as shared libraries if 3427160814Ssimon the "shared" options was given to ./Configure or ./config. 3428160814Ssimon Otherwise, they are inserted in libcrypto.a. 3429160814Ssimon /usr/local/ssl/engines is the default directory for dynamic 3430160814Ssimon engines, but that can be overriden at configure time through 3431160814Ssimon the usual use of --prefix and/or --openssldir, and at run 3432160814Ssimon time with the environment variable OPENSSL_ENGINES. 3433160814Ssimon [Geoff Thorpe and Richard Levitte] 3434160814Ssimon 3435160814Ssimon *) Add Makefile.shared, a helper makefile to build shared 3436160814Ssimon libraries. Addapt Makefile.org. 3437160814Ssimon [Richard Levitte] 3438160814Ssimon 3439160814Ssimon *) Add version info to Win32 DLLs. 3440160814Ssimon [Peter 'Luna' Runestig" <peter@runestig.com>] 3441160814Ssimon 3442160814Ssimon *) Add new 'medium level' PKCS#12 API. Certificates and keys 3443160814Ssimon can be added using this API to created arbitrary PKCS#12 3444160814Ssimon files while avoiding the low level API. 3445160814Ssimon 3446160814Ssimon New options to PKCS12_create(), key or cert can be NULL and 3447160814Ssimon will then be omitted from the output file. The encryption 3448160814Ssimon algorithm NIDs can be set to -1 for no encryption, the mac 3449160814Ssimon iteration count can be set to 0 to omit the mac. 3450160814Ssimon 3451160814Ssimon Enhance pkcs12 utility by making the -nokeys and -nocerts 3452160814Ssimon options work when creating a PKCS#12 file. New option -nomac 3453160814Ssimon to omit the mac, NONE can be set for an encryption algorithm. 3454160814Ssimon New code is modified to use the enhanced PKCS12_create() 3455160814Ssimon instead of the low level API. 3456160814Ssimon [Steve Henson] 3457160814Ssimon 3458160814Ssimon *) Extend ASN1 encoder to support indefinite length constructed 3459160814Ssimon encoding. This can output sequences tags and octet strings in 3460160814Ssimon this form. Modify pk7_asn1.c to support indefinite length 3461160814Ssimon encoding. This is experimental and needs additional code to 3462160814Ssimon be useful, such as an ASN1 bio and some enhanced streaming 3463160814Ssimon PKCS#7 code. 3464160814Ssimon 3465160814Ssimon Extend template encode functionality so that tagging is passed 3466160814Ssimon down to the template encoder. 3467160814Ssimon [Steve Henson] 3468160814Ssimon 3469160814Ssimon *) Let 'openssl req' fail if an argument to '-newkey' is not 3470160814Ssimon recognized instead of using RSA as a default. 3471160814Ssimon [Bodo Moeller] 3472160814Ssimon 3473160814Ssimon *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 3474160814Ssimon As these are not official, they are not included in "ALL"; 3475160814Ssimon the "ECCdraft" ciphersuite group alias can be used to select them. 3476160814Ssimon [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 3477160814Ssimon 3478160814Ssimon *) Add ECDH engine support. 3479160814Ssimon [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 3480160814Ssimon 3481160814Ssimon *) Add ECDH in new directory crypto/ecdh/. 3482160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 3483160814Ssimon 3484160814Ssimon *) Let BN_rand_range() abort with an error after 100 iterations 3485160814Ssimon without success (which indicates a broken PRNG). 3486160814Ssimon [Bodo Moeller] 3487160814Ssimon 3488160814Ssimon *) Change BN_mod_sqrt() so that it verifies that the input value 3489160814Ssimon is really the square of the return value. (Previously, 3490160814Ssimon BN_mod_sqrt would show GIGO behaviour.) 3491160814Ssimon [Bodo Moeller] 3492160814Ssimon 3493160814Ssimon *) Add named elliptic curves over binary fields from X9.62, SECG, 3494160814Ssimon and WAP/WTLS; add OIDs that were still missing. 3495160814Ssimon 3496160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3497160814Ssimon (Sun Microsystems Laboratories)] 3498160814Ssimon 3499160814Ssimon *) Extend the EC library for elliptic curves over binary fields 3500160814Ssimon (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 3501160814Ssimon New EC_METHOD: 3502160814Ssimon 3503160814Ssimon EC_GF2m_simple_method 3504160814Ssimon 3505160814Ssimon New API functions: 3506160814Ssimon 3507160814Ssimon EC_GROUP_new_curve_GF2m 3508160814Ssimon EC_GROUP_set_curve_GF2m 3509160814Ssimon EC_GROUP_get_curve_GF2m 3510160814Ssimon EC_POINT_set_affine_coordinates_GF2m 3511160814Ssimon EC_POINT_get_affine_coordinates_GF2m 3512160814Ssimon EC_POINT_set_compressed_coordinates_GF2m 3513160814Ssimon 3514160814Ssimon Point compression for binary fields is disabled by default for 3515160814Ssimon patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 3516160814Ssimon enable it). 3517160814Ssimon 3518160814Ssimon As binary polynomials are represented as BIGNUMs, various members 3519160814Ssimon of the EC_GROUP and EC_POINT data structures can be shared 3520160814Ssimon between the implementations for prime fields and binary fields; 3521160814Ssimon the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 3522160814Ssimon are essentially identical to their ..._GFp counterparts. 3523160814Ssimon (For simplicity, the '..._GFp' prefix has been dropped from 3524160814Ssimon various internal method names.) 3525160814Ssimon 3526160814Ssimon An internal 'field_div' method (similar to 'field_mul' and 3527160814Ssimon 'field_sqr') has been added; this is used only for binary fields. 3528160814Ssimon 3529160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3530160814Ssimon (Sun Microsystems Laboratories)] 3531160814Ssimon 3532160814Ssimon *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 3533160814Ssimon through methods ('mul', 'precompute_mult'). 3534160814Ssimon 3535160814Ssimon The generic implementations (now internally called 'ec_wNAF_mul' 3536160814Ssimon and 'ec_wNAF_precomputed_mult') remain the default if these 3537160814Ssimon methods are undefined. 3538160814Ssimon 3539160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3540160814Ssimon (Sun Microsystems Laboratories)] 3541160814Ssimon 3542160814Ssimon *) New function EC_GROUP_get_degree, which is defined through 3543160814Ssimon EC_METHOD. For curves over prime fields, this returns the bit 3544160814Ssimon length of the modulus. 3545160814Ssimon 3546160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3547160814Ssimon (Sun Microsystems Laboratories)] 3548160814Ssimon 3549160814Ssimon *) New functions EC_GROUP_dup, EC_POINT_dup. 3550160814Ssimon (These simply call ..._new and ..._copy). 3551160814Ssimon 3552160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3553160814Ssimon (Sun Microsystems Laboratories)] 3554160814Ssimon 3555160814Ssimon *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 3556160814Ssimon Polynomials are represented as BIGNUMs (where the sign bit is not 3557160814Ssimon used) in the following functions [macros]: 3558160814Ssimon 3559160814Ssimon BN_GF2m_add 3560160814Ssimon BN_GF2m_sub [= BN_GF2m_add] 3561160814Ssimon BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 3562160814Ssimon BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 3563160814Ssimon BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 3564160814Ssimon BN_GF2m_mod_inv 3565160814Ssimon BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 3566160814Ssimon BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 3567160814Ssimon BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 3568160814Ssimon BN_GF2m_cmp [= BN_ucmp] 3569160814Ssimon 3570160814Ssimon (Note that only the 'mod' functions are actually for fields GF(2^m). 3571160814Ssimon BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 3572160814Ssimon 3573160814Ssimon For some functions, an the irreducible polynomial defining a 3574160814Ssimon field can be given as an 'unsigned int[]' with strictly 3575160814Ssimon decreasing elements giving the indices of those bits that are set; 3576160814Ssimon i.e., p[] represents the polynomial 3577160814Ssimon f(t) = t^p[0] + t^p[1] + ... + t^p[k] 3578160814Ssimon where 3579160814Ssimon p[0] > p[1] > ... > p[k] = 0. 3580160814Ssimon This applies to the following functions: 3581160814Ssimon 3582160814Ssimon BN_GF2m_mod_arr 3583160814Ssimon BN_GF2m_mod_mul_arr 3584160814Ssimon BN_GF2m_mod_sqr_arr 3585160814Ssimon BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 3586160814Ssimon BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 3587160814Ssimon BN_GF2m_mod_exp_arr 3588160814Ssimon BN_GF2m_mod_sqrt_arr 3589160814Ssimon BN_GF2m_mod_solve_quad_arr 3590160814Ssimon BN_GF2m_poly2arr 3591160814Ssimon BN_GF2m_arr2poly 3592160814Ssimon 3593160814Ssimon Conversion can be performed by the following functions: 3594160814Ssimon 3595160814Ssimon BN_GF2m_poly2arr 3596160814Ssimon BN_GF2m_arr2poly 3597160814Ssimon 3598160814Ssimon bntest.c has additional tests for binary polynomial arithmetic. 3599160814Ssimon 3600160814Ssimon Two implementations for BN_GF2m_mod_div() are available. 3601160814Ssimon The default algorithm simply uses BN_GF2m_mod_inv() and 3602160814Ssimon BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 3603160814Ssimon if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 3604160814Ssimon copyright notice in crypto/bn/bn_gf2m.c before enabling it). 3605160814Ssimon 3606160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3607160814Ssimon (Sun Microsystems Laboratories)] 3608160814Ssimon 3609160814Ssimon *) Add new error code 'ERR_R_DISABLED' that can be used when some 3610160814Ssimon functionality is disabled at compile-time. 3611160814Ssimon [Douglas Stebila <douglas.stebila@sun.com>] 3612160814Ssimon 3613160814Ssimon *) Change default behaviour of 'openssl asn1parse' so that more 3614160814Ssimon information is visible when viewing, e.g., a certificate: 3615160814Ssimon 3616160814Ssimon Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 3617160814Ssimon mode the content of non-printable OCTET STRINGs is output in a 3618160814Ssimon style similar to INTEGERs, but with '[HEX DUMP]' prepended to 3619160814Ssimon avoid the appearance of a printable string. 3620160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3621160814Ssimon 3622160814Ssimon *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 3623160814Ssimon functions 3624160814Ssimon EC_GROUP_set_asn1_flag() 3625160814Ssimon EC_GROUP_get_asn1_flag() 3626160814Ssimon EC_GROUP_set_point_conversion_form() 3627160814Ssimon EC_GROUP_get_point_conversion_form() 3628160814Ssimon These control ASN1 encoding details: 3629160814Ssimon - Curves (i.e., groups) are encoded explicitly unless asn1_flag 3630160814Ssimon has been set to OPENSSL_EC_NAMED_CURVE. 3631160814Ssimon - Points are encoded in uncompressed form by default; options for 3632160814Ssimon asn1_for are as for point2oct, namely 3633160814Ssimon POINT_CONVERSION_COMPRESSED 3634160814Ssimon POINT_CONVERSION_UNCOMPRESSED 3635160814Ssimon POINT_CONVERSION_HYBRID 3636160814Ssimon 3637160814Ssimon Also add 'seed' and 'seed_len' members to EC_GROUP with access 3638160814Ssimon functions 3639160814Ssimon EC_GROUP_set_seed() 3640160814Ssimon EC_GROUP_get0_seed() 3641160814Ssimon EC_GROUP_get_seed_len() 3642160814Ssimon This is used only for ASN1 purposes (so far). 3643160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3644160814Ssimon 3645160814Ssimon *) Add 'field_type' member to EC_METHOD, which holds the NID 3646160814Ssimon of the appropriate field type OID. The new function 3647160814Ssimon EC_METHOD_get_field_type() returns this value. 3648160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3649160814Ssimon 3650160814Ssimon *) Add functions 3651160814Ssimon EC_POINT_point2bn() 3652160814Ssimon EC_POINT_bn2point() 3653160814Ssimon EC_POINT_point2hex() 3654160814Ssimon EC_POINT_hex2point() 3655160814Ssimon providing useful interfaces to EC_POINT_point2oct() and 3656160814Ssimon EC_POINT_oct2point(). 3657160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3658160814Ssimon 3659160814Ssimon *) Change internals of the EC library so that the functions 3660160814Ssimon EC_GROUP_set_generator() 3661160814Ssimon EC_GROUP_get_generator() 3662160814Ssimon EC_GROUP_get_order() 3663160814Ssimon EC_GROUP_get_cofactor() 3664160814Ssimon are implemented directly in crypto/ec/ec_lib.c and not dispatched 3665160814Ssimon to methods, which would lead to unnecessary code duplication when 3666160814Ssimon adding different types of curves. 3667160814Ssimon [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 3668160814Ssimon 3669160814Ssimon *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 3670160814Ssimon arithmetic, and such that modified wNAFs are generated 3671160814Ssimon (which avoid length expansion in many cases). 3672160814Ssimon [Bodo Moeller] 3673160814Ssimon 3674160814Ssimon *) Add a function EC_GROUP_check_discriminant() (defined via 3675160814Ssimon EC_METHOD) that verifies that the curve discriminant is non-zero. 3676160814Ssimon 3677160814Ssimon Add a function EC_GROUP_check() that makes some sanity tests 3678160814Ssimon on a EC_GROUP, its generator and order. This includes 3679160814Ssimon EC_GROUP_check_discriminant(). 3680160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3681160814Ssimon 3682160814Ssimon *) Add ECDSA in new directory crypto/ecdsa/. 3683160814Ssimon 3684160814Ssimon Add applications 'openssl ecparam' and 'openssl ecdsa' 3685160814Ssimon (these are based on 'openssl dsaparam' and 'openssl dsa'). 3686160814Ssimon 3687160814Ssimon ECDSA support is also included in various other files across the 3688160814Ssimon library. Most notably, 3689160814Ssimon - 'openssl req' now has a '-newkey ecdsa:file' option; 3690160814Ssimon - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 3691160814Ssimon - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 3692160814Ssimon d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 3693160814Ssimon them suitable for ECDSA where domain parameters must be 3694160814Ssimon extracted before the specific public key; 3695160814Ssimon - ECDSA engine support has been added. 3696160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3697160814Ssimon 3698160814Ssimon *) Include some named elliptic curves, and add OIDs from X9.62, 3699160814Ssimon SECG, and WAP/WTLS. Each curve can be obtained from the new 3700160814Ssimon function 3701160814Ssimon EC_GROUP_new_by_curve_name(), 3702160814Ssimon and the list of available named curves can be obtained with 3703160814Ssimon EC_get_builtin_curves(). 3704160814Ssimon Also add a 'curve_name' member to EC_GROUP objects, which can be 3705160814Ssimon accessed via 3706160814Ssimon EC_GROUP_set_curve_name() 3707160814Ssimon EC_GROUP_get_curve_name() 3708160814Ssimon [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 3709160814Ssimon 3710160814Ssimon *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 3711160814Ssimon was actually never needed) and in BN_mul(). The removal in BN_mul() 3712160814Ssimon required a small change in bn_mul_part_recursive() and the addition 3713160814Ssimon of the functions bn_cmp_part_words(), bn_sub_part_words() and 3714160814Ssimon bn_add_part_words(), which do the same thing as bn_cmp_words(), 3715160814Ssimon bn_sub_words() and bn_add_words() except they take arrays with 3716160814Ssimon differing sizes. 3717160814Ssimon [Richard Levitte] 3718160814Ssimon 3719194206Ssimon Changes between 0.9.7l and 0.9.7m [23 Feb 2007] 3720194206Ssimon 3721167612Ssimon *) Cleanse PEM buffers before freeing them since they may contain 3722167612Ssimon sensitive data. 3723167612Ssimon [Benjamin Bennett <ben@psc.edu>] 3724167612Ssimon 3725167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 3726167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 3727167612Ssimon authentication-only ciphersuites. 3728167612Ssimon [Bodo Moeller] 3729167612Ssimon 3730167612Ssimon *) Since AES128 and AES256 share a single mask bit in the logic of 3731167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 3732167612Ssimon kludge to work properly if AES128 is available and AES256 isn't. 3733167612Ssimon [Victor Duchovni] 3734167612Ssimon 3735194206Ssimon *) Expand security boundary to match 1.1.1 module. 3736194206Ssimon [Steve Henson] 3737194206Ssimon 3738194206Ssimon *) Remove redundant features: hash file source, editing of test vectors 3739194206Ssimon modify fipsld to use external fips_premain.c signature. 3740194206Ssimon [Steve Henson] 3741194206Ssimon 3742194206Ssimon *) New perl script mkfipsscr.pl to create shell scripts or batch files to 3743194206Ssimon run algorithm test programs. 3744194206Ssimon [Steve Henson] 3745194206Ssimon 3746194206Ssimon *) Make algorithm test programs more tolerant of whitespace. 3747194206Ssimon [Steve Henson] 3748194206Ssimon 3749167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 3750167612Ssimon protocol version while receiving ClientHello even if the 3751167612Ssimon ClientHello is fragmented. (The server can't insist on the 3752167612Ssimon particular protocol version it has chosen before the ServerHello 3753167612Ssimon message has informed the client about his choice.) 3754167612Ssimon [Bodo Moeller] 3755167612Ssimon 3756167612Ssimon *) Load error codes if they are not already present instead of using a 3757167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 3758167612Ssimon [Steve Henson] 3759167612Ssimon 3760167612Ssimon Changes between 0.9.7k and 0.9.7l [28 Sep 2006] 3761167612Ssimon 3762167612Ssimon *) Introduce limits to prevent malicious keys being able to 3763167612Ssimon cause a denial of service. (CVE-2006-2940) 3764167612Ssimon [Steve Henson, Bodo Moeller] 3765167612Ssimon 3766167612Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 3767167612Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 3768167612Ssimon 3769167612Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 3770167612Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 3771167612Ssimon 3772167612Ssimon *) Fix SSL client code which could crash if connecting to a 3773167612Ssimon malicious SSLv2 server. (CVE-2006-4343) 3774167612Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 3775167612Ssimon 3776162911Ssimon *) Change ciphersuite string processing so that an explicit 3777162911Ssimon ciphersuite selects this one ciphersuite (so that "AES256-SHA" 3778162911Ssimon will no longer include "AES128-SHA"), and any other similar 3779162911Ssimon ciphersuite (same bitmap) from *other* protocol versions (so that 3780162911Ssimon "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 3781162911Ssimon SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 3782162911Ssimon changes from 0.9.8b and 0.9.8d. 3783162911Ssimon [Bodo Moeller] 3784162911Ssimon 3785162911Ssimon Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 3786162911Ssimon 3787162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 3788162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 3789162911Ssimon 3790162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 3791162911Ssimon possible instead of select(), since the latter has some 3792162911Ssimon undesirable limitations. 3793162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 3794162911Ssimon 3795162911Ssimon *) Disable rogue ciphersuites: 3796162911Ssimon 3797162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 3798162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 3799162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 3800162911Ssimon 3801162911Ssimon The latter two were purportedly from 3802162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 3803162911Ssimon appear there. 3804162911Ssimon 3805162911Ssimon Also deactive the remaining ciphersuites from 3806162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 3807162911Ssimon unofficial, and the ID has long expired. 3808162911Ssimon [Bodo Moeller] 3809162911Ssimon 3810162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 3811162911Ssimon dual-core machines) and other potential thread-safety issues. 3812162911Ssimon [Bodo Moeller] 3813162911Ssimon 3814162911Ssimon Changes between 0.9.7i and 0.9.7j [04 May 2006] 3815162911Ssimon 3816162911Ssimon *) Adapt fipsld and the build system to link against the validated FIPS 3817162911Ssimon module in FIPS mode. 3818162911Ssimon [Steve Henson] 3819162911Ssimon 3820162911Ssimon *) Fixes for VC++ 2005 build under Windows. 3821162911Ssimon [Steve Henson] 3822162911Ssimon 3823162911Ssimon *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 3824162911Ssimon from a Windows bash shell such as MSYS. It is autodetected from the 3825162911Ssimon "config" script when run from a VC++ environment. Modify standard VC++ 3826162911Ssimon build to use fipscanister.o from the GNU make build. 3827162911Ssimon [Steve Henson] 3828162911Ssimon 3829160814Ssimon Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 3830160814Ssimon 3831160814Ssimon *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 3832160814Ssimon The value now differs depending on if you build for FIPS or not. 3833160814Ssimon BEWARE! A program linked with a shared FIPSed libcrypto can't be 3834160814Ssimon safely run with a non-FIPSed libcrypto, as it may crash because of 3835160814Ssimon the difference induced by this change. 3836160814Ssimon [Andy Polyakov] 3837160814Ssimon 3838160814Ssimon Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 3839160814Ssimon 3840160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 3841160814Ssimon (part of SSL_OP_ALL). This option used to disable the 3842160814Ssimon countermeasure against man-in-the-middle protocol-version 3843160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 3844160814Ssimon idea. (CVE-2005-2969) 3845160814Ssimon 3846160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 3847160814Ssimon for Information Security, National Institute of Advanced Industrial 3848160814Ssimon Science and Technology [AIST], Japan)] 3849160814Ssimon 3850160814Ssimon *) Minimal support for X9.31 signatures and PSS padding modes. This is 3851160814Ssimon mainly for FIPS compliance and not fully integrated at this stage. 3852160814Ssimon [Steve Henson] 3853160814Ssimon 3854160814Ssimon *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 3855160814Ssimon the exponentiation using a fixed-length exponent. (Otherwise, 3856160814Ssimon the information leaked through timing could expose the secret key 3857160814Ssimon after many signatures; cf. Bleichenbacher's attack on DSA with 3858160814Ssimon biased k.) 3859160814Ssimon [Bodo Moeller] 3860160814Ssimon 3861160814Ssimon *) Make a new fixed-window mod_exp implementation the default for 3862160814Ssimon RSA, DSA, and DH private-key operations so that the sequence of 3863160814Ssimon squares and multiplies and the memory access pattern are 3864160814Ssimon independent of the particular secret key. This will mitigate 3865160814Ssimon cache-timing and potential related attacks. 3866160814Ssimon 3867160814Ssimon BN_mod_exp_mont_consttime() is the new exponentiation implementation, 3868160814Ssimon and this is automatically used by BN_mod_exp_mont() if the new flag 3869160814Ssimon BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 3870160814Ssimon will use this BN flag for private exponents unless the flag 3871160814Ssimon RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 3872160814Ssimon DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 3873160814Ssimon 3874160814Ssimon [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 3875160814Ssimon 3876160814Ssimon *) Change the client implementation for SSLv23_method() and 3877160814Ssimon SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 3878160814Ssimon Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 3879160814Ssimon (Previously, the SSL 2.0 backwards compatible Client Hello 3880160814Ssimon message format would be used even with SSL_OP_NO_SSLv2.) 3881160814Ssimon [Bodo Moeller] 3882160814Ssimon 3883160814Ssimon *) Add support for smime-type MIME parameter in S/MIME messages which some 3884160814Ssimon clients need. 3885160814Ssimon [Steve Henson] 3886160814Ssimon 3887160814Ssimon *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 3888160814Ssimon a threadsafe manner. Modify rsa code to use new function and add calls 3889160814Ssimon to dsa and dh code (which had race conditions before). 3890160814Ssimon [Steve Henson] 3891160814Ssimon 3892160814Ssimon *) Include the fixed error library code in the C error file definitions 3893160814Ssimon instead of fixing them up at runtime. This keeps the error code 3894160814Ssimon structures constant. 3895160814Ssimon [Steve Henson] 3896160814Ssimon 3897160814Ssimon Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 3898160814Ssimon 3899160814Ssimon [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 3900160814Ssimon OpenSSL 0.9.8.] 3901160814Ssimon 3902160814Ssimon *) Fixes for newer kerberos headers. NB: the casts are needed because 3903160814Ssimon the 'length' field is signed on one version and unsigned on another 3904160814Ssimon with no (?) obvious way to tell the difference, without these VC++ 3905160814Ssimon complains. Also the "definition" of FAR (blank) is no longer included 3906160814Ssimon nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 3907160814Ssimon some needed definitions. 3908160814Ssimon [Steve Henson] 3909160814Ssimon 3910160814Ssimon *) Undo Cygwin change. 3911160814Ssimon [Ulf M�ller] 3912160814Ssimon 3913160814Ssimon *) Added support for proxy certificates according to RFC 3820. 3914160814Ssimon Because they may be a security thread to unaware applications, 3915160814Ssimon they must be explicitely allowed in run-time. See 3916160814Ssimon docs/HOWTO/proxy_certificates.txt for further information. 3917160814Ssimon [Richard Levitte] 3918160814Ssimon 3919160814Ssimon Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 3920160814Ssimon 3921160814Ssimon *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 3922160814Ssimon server and client random values. Previously 3923160814Ssimon (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 3924160814Ssimon less random data when sizeof(time_t) > 4 (some 64 bit platforms). 3925160814Ssimon 3926160814Ssimon This change has negligible security impact because: 3927160814Ssimon 3928160814Ssimon 1. Server and client random values still have 24 bytes of pseudo random 3929160814Ssimon data. 3930160814Ssimon 3931160814Ssimon 2. Server and client random values are sent in the clear in the initial 3932160814Ssimon handshake. 3933160814Ssimon 3934160814Ssimon 3. The master secret is derived using the premaster secret (48 bytes in 3935160814Ssimon size for static RSA ciphersuites) as well as client server and random 3936160814Ssimon values. 3937160814Ssimon 3938160814Ssimon The OpenSSL team would like to thank the UK NISCC for bringing this issue 3939160814Ssimon to our attention. 3940160814Ssimon 3941160814Ssimon [Stephen Henson, reported by UK NISCC] 3942160814Ssimon 3943160814Ssimon *) Use Windows randomness collection on Cygwin. 3944160814Ssimon [Ulf M�ller] 3945160814Ssimon 3946160814Ssimon *) Fix hang in EGD/PRNGD query when communication socket is closed 3947160814Ssimon prematurely by EGD/PRNGD. 3948160814Ssimon [Darren Tucker <dtucker@zip.com.au> via Lutz J�nicke, resolves #1014] 3949160814Ssimon 3950160814Ssimon *) Prompt for pass phrases when appropriate for PKCS12 input format. 3951160814Ssimon [Steve Henson] 3952160814Ssimon 3953160814Ssimon *) Back-port of selected performance improvements from development 3954160814Ssimon branch, as well as improved support for PowerPC platforms. 3955160814Ssimon [Andy Polyakov] 3956160814Ssimon 3957160814Ssimon *) Add lots of checks for memory allocation failure, error codes to indicate 3958160814Ssimon failure and freeing up memory if a failure occurs. 3959160814Ssimon [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 3960160814Ssimon 3961160814Ssimon *) Add new -passin argument to dgst. 3962160814Ssimon [Steve Henson] 3963160814Ssimon 3964160814Ssimon *) Perform some character comparisons of different types in X509_NAME_cmp: 3965160814Ssimon this is needed for some certificates that reencode DNs into UTF8Strings 3966160814Ssimon (in violation of RFC3280) and can't or wont issue name rollover 3967160814Ssimon certificates. 3968160814Ssimon [Steve Henson] 3969160814Ssimon 3970160814Ssimon *) Make an explicit check during certificate validation to see that 3971160814Ssimon the CA setting in each certificate on the chain is correct. As a 3972160814Ssimon side effect always do the following basic checks on extensions, 3973160814Ssimon not just when there's an associated purpose to the check: 3974160814Ssimon 3975160814Ssimon - if there is an unhandled critical extension (unless the user 3976160814Ssimon has chosen to ignore this fault) 3977160814Ssimon - if the path length has been exceeded (if one is set at all) 3978160814Ssimon - that certain extensions fit the associated purpose (if one has 3979160814Ssimon been given) 3980160814Ssimon [Richard Levitte] 3981160814Ssimon 3982142425Snectar Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 3983142425Snectar 3984142425Snectar *) Avoid a race condition when CRLs are checked in a multi threaded 3985142425Snectar environment. This would happen due to the reordering of the revoked 3986142425Snectar entries during signature checking and serial number lookup. Now the 3987142425Snectar encoding is cached and the serial number sort performed under a lock. 3988142425Snectar Add new STACK function sk_is_sorted(). 3989142425Snectar [Steve Henson] 3990142425Snectar 3991142425Snectar *) Add Delta CRL to the extension code. 3992142425Snectar [Steve Henson] 3993142425Snectar 3994142425Snectar *) Various fixes to s3_pkt.c so alerts are sent properly. 3995142425Snectar [David Holmes <d.holmes@f5.com>] 3996142425Snectar 3997142425Snectar *) Reduce the chances of duplicate issuer name and serial numbers (in 3998142425Snectar violation of RFC3280) using the OpenSSL certificate creation utilities. 3999142425Snectar This is done by creating a random 64 bit value for the initial serial 4000142425Snectar number when a serial number file is created or when a self signed 4001142425Snectar certificate is created using 'openssl req -x509'. The initial serial 4002142425Snectar number file is created using 'openssl x509 -next_serial' in CA.pl 4003142425Snectar rather than being initialized to 1. 4004142425Snectar [Steve Henson] 4005142425Snectar 4006127128Snectar Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 4007127128Snectar 4008127128Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 4009160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 4010127128Snectar [Joe Orton, Steve Henson] 4011127128Snectar 4012127128Snectar *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 4013160814Ssimon (CVE-2004-0112) 4014127128Snectar [Joe Orton, Steve Henson] 4015127128Snectar 4016127128Snectar *) Make it possible to have multiple active certificates with the same 4017127128Snectar subject in the CA index file. This is done only if the keyword 4018127128Snectar 'unique_subject' is set to 'no' in the main CA section (default 4019127128Snectar if 'CA_default') of the configuration file. The value is saved 4020127128Snectar with the database itself in a separate index attribute file, 4021127128Snectar named like the index file with '.attr' appended to the name. 4022127128Snectar [Richard Levitte] 4023127128Snectar 4024127128Snectar *) X509 verify fixes. Disable broken certificate workarounds when 4025127128Snectar X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 4026127128Snectar keyUsage extension present. Don't accept CRLs with unhandled critical 4027127128Snectar extensions: since verify currently doesn't process CRL extensions this 4028127128Snectar rejects a CRL with *any* critical extensions. Add new verify error codes 4029127128Snectar for these cases. 4030127128Snectar [Steve Henson] 4031127128Snectar 4032127128Snectar *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 4033127128Snectar A clarification of RFC2560 will require the use of OCTET STRINGs and 4034127128Snectar some implementations cannot handle the current raw format. Since OpenSSL 4035127128Snectar copies and compares OCSP nonces as opaque blobs without any attempt at 4036127128Snectar parsing them this should not create any compatibility issues. 4037127128Snectar [Steve Henson] 4038127128Snectar 4039127128Snectar *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 4040127128Snectar calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 4041127128Snectar this HMAC (and other) operations are several times slower than OpenSSL 4042127128Snectar < 0.9.7. 4043127128Snectar [Steve Henson] 4044127128Snectar 4045127128Snectar *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 4046127128Snectar [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 4047127128Snectar 4048127128Snectar *) Use the correct content when signing type "other". 4049127128Snectar [Steve Henson] 4050127128Snectar 4051120631Snectar Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 4052120631Snectar 4053120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 4054120631Snectar 4055120631Snectar Stop out of bounds reads in the ASN1 code when presented with 4056160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 4057120631Snectar 4058160814Ssimon Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 4059120631Snectar 4060120631Snectar If verify callback ignores invalid public key errors don't try to check 4061120631Snectar certificate signature with the NULL public key. 4062120631Snectar 4063120631Snectar [Steve Henson] 4064120631Snectar 4065120631Snectar *) New -ignore_err option in ocsp application to stop the server 4066120631Snectar exiting on the first error in a request. 4067120631Snectar [Steve Henson] 4068120631Snectar 4069120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 4070120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 4071120631Snectar specifications. 4072120631Snectar [Steve Henson] 4073120631Snectar 4074120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 4075120631Snectar extra data after the compression methods not only for TLS 1.0 4076120631Snectar but also for SSL 3.0 (as required by the specification). 4077120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 4078120631Snectar 4079120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 4080120631Snectar when it's 512 *bits* long, not 512 bytes. 4081120631Snectar [Richard Levitte] 4082120631Snectar 4083120631Snectar *) Change AES_cbc_encrypt() so it outputs exact multiple of 4084120631Snectar blocks during encryption. 4085120631Snectar [Richard Levitte] 4086120631Snectar 4087120631Snectar *) Various fixes to base64 BIO and non blocking I/O. On write 4088120631Snectar flushes were not handled properly if the BIO retried. On read 4089120631Snectar data was not being buffered properly and had various logic bugs. 4090120631Snectar This also affects blocking I/O when the data being decoded is a 4091120631Snectar certain size. 4092120631Snectar [Steve Henson] 4093120631Snectar 4094120631Snectar *) Various S/MIME bugfixes and compatibility changes: 4095120631Snectar output correct application/pkcs7 MIME type if 4096120631Snectar PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 4097120631Snectar Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 4098120631Snectar of files as .eml work). Correctly handle very long lines in MIME 4099120631Snectar parser. 4100120631Snectar [Steve Henson] 4101120631Snectar 4102120631Snectar Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 4103120631Snectar 4104120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 4105120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 4106120631Snectar a protocol version number mismatch like a decryption error 4107120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 4108120631Snectar [Bodo Moeller] 4109120631Snectar 4110120631Snectar *) Turn on RSA blinding by default in the default implementation 4111120631Snectar to avoid a timing attack. Applications that don't want it can call 4112120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 4113120631Snectar They would be ill-advised to do so in most cases. 4114120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 4115120631Snectar 4116120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 4117120631Snectar seeded (in this case, the secret RSA exponent is abused as 4118120631Snectar an unpredictable seed -- if it is not unpredictable, there 4119120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 4120120631Snectar by remembering the creator's thread ID in rsa->blinding and 4121120631Snectar having all other threads use local one-time blinding factors 4122120631Snectar (this requires more computation than sharing rsa->blinding, but 4123120631Snectar avoids excessive locking; and if an RSA object is not shared 4124120631Snectar between threads, blinding will still be very fast). 4125120631Snectar [Bodo Moeller] 4126120631Snectar 4127120631Snectar *) Fixed a typo bug that would cause ENGINE_set_default() to set an 4128120631Snectar ENGINE as defaults for all supported algorithms irrespective of 4129120631Snectar the 'flags' parameter. 'flags' is now honoured, so applications 4130120631Snectar should make sure they are passing it correctly. 4131120631Snectar [Geoff Thorpe] 4132120631Snectar 4133120631Snectar *) Target "mingw" now allows native Windows code to be generated in 4134120631Snectar the Cygwin environment as well as with the MinGW compiler. 4135120631Snectar [Ulf Moeller] 4136120631Snectar 4137111147Snectar Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 4138111147Snectar 4139111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 4140111147Snectar via timing by performing a MAC computation even if incorrrect 4141111147Snectar block cipher padding has been found. This is a countermeasure 4142111147Snectar against active attacks where the attacker has to distinguish 4143160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 4144111147Snectar 4145111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 4146111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 4147111147Snectar Martin Vuagnoux (EPFL, Ilion)] 4148111147Snectar 4149111147Snectar *) Make the no-err option work as intended. The intention with no-err 4150111147Snectar is not to have the whole error stack handling routines removed from 4151111147Snectar libcrypto, it's only intended to remove all the function name and 4152111147Snectar reason texts, thereby removing some of the footprint that may not 4153111147Snectar be interesting if those errors aren't displayed anyway. 4154111147Snectar 4155111147Snectar NOTE: it's still possible for any application or module to have it's 4156111147Snectar own set of error texts inserted. The routines are there, just not 4157111147Snectar used by default when no-err is given. 4158111147Snectar [Richard Levitte] 4159111147Snectar 4160111147Snectar *) Add support for FreeBSD on IA64. 4161111147Snectar [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 4162111147Snectar 4163111147Snectar *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 4164111147Snectar Kerberos function mit_des_cbc_cksum(). Before this change, 4165111147Snectar the value returned by DES_cbc_cksum() was like the one from 4166111147Snectar mit_des_cbc_cksum(), except the bytes were swapped. 4167111147Snectar [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 4168111147Snectar 4169111147Snectar *) Allow an application to disable the automatic SSL chain building. 4170111147Snectar Before this a rather primitive chain build was always performed in 4171111147Snectar ssl3_output_cert_chain(): an application had no way to send the 4172111147Snectar correct chain if the automatic operation produced an incorrect result. 4173111147Snectar 4174111147Snectar Now the chain builder is disabled if either: 4175111147Snectar 4176111147Snectar 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 4177111147Snectar 4178111147Snectar 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 4179111147Snectar 4180111147Snectar The reasoning behind this is that an application would not want the 4181111147Snectar auto chain building to take place if extra chain certificates are 4182111147Snectar present and it might also want a means of sending no additional 4183111147Snectar certificates (for example the chain has two certificates and the 4184111147Snectar root is omitted). 4185111147Snectar [Steve Henson] 4186111147Snectar 4187111147Snectar *) Add the possibility to build without the ENGINE framework. 4188111147Snectar [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 4189111147Snectar 4190111147Snectar *) Under Win32 gmtime() can return NULL: check return value in 4191111147Snectar OPENSSL_gmtime(). Add error code for case where gmtime() fails. 4192111147Snectar [Steve Henson] 4193111147Snectar 4194111147Snectar *) DSA routines: under certain error conditions uninitialized BN objects 4195111147Snectar could be freed. Solution: make sure initialization is performed early 4196111147Snectar enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 4197111147Snectar Nils Larsch <nla@trustcenter.de> via PR#459) 4198111147Snectar [Lutz Jaenicke] 4199111147Snectar 4200111147Snectar *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 4201111147Snectar checked on reconnect on the client side, therefore session resumption 4202111147Snectar could still fail with a "ssl session id is different" error. This 4203111147Snectar behaviour is masked when SSL_OP_ALL is used due to 4204111147Snectar SSL_OP_MICROSOFT_SESS_ID_BUG being set. 4205111147Snectar Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4206111147Snectar followup to PR #377. 4207111147Snectar [Lutz Jaenicke] 4208111147Snectar 4209111147Snectar *) IA-32 assembler support enhancements: unified ELF targets, support 4210111147Snectar for SCO/Caldera platforms, fix for Cygwin shared build. 4211111147Snectar [Andy Polyakov] 4212111147Snectar 4213111147Snectar *) Add support for FreeBSD on sparc64. As a consequence, support for 4214111147Snectar FreeBSD on non-x86 processors is separate from x86 processors on 4215111147Snectar the config script, much like the NetBSD support. 4216111147Snectar [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 4217111147Snectar 4218109998Smarkm Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 4219109998Smarkm 4220120631Snectar [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 4221120631Snectar OpenSSL 0.9.7.] 4222120631Snectar 4223109998Smarkm *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 4224109998Smarkm code (06) was taken as the first octet of the session ID and the last 4225109998Smarkm octet was ignored consequently. As a result SSLv2 client side session 4226109998Smarkm caching could not have worked due to the session ID mismatch between 4227109998Smarkm client and server. 4228109998Smarkm Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4229109998Smarkm PR #377. 4230109998Smarkm [Lutz Jaenicke] 4231109998Smarkm 4232109998Smarkm *) Change the declaration of needed Kerberos libraries to use EX_LIBS 4233109998Smarkm instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 4234109998Smarkm removed entirely. 4235109998Smarkm [Richard Levitte] 4236109998Smarkm 4237109998Smarkm *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 4238109998Smarkm seems that in spite of existing for more than a year, many application 4239109998Smarkm author have done nothing to provide the necessary callbacks, which 4240109998Smarkm means that this particular engine will not work properly anywhere. 4241109998Smarkm This is a very unfortunate situation which forces us, in the name 4242109998Smarkm of usability, to give the hw_ncipher.c a static lock, which is part 4243109998Smarkm of libcrypto. 4244109998Smarkm NOTE: This is for the 0.9.7 series ONLY. This hack will never 4245109998Smarkm appear in 0.9.8 or later. We EXPECT application authors to have 4246109998Smarkm dealt properly with this when 0.9.8 is released (unless we actually 4247109998Smarkm make such changes in the libcrypto locking code that changes will 4248109998Smarkm have to be made anyway). 4249109998Smarkm [Richard Levitte] 4250109998Smarkm 4251109998Smarkm *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 4252109998Smarkm octets have been read, EOF or an error occurs. Without this change 4253109998Smarkm some truncated ASN1 structures will not produce an error. 4254109998Smarkm [Steve Henson] 4255109998Smarkm 4256109998Smarkm *) Disable Heimdal support, since it hasn't been fully implemented. 4257109998Smarkm Still give the possibility to force the use of Heimdal, but with 4258109998Smarkm warnings and a request that patches get sent to openssl-dev. 4259109998Smarkm [Richard Levitte] 4260109998Smarkm 4261109998Smarkm *) Add the VC-CE target, introduce the WINCE sysname, and add 4262109998Smarkm INSTALL.WCE and appropriate conditionals to make it build. 4263109998Smarkm [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 4264109998Smarkm 4265109998Smarkm *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 4266109998Smarkm cygssl-x.y.z.dll, where x, y and z are the major, minor and 4267109998Smarkm edit numbers of the version. 4268109998Smarkm [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 4269109998Smarkm 4270109998Smarkm *) Introduce safe string copy and catenation functions 4271109998Smarkm (BUF_strlcpy() and BUF_strlcat()). 4272109998Smarkm [Ben Laurie (CHATS) and Richard Levitte] 4273109998Smarkm 4274109998Smarkm *) Avoid using fixed-size buffers for one-line DNs. 4275109998Smarkm [Ben Laurie (CHATS)] 4276109998Smarkm 4277109998Smarkm *) Add BUF_MEM_grow_clean() to avoid information leakage when 4278109998Smarkm resizing buffers containing secrets, and use where appropriate. 4279109998Smarkm [Ben Laurie (CHATS)] 4280109998Smarkm 4281109998Smarkm *) Avoid using fixed size buffers for configuration file location. 4282109998Smarkm [Ben Laurie (CHATS)] 4283109998Smarkm 4284109998Smarkm *) Avoid filename truncation for various CA files. 4285109998Smarkm [Ben Laurie (CHATS)] 4286109998Smarkm 4287109998Smarkm *) Use sizeof in preference to magic numbers. 4288109998Smarkm [Ben Laurie (CHATS)] 4289109998Smarkm 4290109998Smarkm *) Avoid filename truncation in cert requests. 4291109998Smarkm [Ben Laurie (CHATS)] 4292109998Smarkm 4293109998Smarkm *) Add assertions to check for (supposedly impossible) buffer 4294109998Smarkm overflows. 4295109998Smarkm [Ben Laurie (CHATS)] 4296109998Smarkm 4297109998Smarkm *) Don't cache truncated DNS entries in the local cache (this could 4298109998Smarkm potentially lead to a spoofing attack). 4299109998Smarkm [Ben Laurie (CHATS)] 4300109998Smarkm 4301109998Smarkm *) Fix various buffers to be large enough for hex/decimal 4302109998Smarkm representations in a platform independent manner. 4303109998Smarkm [Ben Laurie (CHATS)] 4304109998Smarkm 4305109998Smarkm *) Add CRYPTO_realloc_clean() to avoid information leakage when 4306109998Smarkm resizing buffers containing secrets, and use where appropriate. 4307109998Smarkm [Ben Laurie (CHATS)] 4308109998Smarkm 4309109998Smarkm *) Add BIO_indent() to avoid much slightly worrying code to do 4310109998Smarkm indents. 4311109998Smarkm [Ben Laurie (CHATS)] 4312109998Smarkm 4313109998Smarkm *) Convert sprintf()/BIO_puts() to BIO_printf(). 4314109998Smarkm [Ben Laurie (CHATS)] 4315109998Smarkm 4316109998Smarkm *) buffer_gets() could terminate with the buffer only half 4317109998Smarkm full. Fixed. 4318109998Smarkm [Ben Laurie (CHATS)] 4319109998Smarkm 4320109998Smarkm *) Add assertions to prevent user-supplied crypto functions from 4321109998Smarkm overflowing internal buffers by having large block sizes, etc. 4322109998Smarkm [Ben Laurie (CHATS)] 4323109998Smarkm 4324109998Smarkm *) New OPENSSL_assert() macro (similar to assert(), but enabled 4325109998Smarkm unconditionally). 4326109998Smarkm [Ben Laurie (CHATS)] 4327109998Smarkm 4328109998Smarkm *) Eliminate unused copy of key in RC4. 4329109998Smarkm [Ben Laurie (CHATS)] 4330109998Smarkm 4331109998Smarkm *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 4332109998Smarkm [Ben Laurie (CHATS)] 4333109998Smarkm 4334109998Smarkm *) Fix off-by-one error in EGD path. 4335109998Smarkm [Ben Laurie (CHATS)] 4336109998Smarkm 4337109998Smarkm *) If RANDFILE path is too long, ignore instead of truncating. 4338109998Smarkm [Ben Laurie (CHATS)] 4339109998Smarkm 4340109998Smarkm *) Eliminate unused and incorrectly sized X.509 structure 4341109998Smarkm CBCParameter. 4342109998Smarkm [Ben Laurie (CHATS)] 4343109998Smarkm 4344109998Smarkm *) Eliminate unused and dangerous function knumber(). 4345109998Smarkm [Ben Laurie (CHATS)] 4346109998Smarkm 4347109998Smarkm *) Eliminate unused and dangerous structure, KSSL_ERR. 4348109998Smarkm [Ben Laurie (CHATS)] 4349109998Smarkm 4350109998Smarkm *) Protect against overlong session ID context length in an encoded 4351109998Smarkm session object. Since these are local, this does not appear to be 4352109998Smarkm exploitable. 4353109998Smarkm [Ben Laurie (CHATS)] 4354109998Smarkm 4355109998Smarkm *) Change from security patch (see 0.9.6e below) that did not affect 4356109998Smarkm the 0.9.6 release series: 4357109998Smarkm 4358109998Smarkm Remote buffer overflow in SSL3 protocol - an attacker could 4359109998Smarkm supply an oversized master key in Kerberos-enabled versions. 4360160814Ssimon (CVE-2002-0657) 4361109998Smarkm [Ben Laurie (CHATS)] 4362109998Smarkm 4363109998Smarkm *) Change the SSL kerb5 codes to match RFC 2712. 4364109998Smarkm [Richard Levitte] 4365109998Smarkm 4366109998Smarkm *) Make -nameopt work fully for req and add -reqopt switch. 4367109998Smarkm [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 4368109998Smarkm 4369109998Smarkm *) The "block size" for block ciphers in CFB and OFB mode should be 1. 4370109998Smarkm [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 4371109998Smarkm 4372109998Smarkm *) Make sure tests can be performed even if the corresponding algorithms 4373109998Smarkm have been removed entirely. This was also the last step to make 4374109998Smarkm OpenSSL compilable with DJGPP under all reasonable conditions. 4375109998Smarkm [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 4376109998Smarkm 4377109998Smarkm *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 4378109998Smarkm to allow version independent disabling of normally unselected ciphers, 4379109998Smarkm which may be activated as a side-effect of selecting a single cipher. 4380109998Smarkm 4381109998Smarkm (E.g., cipher list string "RSA" enables ciphersuites that are left 4382109998Smarkm out of "ALL" because they do not provide symmetric encryption. 4383109998Smarkm "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 4384109998Smarkm [Lutz Jaenicke, Bodo Moeller] 4385109998Smarkm 4386109998Smarkm *) Add appropriate support for separate platform-dependent build 4387109998Smarkm directories. The recommended way to make a platform-dependent 4388109998Smarkm build directory is the following (tested on Linux), maybe with 4389109998Smarkm some local tweaks: 4390109998Smarkm 4391109998Smarkm # Place yourself outside of the OpenSSL source tree. In 4392109998Smarkm # this example, the environment variable OPENSSL_SOURCE 4393109998Smarkm # is assumed to contain the absolute OpenSSL source directory. 4394109998Smarkm mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 4395109998Smarkm cd objtree/"`uname -s`-`uname -r`-`uname -m`" 4396111147Snectar (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 4397109998Smarkm mkdir -p `dirname $F` 4398109998Smarkm ln -s $OPENSSL_SOURCE/$F $F 4399109998Smarkm done 4400109998Smarkm 4401109998Smarkm To be absolutely sure not to disturb the source tree, a "make clean" 4402109998Smarkm is a good thing. If it isn't successfull, don't worry about it, 4403109998Smarkm it probably means the source directory is very clean. 4404109998Smarkm [Richard Levitte] 4405109998Smarkm 4406109998Smarkm *) Make sure any ENGINE control commands make local copies of string 4407109998Smarkm pointers passed to them whenever necessary. Otherwise it is possible 4408109998Smarkm the caller may have overwritten (or deallocated) the original string 4409109998Smarkm data when a later ENGINE operation tries to use the stored values. 4410109998Smarkm [G�tz Babin-Ebell <babinebell@trustcenter.de>] 4411109998Smarkm 4412109998Smarkm *) Improve diagnostics in file reading and command-line digests. 4413109998Smarkm [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 4414109998Smarkm 4415109998Smarkm *) Add AES modes CFB and OFB to the object database. Correct an 4416109998Smarkm error in AES-CFB decryption. 4417109998Smarkm [Richard Levitte] 4418109998Smarkm 4419109998Smarkm *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 4420109998Smarkm allows existing EVP_CIPHER_CTX structures to be reused after 4421109998Smarkm calling EVP_*Final(). This behaviour is used by encryption 4422109998Smarkm BIOs and some applications. This has the side effect that 4423109998Smarkm applications must explicitly clean up cipher contexts with 4424109998Smarkm EVP_CIPHER_CTX_cleanup() or they will leak memory. 4425109998Smarkm [Steve Henson] 4426109998Smarkm 4427109998Smarkm *) Check the values of dna and dnb in bn_mul_recursive before calling 4428109998Smarkm bn_mul_comba (a non zero value means the a or b arrays do not contain 4429109998Smarkm n2 elements) and fallback to bn_mul_normal if either is not zero. 4430109998Smarkm [Steve Henson] 4431109998Smarkm 4432109998Smarkm *) Fix escaping of non-ASCII characters when using the -subj option 4433109998Smarkm of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 4434109998Smarkm [Lutz Jaenicke] 4435109998Smarkm 4436109998Smarkm *) Make object definitions compliant to LDAP (RFC2256): SN is the short 4437109998Smarkm form for "surname", serialNumber has no short form. 4438109998Smarkm Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 4439109998Smarkm therefore remove "mail" short name for "internet 7". 4440109998Smarkm The OID for unique identifiers in X509 certificates is 4441109998Smarkm x500UniqueIdentifier, not uniqueIdentifier. 4442109998Smarkm Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 4443109998Smarkm [Lutz Jaenicke] 4444109998Smarkm 4445109998Smarkm *) Add an "init" command to the ENGINE config module and auto initialize 4446109998Smarkm ENGINEs. Without any "init" command the ENGINE will be initialized 4447109998Smarkm after all ctrl commands have been executed on it. If init=1 the 4448109998Smarkm ENGINE is initailized at that point (ctrls before that point are run 4449109998Smarkm on the uninitialized ENGINE and after on the initialized one). If 4450109998Smarkm init=0 then the ENGINE will not be iniatialized at all. 4451109998Smarkm [Steve Henson] 4452109998Smarkm 4453109998Smarkm *) Fix the 'app_verify_callback' interface so that the user-defined 4454109998Smarkm argument is actually passed to the callback: In the 4455109998Smarkm SSL_CTX_set_cert_verify_callback() prototype, the callback 4456109998Smarkm declaration has been changed from 4457109998Smarkm int (*cb)() 4458109998Smarkm into 4459109998Smarkm int (*cb)(X509_STORE_CTX *,void *); 4460109998Smarkm in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 4461109998Smarkm i=s->ctx->app_verify_callback(&ctx) 4462109998Smarkm has been changed into 4463109998Smarkm i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 4464109998Smarkm 4465109998Smarkm To update applications using SSL_CTX_set_cert_verify_callback(), 4466109998Smarkm a dummy argument can be added to their callback functions. 4467109998Smarkm [D. K. Smetters <smetters@parc.xerox.com>] 4468109998Smarkm 4469109998Smarkm *) Added the '4758cca' ENGINE to support IBM 4758 cards. 4470109998Smarkm [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 4471109998Smarkm 4472109998Smarkm *) Add and OPENSSL_LOAD_CONF define which will cause 4473109998Smarkm OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 4474109998Smarkm This allows older applications to transparently support certain 4475109998Smarkm OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 4476109998Smarkm Two new functions OPENSSL_add_all_algorithms_noconf() which will never 4477109998Smarkm load the config file and OPENSSL_add_all_algorithms_conf() which will 4478109998Smarkm always load it have also been added. 4479109998Smarkm [Steve Henson] 4480109998Smarkm 4481109998Smarkm *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 4482109998Smarkm Adjust NIDs and EVP layer. 4483109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4484109998Smarkm 4485109998Smarkm *) Config modules support in openssl utility. 4486109998Smarkm 4487109998Smarkm Most commands now load modules from the config file, 4488109998Smarkm though in a few (such as version) this isn't done 4489109998Smarkm because it couldn't be used for anything. 4490109998Smarkm 4491109998Smarkm In the case of ca and req the config file used is 4492109998Smarkm the same as the utility itself: that is the -config 4493109998Smarkm command line option can be used to specify an 4494109998Smarkm alternative file. 4495109998Smarkm [Steve Henson] 4496109998Smarkm 4497109998Smarkm *) Move default behaviour from OPENSSL_config(). If appname is NULL 4498109998Smarkm use "openssl_conf" if filename is NULL use default openssl config file. 4499109998Smarkm [Steve Henson] 4500109998Smarkm 4501109998Smarkm *) Add an argument to OPENSSL_config() to allow the use of an alternative 4502109998Smarkm config section name. Add a new flag to tolerate a missing config file 4503109998Smarkm and move code to CONF_modules_load_file(). 4504109998Smarkm [Steve Henson] 4505109998Smarkm 4506109998Smarkm *) Support for crypto accelerator cards from Accelerated Encryption 4507109998Smarkm Processing, www.aep.ie. (Use engine 'aep') 4508109998Smarkm The support was copied from 0.9.6c [engine] and adapted/corrected 4509109998Smarkm to work with the new engine framework. 4510109998Smarkm [AEP Inc. and Richard Levitte] 4511109998Smarkm 4512109998Smarkm *) Support for SureWare crypto accelerator cards from Baltimore 4513109998Smarkm Technologies. (Use engine 'sureware') 4514109998Smarkm The support was copied from 0.9.6c [engine] and adapted 4515109998Smarkm to work with the new engine framework. 4516109998Smarkm [Richard Levitte] 4517109998Smarkm 4518109998Smarkm *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 4519109998Smarkm make the newer ENGINE framework commands for the CHIL engine work. 4520109998Smarkm [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 4521109998Smarkm 4522109998Smarkm *) Make it possible to produce shared libraries on ReliantUNIX. 4523109998Smarkm [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 4524109998Smarkm 4525109998Smarkm *) Add the configuration target debug-linux-ppro. 4526109998Smarkm Make 'openssl rsa' use the general key loading routines 4527109998Smarkm implemented in apps.c, and make those routines able to 4528109998Smarkm handle the key format FORMAT_NETSCAPE and the variant 4529109998Smarkm FORMAT_IISSGC. 4530109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4531109998Smarkm 4532109998Smarkm *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 4533109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4534109998Smarkm 4535109998Smarkm *) Add -keyform to rsautl, and document -engine. 4536109998Smarkm [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 4537109998Smarkm 4538109998Smarkm *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 4539109998Smarkm BIO_R_NO_SUCH_FILE error code rather than the generic 4540109998Smarkm ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 4541109998Smarkm [Ben Laurie] 4542109998Smarkm 4543109998Smarkm *) Add new functions 4544109998Smarkm ERR_peek_last_error 4545109998Smarkm ERR_peek_last_error_line 4546109998Smarkm ERR_peek_last_error_line_data. 4547109998Smarkm These are similar to 4548109998Smarkm ERR_peek_error 4549109998Smarkm ERR_peek_error_line 4550109998Smarkm ERR_peek_error_line_data, 4551109998Smarkm but report on the latest error recorded rather than the first one 4552109998Smarkm still in the error queue. 4553109998Smarkm [Ben Laurie, Bodo Moeller] 4554109998Smarkm 4555109998Smarkm *) default_algorithms option in ENGINE config module. This allows things 4556109998Smarkm like: 4557109998Smarkm default_algorithms = ALL 4558109998Smarkm default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 4559109998Smarkm [Steve Henson] 4560109998Smarkm 4561109998Smarkm *) Prelminary ENGINE config module. 4562109998Smarkm [Steve Henson] 4563109998Smarkm 4564109998Smarkm *) New experimental application configuration code. 4565109998Smarkm [Steve Henson] 4566109998Smarkm 4567109998Smarkm *) Change the AES code to follow the same name structure as all other 4568109998Smarkm symmetric ciphers, and behave the same way. Move everything to 4569109998Smarkm the directory crypto/aes, thereby obsoleting crypto/rijndael. 4570109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4571109998Smarkm 4572109998Smarkm *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 4573109998Smarkm [Ben Laurie and Theo de Raadt] 4574109998Smarkm 4575109998Smarkm *) Add option to output public keys in req command. 4576109998Smarkm [Massimiliano Pala madwolf@openca.org] 4577109998Smarkm 4578109998Smarkm *) Use wNAFs in EC_POINTs_mul() for improved efficiency 4579109998Smarkm (up to about 10% better than before for P-192 and P-224). 4580109998Smarkm [Bodo Moeller] 4581109998Smarkm 4582109998Smarkm *) New functions/macros 4583109998Smarkm 4584109998Smarkm SSL_CTX_set_msg_callback(ctx, cb) 4585109998Smarkm SSL_CTX_set_msg_callback_arg(ctx, arg) 4586109998Smarkm SSL_set_msg_callback(ssl, cb) 4587109998Smarkm SSL_set_msg_callback_arg(ssl, arg) 4588109998Smarkm 4589109998Smarkm to request calling a callback function 4590109998Smarkm 4591109998Smarkm void cb(int write_p, int version, int content_type, 4592109998Smarkm const void *buf, size_t len, SSL *ssl, void *arg) 4593109998Smarkm 4594109998Smarkm whenever a protocol message has been completely received 4595109998Smarkm (write_p == 0) or sent (write_p == 1). Here 'version' is the 4596109998Smarkm protocol version according to which the SSL library interprets 4597109998Smarkm the current protocol message (SSL2_VERSION, SSL3_VERSION, or 4598109998Smarkm TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 4599109998Smarkm the content type as defined in the SSL 3.0/TLS 1.0 protocol 4600109998Smarkm specification (change_cipher_spec(20), alert(21), handshake(22)). 4601109998Smarkm 'buf' and 'len' point to the actual message, 'ssl' to the 4602109998Smarkm SSL object, and 'arg' is the application-defined value set by 4603109998Smarkm SSL[_CTX]_set_msg_callback_arg(). 4604109998Smarkm 4605109998Smarkm 'openssl s_client' and 'openssl s_server' have new '-msg' options 4606109998Smarkm to enable a callback that displays all protocol messages. 4607109998Smarkm [Bodo Moeller] 4608109998Smarkm 4609109998Smarkm *) Change the shared library support so shared libraries are built as 4610109998Smarkm soon as the corresponding static library is finished, and thereby get 4611109998Smarkm openssl and the test programs linked against the shared library. 4612109998Smarkm This still only happens when the keyword "shard" has been given to 4613109998Smarkm the configuration scripts. 4614109998Smarkm 4615109998Smarkm NOTE: shared library support is still an experimental thing, and 4616109998Smarkm backward binary compatibility is still not guaranteed. 4617109998Smarkm ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 4618109998Smarkm 4619109998Smarkm *) Add support for Subject Information Access extension. 4620109998Smarkm [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 4621109998Smarkm 4622109998Smarkm *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 4623109998Smarkm additional bytes when new memory had to be allocated, not just 4624109998Smarkm when reusing an existing buffer. 4625109998Smarkm [Bodo Moeller] 4626109998Smarkm 4627109998Smarkm *) New command line and configuration option 'utf8' for the req command. 4628109998Smarkm This allows field values to be specified as UTF8 strings. 4629109998Smarkm [Steve Henson] 4630109998Smarkm 4631109998Smarkm *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 4632109998Smarkm runs for the former and machine-readable output for the latter. 4633109998Smarkm [Ben Laurie] 4634109998Smarkm 4635109998Smarkm *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 4636109998Smarkm of the e-mail address in the DN (i.e., it will go into a certificate 4637109998Smarkm extension only). The new configuration file option 'email_in_dn = no' 4638109998Smarkm has the same effect. 4639109998Smarkm [Massimiliano Pala madwolf@openca.org] 4640109998Smarkm 4641109998Smarkm *) Change all functions with names starting with des_ to be starting 4642109998Smarkm with DES_ instead. Add wrappers that are compatible with libdes, 4643109998Smarkm but are named _ossl_old_des_*. Finally, add macros that map the 4644109998Smarkm des_* symbols to the corresponding _ossl_old_des_* if libdes 4645109998Smarkm compatibility is desired. If OpenSSL 0.9.6c compatibility is 4646109998Smarkm desired, the des_* symbols will be mapped to DES_*, with one 4647109998Smarkm exception. 4648109998Smarkm 4649109998Smarkm Since we provide two compatibility mappings, the user needs to 4650109998Smarkm define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 4651109998Smarkm compatibility is desired. The default (i.e., when that macro 4652109998Smarkm isn't defined) is OpenSSL 0.9.6c compatibility. 4653109998Smarkm 4654109998Smarkm There are also macros that enable and disable the support of old 4655109998Smarkm des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 4656109998Smarkm and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 4657109998Smarkm are defined, the default will apply: to support the old des routines. 4658109998Smarkm 4659109998Smarkm In either case, one must include openssl/des.h to get the correct 4660109998Smarkm definitions. Do not try to just include openssl/des_old.h, that 4661109998Smarkm won't work. 4662109998Smarkm 4663109998Smarkm NOTE: This is a major break of an old API into a new one. Software 4664109998Smarkm authors are encouraged to switch to the DES_ style functions. Some 4665109998Smarkm time in the future, des_old.h and the libdes compatibility functions 4666109998Smarkm will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 4667109998Smarkm default), and then completely removed. 4668109998Smarkm [Richard Levitte] 4669109998Smarkm 4670109998Smarkm *) Test for certificates which contain unsupported critical extensions. 4671109998Smarkm If such a certificate is found during a verify operation it is 4672109998Smarkm rejected by default: this behaviour can be overridden by either 4673109998Smarkm handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 4674109998Smarkm by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 4675109998Smarkm X509_supported_extension() has also been added which returns 1 if a 4676109998Smarkm particular extension is supported. 4677109998Smarkm [Steve Henson] 4678109998Smarkm 4679109998Smarkm *) Modify the behaviour of EVP cipher functions in similar way to digests 4680109998Smarkm to retain compatibility with existing code. 4681109998Smarkm [Steve Henson] 4682109998Smarkm 4683109998Smarkm *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 4684109998Smarkm compatibility with existing code. In particular the 'ctx' parameter does 4685109998Smarkm not have to be to be initialized before the call to EVP_DigestInit() and 4686109998Smarkm it is tidied up after a call to EVP_DigestFinal(). New function 4687109998Smarkm EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 4688109998Smarkm EVP_MD_CTX_copy() changed to not require the destination to be 4689109998Smarkm initialized valid and new function EVP_MD_CTX_copy_ex() added which 4690109998Smarkm requires the destination to be valid. 4691109998Smarkm 4692109998Smarkm Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 4693109998Smarkm EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 4694109998Smarkm [Steve Henson] 4695109998Smarkm 4696109998Smarkm *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 4697109998Smarkm so that complete 'Handshake' protocol structures are kept in memory 4698109998Smarkm instead of overwriting 'msg_type' and 'length' with 'body' data. 4699109998Smarkm [Bodo Moeller] 4700109998Smarkm 4701109998Smarkm *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 4702109998Smarkm [Massimo Santin via Richard Levitte] 4703109998Smarkm 4704109998Smarkm *) Major restructuring to the underlying ENGINE code. This includes 4705109998Smarkm reduction of linker bloat, separation of pure "ENGINE" manipulation 4706109998Smarkm (initialisation, etc) from functionality dealing with implementations 4707109998Smarkm of specific crypto iterfaces. This change also introduces integrated 4708109998Smarkm support for symmetric ciphers and digest implementations - so ENGINEs 4709109998Smarkm can now accelerate these by providing EVP_CIPHER and EVP_MD 4710109998Smarkm implementations of their own. This is detailed in crypto/engine/README 4711109998Smarkm as it couldn't be adequately described here. However, there are a few 4712109998Smarkm API changes worth noting - some RSA, DSA, DH, and RAND functions that 4713109998Smarkm were changed in the original introduction of ENGINE code have now 4714109998Smarkm reverted back - the hooking from this code to ENGINE is now a good 4715109998Smarkm deal more passive and at run-time, operations deal directly with 4716109998Smarkm RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 4717109998Smarkm dereferencing through an ENGINE pointer any more. Also, the ENGINE 4718109998Smarkm functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 4719109998Smarkm they were not being used by the framework as there is no concept of a 4720109998Smarkm BIGNUM_METHOD and they could not be generalised to the new 4721109998Smarkm 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 4722109998Smarkm ENGINE_cpy() has been removed as it cannot be consistently defined in 4723109998Smarkm the new code. 4724109998Smarkm [Geoff Thorpe] 4725109998Smarkm 4726109998Smarkm *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 4727109998Smarkm [Steve Henson] 4728109998Smarkm 4729109998Smarkm *) Change mkdef.pl to sort symbols that get the same entry number, 4730109998Smarkm and make sure the automatically generated functions ERR_load_* 4731109998Smarkm become part of libeay.num as well. 4732109998Smarkm [Richard Levitte] 4733109998Smarkm 4734109998Smarkm *) New function SSL_renegotiate_pending(). This returns true once 4735109998Smarkm renegotiation has been requested (either SSL_renegotiate() call 4736109998Smarkm or HelloRequest/ClientHello receveived from the peer) and becomes 4737109998Smarkm false once a handshake has been completed. 4738109998Smarkm (For servers, SSL_renegotiate() followed by SSL_do_handshake() 4739109998Smarkm sends a HelloRequest, but does not ensure that a handshake takes 4740109998Smarkm place. SSL_renegotiate_pending() is useful for checking if the 4741109998Smarkm client has followed the request.) 4742109998Smarkm [Bodo Moeller] 4743109998Smarkm 4744109998Smarkm *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 4745109998Smarkm By default, clients may request session resumption even during 4746109998Smarkm renegotiation (if session ID contexts permit); with this option, 4747109998Smarkm session resumption is possible only in the first handshake. 4748109998Smarkm 4749109998Smarkm SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 4750109998Smarkm more bits available for options that should not be part of 4751109998Smarkm SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 4752109998Smarkm [Bodo Moeller] 4753109998Smarkm 4754109998Smarkm *) Add some demos for certificate and certificate request creation. 4755109998Smarkm [Steve Henson] 4756109998Smarkm 4757109998Smarkm *) Make maximum certificate chain size accepted from the peer application 4758109998Smarkm settable (SSL*_get/set_max_cert_list()), as proposed by 4759109998Smarkm "Douglas E. Engert" <deengert@anl.gov>. 4760109998Smarkm [Lutz Jaenicke] 4761109998Smarkm 4762109998Smarkm *) Add support for shared libraries for Unixware-7 4763109998Smarkm (Boyd Lynn Gerber <gerberb@zenez.com>). 4764109998Smarkm [Lutz Jaenicke] 4765109998Smarkm 4766109998Smarkm *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 4767109998Smarkm be done prior to destruction. Use this to unload error strings from 4768109998Smarkm ENGINEs that load their own error strings. NB: This adds two new API 4769109998Smarkm functions to "get" and "set" this destroy handler in an ENGINE. 4770109998Smarkm [Geoff Thorpe] 4771109998Smarkm 4772109998Smarkm *) Alter all existing ENGINE implementations (except "openssl" and 4773109998Smarkm "openbsd") to dynamically instantiate their own error strings. This 4774109998Smarkm makes them more flexible to be built both as statically-linked ENGINEs 4775109998Smarkm and self-contained shared-libraries loadable via the "dynamic" ENGINE. 4776109998Smarkm Also, add stub code to each that makes building them as self-contained 4777109998Smarkm shared-libraries easier (see README.ENGINE). 4778109998Smarkm [Geoff Thorpe] 4779109998Smarkm 4780109998Smarkm *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 4781109998Smarkm implementations into applications that are completely implemented in 4782109998Smarkm self-contained shared-libraries. The "dynamic" ENGINE exposes control 4783109998Smarkm commands that can be used to configure what shared-library to load and 4784109998Smarkm to control aspects of the way it is handled. Also, made an update to 4785109998Smarkm the README.ENGINE file that brings its information up-to-date and 4786109998Smarkm provides some information and instructions on the "dynamic" ENGINE 4787109998Smarkm (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 4788109998Smarkm [Geoff Thorpe] 4789109998Smarkm 4790109998Smarkm *) Make it possible to unload ranges of ERR strings with a new 4791109998Smarkm "ERR_unload_strings" function. 4792109998Smarkm [Geoff Thorpe] 4793109998Smarkm 4794109998Smarkm *) Add a copy() function to EVP_MD. 4795109998Smarkm [Ben Laurie] 4796109998Smarkm 4797109998Smarkm *) Make EVP_MD routines take a context pointer instead of just the 4798109998Smarkm md_data void pointer. 4799109998Smarkm [Ben Laurie] 4800109998Smarkm 4801109998Smarkm *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 4802109998Smarkm that the digest can only process a single chunk of data 4803109998Smarkm (typically because it is provided by a piece of 4804109998Smarkm hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 4805109998Smarkm is only going to provide a single chunk of data, and hence the 4806109998Smarkm framework needn't accumulate the data for oneshot drivers. 4807109998Smarkm [Ben Laurie] 4808109998Smarkm 4809109998Smarkm *) As with "ERR", make it possible to replace the underlying "ex_data" 4810109998Smarkm functions. This change also alters the storage and management of global 4811109998Smarkm ex_data state - it's now all inside ex_data.c and all "class" code (eg. 4812109998Smarkm RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 4813109998Smarkm index counters. The API functions that use this state have been changed 4814109998Smarkm to take a "class_index" rather than pointers to the class's local STACK 4815109998Smarkm and counter, and there is now an API function to dynamically create new 4816109998Smarkm classes. This centralisation allows us to (a) plug a lot of the 4817109998Smarkm thread-safety problems that existed, and (b) makes it possible to clean 4818109998Smarkm up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 4819109998Smarkm such data would previously have always leaked in application code and 4820109998Smarkm workarounds were in place to make the memory debugging turn a blind eye 4821109998Smarkm to it. Application code that doesn't use this new function will still 4822109998Smarkm leak as before, but their memory debugging output will announce it now 4823109998Smarkm rather than letting it slide. 4824109998Smarkm 4825109998Smarkm Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 4826109998Smarkm induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 4827109998Smarkm has a return value to indicate success or failure. 4828109998Smarkm [Geoff Thorpe] 4829109998Smarkm 4830109998Smarkm *) Make it possible to replace the underlying "ERR" functions such that the 4831109998Smarkm global state (2 LHASH tables and 2 locks) is only used by the "default" 4832109998Smarkm implementation. This change also adds two functions to "get" and "set" 4833109998Smarkm the implementation prior to it being automatically set the first time 4834109998Smarkm any other ERR function takes place. Ie. an application can call "get", 4835109998Smarkm pass the return value to a module it has just loaded, and that module 4836109998Smarkm can call its own "set" function using that value. This means the 4837109998Smarkm module's "ERR" operations will use (and modify) the error state in the 4838109998Smarkm application and not in its own statically linked copy of OpenSSL code. 4839109998Smarkm [Geoff Thorpe] 4840109998Smarkm 4841109998Smarkm *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 4842109998Smarkm reference counts. This performs normal REF_PRINT/REF_CHECK macros on 4843109998Smarkm the operation, and provides a more encapsulated way for external code 4844109998Smarkm (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 4845109998Smarkm to use these functions rather than manually incrementing the counts. 4846109998Smarkm 4847109998Smarkm Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 4848109998Smarkm [Geoff Thorpe] 4849109998Smarkm 4850109998Smarkm *) Add EVP test program. 4851109998Smarkm [Ben Laurie] 4852109998Smarkm 4853109998Smarkm *) Add symmetric cipher support to ENGINE. Expect the API to change! 4854109998Smarkm [Ben Laurie] 4855109998Smarkm 4856109998Smarkm *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 4857109998Smarkm X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 4858109998Smarkm X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 4859109998Smarkm These allow a CRL to be built without having to access X509_CRL fields 4860109998Smarkm directly. Modify 'ca' application to use new functions. 4861109998Smarkm [Steve Henson] 4862109998Smarkm 4863109998Smarkm *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 4864109998Smarkm bug workarounds. Rollback attack detection is a security feature. 4865109998Smarkm The problem will only arise on OpenSSL servers when TLSv1 is not 4866109998Smarkm available (sslv3_server_method() or SSL_OP_NO_TLSv1). 4867109998Smarkm Software authors not wanting to support TLSv1 will have special reasons 4868109998Smarkm for their choice and can explicitly enable this option. 4869109998Smarkm [Bodo Moeller, Lutz Jaenicke] 4870109998Smarkm 4871109998Smarkm *) Rationalise EVP so it can be extended: don't include a union of 4872109998Smarkm cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 4873109998Smarkm (similar to those existing for EVP_CIPHER_CTX). 4874109998Smarkm Usage example: 4875109998Smarkm 4876109998Smarkm EVP_MD_CTX md; 4877109998Smarkm 4878109998Smarkm EVP_MD_CTX_init(&md); /* new function call */ 4879109998Smarkm EVP_DigestInit(&md, EVP_sha1()); 4880109998Smarkm EVP_DigestUpdate(&md, in, len); 4881109998Smarkm EVP_DigestFinal(&md, out, NULL); 4882109998Smarkm EVP_MD_CTX_cleanup(&md); /* new function call */ 4883109998Smarkm 4884109998Smarkm [Ben Laurie] 4885109998Smarkm 4886109998Smarkm *) Make DES key schedule conform to the usual scheme, as well as 4887109998Smarkm correcting its structure. This means that calls to DES functions 4888109998Smarkm now have to pass a pointer to a des_key_schedule instead of a 4889109998Smarkm plain des_key_schedule (which was actually always a pointer 4890109998Smarkm anyway): E.g., 4891109998Smarkm 4892109998Smarkm des_key_schedule ks; 4893109998Smarkm 4894109998Smarkm des_set_key_checked(..., &ks); 4895109998Smarkm des_ncbc_encrypt(..., &ks, ...); 4896109998Smarkm 4897109998Smarkm (Note that a later change renames 'des_...' into 'DES_...'.) 4898109998Smarkm [Ben Laurie] 4899109998Smarkm 4900109998Smarkm *) Initial reduction of linker bloat: the use of some functions, such as 4901109998Smarkm PEM causes large amounts of unused functions to be linked in due to 4902109998Smarkm poor organisation. For example pem_all.c contains every PEM function 4903109998Smarkm which has a knock on effect of linking in large amounts of (unused) 4904109998Smarkm ASN1 code. Grouping together similar functions and splitting unrelated 4905109998Smarkm functions prevents this. 4906109998Smarkm [Steve Henson] 4907109998Smarkm 4908109998Smarkm *) Cleanup of EVP macros. 4909109998Smarkm [Ben Laurie] 4910109998Smarkm 4911109998Smarkm *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 4912109998Smarkm correct _ecb suffix. 4913109998Smarkm [Ben Laurie] 4914109998Smarkm 4915109998Smarkm *) Add initial OCSP responder support to ocsp application. The 4916109998Smarkm revocation information is handled using the text based index 4917109998Smarkm use by the ca application. The responder can either handle 4918109998Smarkm requests generated internally, supplied in files (for example 4919109998Smarkm via a CGI script) or using an internal minimal server. 4920109998Smarkm [Steve Henson] 4921109998Smarkm 4922109998Smarkm *) Add configuration choices to get zlib compression for TLS. 4923109998Smarkm [Richard Levitte] 4924109998Smarkm 4925109998Smarkm *) Changes to Kerberos SSL for RFC 2712 compliance: 4926109998Smarkm 1. Implemented real KerberosWrapper, instead of just using 4927109998Smarkm KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 4928109998Smarkm 2. Implemented optional authenticator field of KerberosWrapper. 4929109998Smarkm 4930109998Smarkm Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 4931109998Smarkm and authenticator structs; see crypto/krb5/. 4932109998Smarkm 4933109998Smarkm Generalized Kerberos calls to support multiple Kerberos libraries. 4934109998Smarkm [Vern Staats <staatsvr@asc.hpc.mil>, 4935109998Smarkm Jeffrey Altman <jaltman@columbia.edu> 4936109998Smarkm via Richard Levitte] 4937109998Smarkm 4938109998Smarkm *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 4939109998Smarkm already does with RSA. testdsa.h now has 'priv_key/pub_key' 4940109998Smarkm values for each of the key sizes rather than having just 4941109998Smarkm parameters (and 'speed' generating keys each time). 4942109998Smarkm [Geoff Thorpe] 4943109998Smarkm 4944109998Smarkm *) Speed up EVP routines. 4945109998Smarkm Before: 4946109998Smarkmencrypt 4947109998Smarkmtype 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 4948109998Smarkmdes-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 4949109998Smarkmdes-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 4950109998Smarkmdes-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 4951109998Smarkmdecrypt 4952109998Smarkmdes-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 4953109998Smarkmdes-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 4954109998Smarkmdes-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 4955109998Smarkm After: 4956109998Smarkmencrypt 4957109998Smarkmdes-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 4958109998Smarkmdecrypt 4959109998Smarkmdes-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 4960109998Smarkm [Ben Laurie] 4961109998Smarkm 4962109998Smarkm *) Added the OS2-EMX target. 4963109998Smarkm ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 4964109998Smarkm 4965109998Smarkm *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 4966109998Smarkm to support NCONF routines in extension code. New function CONF_set_nconf() 4967109998Smarkm to allow functions which take an NCONF to also handle the old LHASH 4968109998Smarkm structure: this means that the old CONF compatible routines can be 4969109998Smarkm retained (in particular wrt extensions) without having to duplicate the 4970109998Smarkm code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 4971109998Smarkm [Steve Henson] 4972109998Smarkm 4973109998Smarkm *) Enhance the general user interface with mechanisms for inner control 4974109998Smarkm and with possibilities to have yes/no kind of prompts. 4975109998Smarkm [Richard Levitte] 4976109998Smarkm 4977109998Smarkm *) Change all calls to low level digest routines in the library and 4978109998Smarkm applications to use EVP. Add missing calls to HMAC_cleanup() and 4979109998Smarkm don't assume HMAC_CTX can be copied using memcpy(). 4980109998Smarkm [Verdon Walker <VWalker@novell.com>, Steve Henson] 4981109998Smarkm 4982109998Smarkm *) Add the possibility to control engines through control names but with 4983109998Smarkm arbitrary arguments instead of just a string. 4984109998Smarkm Change the key loaders to take a UI_METHOD instead of a callback 4985109998Smarkm function pointer. NOTE: this breaks binary compatibility with earlier 4986109998Smarkm versions of OpenSSL [engine]. 4987109998Smarkm Adapt the nCipher code for these new conditions and add a card insertion 4988109998Smarkm callback. 4989109998Smarkm [Richard Levitte] 4990109998Smarkm 4991109998Smarkm *) Enhance the general user interface with mechanisms to better support 4992109998Smarkm dialog box interfaces, application-defined prompts, the possibility 4993109998Smarkm to use defaults (for example default passwords from somewhere else) 4994109998Smarkm and interrupts/cancellations. 4995109998Smarkm [Richard Levitte] 4996109998Smarkm 4997109998Smarkm *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 4998109998Smarkm attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 4999109998Smarkm [Steve Henson] 5000109998Smarkm 5001109998Smarkm *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 5002109998Smarkm tidy up some unnecessarily weird code in 'sk_new()'). 5003109998Smarkm [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 5004109998Smarkm 5005109998Smarkm *) Change the key loading routines for ENGINEs to use the same kind 5006109998Smarkm callback (pem_password_cb) as all other routines that need this 5007109998Smarkm kind of callback. 5008109998Smarkm [Richard Levitte] 5009109998Smarkm 5010109998Smarkm *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 5011109998Smarkm 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 5012109998Smarkm than this minimum value is recommended. 5013109998Smarkm [Lutz Jaenicke] 5014109998Smarkm 5015109998Smarkm *) New random seeder for OpenVMS, using the system process statistics 5016109998Smarkm that are easily reachable. 5017109998Smarkm [Richard Levitte] 5018109998Smarkm 5019109998Smarkm *) Windows apparently can't transparently handle global 5020109998Smarkm variables defined in DLLs. Initialisations such as: 5021109998Smarkm 5022109998Smarkm const ASN1_ITEM *it = &ASN1_INTEGER_it; 5023109998Smarkm 5024109998Smarkm wont compile. This is used by the any applications that need to 5025109998Smarkm declare their own ASN1 modules. This was fixed by adding the option 5026109998Smarkm EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 5027109998Smarkm needed for static libraries under Win32. 5028109998Smarkm [Steve Henson] 5029109998Smarkm 5030109998Smarkm *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 5031109998Smarkm setting of purpose and trust fields. New X509_STORE trust and 5032109998Smarkm purpose functions and tidy up setting in other SSL functions. 5033109998Smarkm [Steve Henson] 5034109998Smarkm 5035109998Smarkm *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 5036109998Smarkm structure. These are inherited by X509_STORE_CTX when it is 5037109998Smarkm initialised. This allows various defaults to be set in the 5038109998Smarkm X509_STORE structure (such as flags for CRL checking and custom 5039109998Smarkm purpose or trust settings) for functions which only use X509_STORE_CTX 5040109998Smarkm internally such as S/MIME. 5041109998Smarkm 5042109998Smarkm Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 5043109998Smarkm trust settings if they are not set in X509_STORE. This allows X509_STORE 5044109998Smarkm purposes and trust (in S/MIME for example) to override any set by default. 5045109998Smarkm 5046109998Smarkm Add command line options for CRL checking to smime, s_client and s_server 5047109998Smarkm applications. 5048109998Smarkm [Steve Henson] 5049109998Smarkm 5050109998Smarkm *) Initial CRL based revocation checking. If the CRL checking flag(s) 5051109998Smarkm are set then the CRL is looked up in the X509_STORE structure and 5052109998Smarkm its validity and signature checked, then if the certificate is found 5053109998Smarkm in the CRL the verify fails with a revoked error. 5054109998Smarkm 5055109998Smarkm Various new CRL related callbacks added to X509_STORE_CTX structure. 5056109998Smarkm 5057109998Smarkm Command line options added to 'verify' application to support this. 5058109998Smarkm 5059109998Smarkm This needs some additional work, such as being able to handle multiple 5060109998Smarkm CRLs with different times, extension based lookup (rather than just 5061109998Smarkm by subject name) and ultimately more complete V2 CRL extension 5062109998Smarkm handling. 5063109998Smarkm [Steve Henson] 5064109998Smarkm 5065109998Smarkm *) Add a general user interface API (crypto/ui/). This is designed 5066109998Smarkm to replace things like des_read_password and friends (backward 5067109998Smarkm compatibility functions using this new API are provided). 5068109998Smarkm The purpose is to remove prompting functions from the DES code 5069109998Smarkm section as well as provide for prompting through dialog boxes in 5070109998Smarkm a window system and the like. 5071109998Smarkm [Richard Levitte] 5072109998Smarkm 5073109998Smarkm *) Add "ex_data" support to ENGINE so implementations can add state at a 5074109998Smarkm per-structure level rather than having to store it globally. 5075109998Smarkm [Geoff] 5076109998Smarkm 5077109998Smarkm *) Make it possible for ENGINE structures to be copied when retrieved by 5078109998Smarkm ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 5079109998Smarkm This causes the "original" ENGINE structure to act like a template, 5080109998Smarkm analogous to the RSA vs. RSA_METHOD type of separation. Because of this 5081109998Smarkm operational state can be localised to each ENGINE structure, despite the 5082109998Smarkm fact they all share the same "methods". New ENGINE structures returned in 5083109998Smarkm this case have no functional references and the return value is the single 5084109998Smarkm structural reference. This matches the single structural reference returned 5085109998Smarkm by ENGINE_by_id() normally, when it is incremented on the pre-existing 5086109998Smarkm ENGINE structure. 5087109998Smarkm [Geoff] 5088109998Smarkm 5089109998Smarkm *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 5090109998Smarkm needs to match any other type at all we need to manually clear the 5091109998Smarkm tag cache. 5092109998Smarkm [Steve Henson] 5093109998Smarkm 5094109998Smarkm *) Changes to the "openssl engine" utility to include; 5095109998Smarkm - verbosity levels ('-v', '-vv', and '-vvv') that provide information 5096109998Smarkm about an ENGINE's available control commands. 5097109998Smarkm - executing control commands from command line arguments using the 5098109998Smarkm '-pre' and '-post' switches. '-post' is only used if '-t' is 5099109998Smarkm specified and the ENGINE is successfully initialised. The syntax for 5100109998Smarkm the individual commands are colon-separated, for example; 5101109998Smarkm openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 5102109998Smarkm [Geoff] 5103109998Smarkm 5104109998Smarkm *) New dynamic control command support for ENGINEs. ENGINEs can now 5105109998Smarkm declare their own commands (numbers), names (strings), descriptions, 5106109998Smarkm and input types for run-time discovery by calling applications. A 5107109998Smarkm subset of these commands are implicitly classed as "executable" 5108109998Smarkm depending on their input type, and only these can be invoked through 5109109998Smarkm the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 5110109998Smarkm can be based on user input, config files, etc). The distinction is 5111109998Smarkm that "executable" commands cannot return anything other than a boolean 5112109998Smarkm result and can only support numeric or string input, whereas some 5113109998Smarkm discoverable commands may only be for direct use through 5114109998Smarkm ENGINE_ctrl(), eg. supporting the exchange of binary data, function 5115109998Smarkm pointers, or other custom uses. The "executable" commands are to 5116109998Smarkm support parameterisations of ENGINE behaviour that can be 5117109998Smarkm unambiguously defined by ENGINEs and used consistently across any 5118109998Smarkm OpenSSL-based application. Commands have been added to all the 5119109998Smarkm existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 5120109998Smarkm control over shared-library paths without source code alterations. 5121109998Smarkm [Geoff] 5122109998Smarkm 5123109998Smarkm *) Changed all ENGINE implementations to dynamically allocate their 5124109998Smarkm ENGINEs rather than declaring them statically. Apart from this being 5125109998Smarkm necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 5126109998Smarkm this also allows the implementations to compile without using the 5127109998Smarkm internal engine_int.h header. 5128109998Smarkm [Geoff] 5129109998Smarkm 5130109998Smarkm *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 5131109998Smarkm 'const' value. Any code that should be able to modify a RAND_METHOD 5132109998Smarkm should already have non-const pointers to it (ie. they should only 5133109998Smarkm modify their own ones). 5134109998Smarkm [Geoff] 5135109998Smarkm 5136109998Smarkm *) Made a variety of little tweaks to the ENGINE code. 5137109998Smarkm - "atalla" and "ubsec" string definitions were moved from header files 5138109998Smarkm to C code. "nuron" string definitions were placed in variables 5139109998Smarkm rather than hard-coded - allowing parameterisation of these values 5140109998Smarkm later on via ctrl() commands. 5141109998Smarkm - Removed unused "#if 0"'d code. 5142109998Smarkm - Fixed engine list iteration code so it uses ENGINE_free() to release 5143109998Smarkm structural references. 5144109998Smarkm - Constified the RAND_METHOD element of ENGINE structures. 5145109998Smarkm - Constified various get/set functions as appropriate and added 5146109998Smarkm missing functions (including a catch-all ENGINE_cpy that duplicates 5147109998Smarkm all ENGINE values onto a new ENGINE except reference counts/state). 5148109998Smarkm - Removed NULL parameter checks in get/set functions. Setting a method 5149109998Smarkm or function to NULL is a way of cancelling out a previously set 5150109998Smarkm value. Passing a NULL ENGINE parameter is just plain stupid anyway 5151109998Smarkm and doesn't justify the extra error symbols and code. 5152109998Smarkm - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 5153109998Smarkm flags from engine_int.h to engine.h. 5154109998Smarkm - Changed prototypes for ENGINE handler functions (init(), finish(), 5155109998Smarkm ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 5156109998Smarkm [Geoff] 5157109998Smarkm 5158109998Smarkm *) Implement binary inversion algorithm for BN_mod_inverse in addition 5159109998Smarkm to the algorithm using long division. The binary algorithm can be 5160109998Smarkm used only if the modulus is odd. On 32-bit systems, it is faster 5161109998Smarkm only for relatively small moduli (roughly 20-30% for 128-bit moduli, 5162109998Smarkm roughly 5-15% for 256-bit moduli), so we use it only for moduli 5163109998Smarkm up to 450 bits. In 64-bit environments, the binary algorithm 5164109998Smarkm appears to be advantageous for much longer moduli; here we use it 5165109998Smarkm for moduli up to 2048 bits. 5166109998Smarkm [Bodo Moeller] 5167109998Smarkm 5168109998Smarkm *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 5169109998Smarkm could not support the combine flag in choice fields. 5170109998Smarkm [Steve Henson] 5171109998Smarkm 5172109998Smarkm *) Add a 'copy_extensions' option to the 'ca' utility. This copies 5173109998Smarkm extensions from a certificate request to the certificate. 5174109998Smarkm [Steve Henson] 5175109998Smarkm 5176109998Smarkm *) Allow multiple 'certopt' and 'nameopt' options to be separated 5177109998Smarkm by commas. Add 'namopt' and 'certopt' options to the 'ca' config 5178109998Smarkm file: this allows the display of the certificate about to be 5179109998Smarkm signed to be customised, to allow certain fields to be included 5180109998Smarkm or excluded and extension details. The old system didn't display 5181109998Smarkm multicharacter strings properly, omitted fields not in the policy 5182109998Smarkm and couldn't display additional details such as extensions. 5183109998Smarkm [Steve Henson] 5184109998Smarkm 5185109998Smarkm *) Function EC_POINTs_mul for multiple scalar multiplication 5186109998Smarkm of an arbitrary number of elliptic curve points 5187109998Smarkm \sum scalars[i]*points[i], 5188109998Smarkm optionally including the generator defined for the EC_GROUP: 5189109998Smarkm scalar*generator + \sum scalars[i]*points[i]. 5190109998Smarkm 5191109998Smarkm EC_POINT_mul is a simple wrapper function for the typical case 5192109998Smarkm that the point list has just one item (besides the optional 5193109998Smarkm generator). 5194109998Smarkm [Bodo Moeller] 5195109998Smarkm 5196109998Smarkm *) First EC_METHODs for curves over GF(p): 5197109998Smarkm 5198109998Smarkm EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 5199109998Smarkm operations and provides various method functions that can also 5200109998Smarkm operate with faster implementations of modular arithmetic. 5201109998Smarkm 5202109998Smarkm EC_GFp_mont_method() reuses most functions that are part of 5203109998Smarkm EC_GFp_simple_method, but uses Montgomery arithmetic. 5204109998Smarkm 5205109998Smarkm [Bodo Moeller; point addition and point doubling 5206109998Smarkm implementation directly derived from source code provided by 5207109998Smarkm Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 5208109998Smarkm 5209109998Smarkm *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 5210109998Smarkm crypto/ec/ec_lib.c): 5211109998Smarkm 5212109998Smarkm Curves are EC_GROUP objects (with an optional group generator) 5213109998Smarkm based on EC_METHODs that are built into the library. 5214109998Smarkm 5215109998Smarkm Points are EC_POINT objects based on EC_GROUP objects. 5216109998Smarkm 5217109998Smarkm Most of the framework would be able to handle curves over arbitrary 5218109998Smarkm finite fields, but as there are no obvious types for fields other 5219109998Smarkm than GF(p), some functions are limited to that for now. 5220109998Smarkm [Bodo Moeller] 5221109998Smarkm 5222109998Smarkm *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 5223109998Smarkm that the file contains a complete HTTP response. 5224109998Smarkm [Richard Levitte] 5225109998Smarkm 5226109998Smarkm *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 5227109998Smarkm change the def and num file printf format specifier from "%-40sXXX" 5228109998Smarkm to "%-39s XXX". The latter will always guarantee a space after the 5229109998Smarkm field while the former will cause them to run together if the field 5230109998Smarkm is 40 of more characters long. 5231109998Smarkm [Steve Henson] 5232109998Smarkm 5233109998Smarkm *) Constify the cipher and digest 'method' functions and structures 5234109998Smarkm and modify related functions to take constant EVP_MD and EVP_CIPHER 5235109998Smarkm pointers. 5236109998Smarkm [Steve Henson] 5237109998Smarkm 5238109998Smarkm *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 5239109998Smarkm in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 5240109998Smarkm [Bodo Moeller] 5241109998Smarkm 5242109998Smarkm *) Modify EVP_Digest*() routines so they now return values. Although the 5243109998Smarkm internal software routines can never fail additional hardware versions 5244109998Smarkm might. 5245109998Smarkm [Steve Henson] 5246109998Smarkm 5247109998Smarkm *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 5248109998Smarkm 5249109998Smarkm Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 5250109998Smarkm (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 5251109998Smarkm 5252109998Smarkm ASN1 error codes 5253109998Smarkm ERR_R_NESTED_ASN1_ERROR 5254109998Smarkm ... 5255109998Smarkm ERR_R_MISSING_ASN1_EOS 5256109998Smarkm were 4 .. 9, conflicting with 5257109998Smarkm ERR_LIB_RSA (= ERR_R_RSA_LIB) 5258109998Smarkm ... 5259109998Smarkm ERR_LIB_PEM (= ERR_R_PEM_LIB). 5260109998Smarkm They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 5261109998Smarkm 5262109998Smarkm Add new error code 'ERR_R_INTERNAL_ERROR'. 5263109998Smarkm [Bodo Moeller] 5264109998Smarkm 5265109998Smarkm *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 5266109998Smarkm suffices. 5267109998Smarkm [Bodo Moeller] 5268109998Smarkm 5269109998Smarkm *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 5270109998Smarkm sets the subject name for a new request or supersedes the 5271109998Smarkm subject name in a given request. Formats that can be parsed are 5272109998Smarkm 'CN=Some Name, OU=myOU, C=IT' 5273109998Smarkm and 5274109998Smarkm 'CN=Some Name/OU=myOU/C=IT'. 5275109998Smarkm 5276109998Smarkm Add options '-batch' and '-verbose' to 'openssl req'. 5277109998Smarkm [Massimiliano Pala <madwolf@hackmasters.net>] 5278109998Smarkm 5279109998Smarkm *) Introduce the possibility to access global variables through 5280109998Smarkm functions on platform were that's the best way to handle exporting 5281109998Smarkm global variables in shared libraries. To enable this functionality, 5282109998Smarkm one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 5283109998Smarkm "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 5284109998Smarkm is normally done by Configure or something similar). 5285109998Smarkm 5286109998Smarkm To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 5287109998Smarkm in the source file (foo.c) like this: 5288109998Smarkm 5289109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 5290109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(double,bar); 5291109998Smarkm 5292109998Smarkm To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 5293109998Smarkm and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 5294109998Smarkm 5295109998Smarkm OPENSSL_DECLARE_GLOBAL(int,foo); 5296109998Smarkm #define foo OPENSSL_GLOBAL_REF(foo) 5297109998Smarkm OPENSSL_DECLARE_GLOBAL(double,bar); 5298109998Smarkm #define bar OPENSSL_GLOBAL_REF(bar) 5299109998Smarkm 5300109998Smarkm The #defines are very important, and therefore so is including the 5301109998Smarkm header file everywhere where the defined globals are used. 5302109998Smarkm 5303109998Smarkm The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 5304109998Smarkm of ASN.1 items, but that structure is a bit different. 5305109998Smarkm 5306109998Smarkm The largest change is in util/mkdef.pl which has been enhanced with 5307109998Smarkm better and easier to understand logic to choose which symbols should 5308109998Smarkm go into the Windows .def files as well as a number of fixes and code 5309109998Smarkm cleanup (among others, algorithm keywords are now sorted 5310109998Smarkm lexicographically to avoid constant rewrites). 5311109998Smarkm [Richard Levitte] 5312109998Smarkm 5313109998Smarkm *) In BN_div() keep a copy of the sign of 'num' before writing the 5314109998Smarkm result to 'rm' because if rm==num the value will be overwritten 5315109998Smarkm and produce the wrong result if 'num' is negative: this caused 5316109998Smarkm problems with BN_mod() and BN_nnmod(). 5317109998Smarkm [Steve Henson] 5318109998Smarkm 5319109998Smarkm *) Function OCSP_request_verify(). This checks the signature on an 5320109998Smarkm OCSP request and verifies the signer certificate. The signer 5321109998Smarkm certificate is just checked for a generic purpose and OCSP request 5322109998Smarkm trust settings. 5323109998Smarkm [Steve Henson] 5324109998Smarkm 5325109998Smarkm *) Add OCSP_check_validity() function to check the validity of OCSP 5326109998Smarkm responses. OCSP responses are prepared in real time and may only 5327109998Smarkm be a few seconds old. Simply checking that the current time lies 5328109998Smarkm between thisUpdate and nextUpdate max reject otherwise valid responses 5329109998Smarkm caused by either OCSP responder or client clock inaccuracy. Instead 5330109998Smarkm we allow thisUpdate and nextUpdate to fall within a certain period of 5331109998Smarkm the current time. The age of the response can also optionally be 5332109998Smarkm checked. Two new options -validity_period and -status_age added to 5333109998Smarkm ocsp utility. 5334109998Smarkm [Steve Henson] 5335109998Smarkm 5336109998Smarkm *) If signature or public key algorithm is unrecognized print out its 5337109998Smarkm OID rather that just UNKNOWN. 5338109998Smarkm [Steve Henson] 5339109998Smarkm 5340109998Smarkm *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 5341109998Smarkm OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 5342109998Smarkm ID to be generated from the issuer certificate alone which can then be 5343109998Smarkm passed to OCSP_id_issuer_cmp(). 5344109998Smarkm [Steve Henson] 5345109998Smarkm 5346109998Smarkm *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 5347109998Smarkm ASN1 modules to export functions returning ASN1_ITEM pointers 5348109998Smarkm instead of the ASN1_ITEM structures themselves. This adds several 5349109998Smarkm new macros which allow the underlying ASN1 function/structure to 5350109998Smarkm be accessed transparently. As a result code should not use ASN1_ITEM 5351109998Smarkm references directly (such as &X509_it) but instead use the relevant 5352109998Smarkm macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 5353109998Smarkm use of the new ASN1 code on platforms where exporting structures 5354109998Smarkm is problematical (for example in shared libraries) but exporting 5355109998Smarkm functions returning pointers to structures is not. 5356109998Smarkm [Steve Henson] 5357109998Smarkm 5358109998Smarkm *) Add support for overriding the generation of SSL/TLS session IDs. 5359109998Smarkm These callbacks can be registered either in an SSL_CTX or per SSL. 5360109998Smarkm The purpose of this is to allow applications to control, if they wish, 5361109998Smarkm the arbitrary values chosen for use as session IDs, particularly as it 5362109998Smarkm can be useful for session caching in multiple-server environments. A 5363109998Smarkm command-line switch for testing this (and any client code that wishes 5364109998Smarkm to use such a feature) has been added to "s_server". 5365109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 5366109998Smarkm 5367109998Smarkm *) Modify mkdef.pl to recognise and parse preprocessor conditionals 5368109998Smarkm of the form '#if defined(...) || defined(...) || ...' and 5369109998Smarkm '#if !defined(...) && !defined(...) && ...'. This also avoids 5370109998Smarkm the growing number of special cases it was previously handling. 5371109998Smarkm [Richard Levitte] 5372109998Smarkm 5373109998Smarkm *) Make all configuration macros available for application by making 5374109998Smarkm sure they are available in opensslconf.h, by giving them names starting 5375109998Smarkm with "OPENSSL_" to avoid conflicts with other packages and by making 5376109998Smarkm sure e_os2.h will cover all platform-specific cases together with 5377109998Smarkm opensslconf.h. 5378109998Smarkm Additionally, it is now possible to define configuration/platform- 5379109998Smarkm specific names (called "system identities"). In the C code, these 5380109998Smarkm are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 5381109998Smarkm macro with the name beginning with "OPENSSL_SYS_", which is determined 5382109998Smarkm from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 5383109998Smarkm what is available. 5384109998Smarkm [Richard Levitte] 5385109998Smarkm 5386109998Smarkm *) New option -set_serial to 'req' and 'x509' this allows the serial 5387109998Smarkm number to use to be specified on the command line. Previously self 5388109998Smarkm signed certificates were hard coded with serial number 0 and the 5389109998Smarkm CA options of 'x509' had to use a serial number in a file which was 5390109998Smarkm auto incremented. 5391109998Smarkm [Steve Henson] 5392109998Smarkm 5393109998Smarkm *) New options to 'ca' utility to support V2 CRL entry extensions. 5394109998Smarkm Currently CRL reason, invalidity date and hold instruction are 5395109998Smarkm supported. Add new CRL extensions to V3 code and some new objects. 5396109998Smarkm [Steve Henson] 5397109998Smarkm 5398109998Smarkm *) New function EVP_CIPHER_CTX_set_padding() this is used to 5399109998Smarkm disable standard block padding (aka PKCS#5 padding) in the EVP 5400109998Smarkm API, which was previously mandatory. This means that the data is 5401109998Smarkm not padded in any way and so the total length much be a multiple 5402109998Smarkm of the block size, otherwise an error occurs. 5403109998Smarkm [Steve Henson] 5404109998Smarkm 5405109998Smarkm *) Initial (incomplete) OCSP SSL support. 5406109998Smarkm [Steve Henson] 5407109998Smarkm 5408109998Smarkm *) New function OCSP_parse_url(). This splits up a URL into its host, 5409109998Smarkm port and path components: primarily to parse OCSP URLs. New -url 5410109998Smarkm option to ocsp utility. 5411109998Smarkm [Steve Henson] 5412109998Smarkm 5413109998Smarkm *) New nonce behavior. The return value of OCSP_check_nonce() now 5414109998Smarkm reflects the various checks performed. Applications can decide 5415109998Smarkm whether to tolerate certain situations such as an absent nonce 5416109998Smarkm in a response when one was present in a request: the ocsp application 5417109998Smarkm just prints out a warning. New function OCSP_add1_basic_nonce() 5418109998Smarkm this is to allow responders to include a nonce in a response even if 5419109998Smarkm the request is nonce-less. 5420109998Smarkm [Steve Henson] 5421109998Smarkm 5422109998Smarkm *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 5423109998Smarkm skipped when using openssl x509 multiple times on a single input file, 5424109998Smarkm e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 5425109998Smarkm [Bodo Moeller] 5426109998Smarkm 5427109998Smarkm *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 5428109998Smarkm set string type: to handle setting ASN1_TIME structures. Fix ca 5429109998Smarkm utility to correctly initialize revocation date of CRLs. 5430109998Smarkm [Steve Henson] 5431109998Smarkm 5432109998Smarkm *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 5433109998Smarkm the clients preferred ciphersuites and rather use its own preferences. 5434109998Smarkm Should help to work around M$ SGC (Server Gated Cryptography) bug in 5435109998Smarkm Internet Explorer by ensuring unchanged hash method during stepup. 5436109998Smarkm (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 5437109998Smarkm [Lutz Jaenicke] 5438109998Smarkm 5439109998Smarkm *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 5440109998Smarkm to aes and add a new 'exist' option to print out symbols that don't 5441109998Smarkm appear to exist. 5442109998Smarkm [Steve Henson] 5443109998Smarkm 5444109998Smarkm *) Additional options to ocsp utility to allow flags to be set and 5445109998Smarkm additional certificates supplied. 5446109998Smarkm [Steve Henson] 5447109998Smarkm 5448109998Smarkm *) Add the option -VAfile to 'openssl ocsp', so the user can give the 5449109998Smarkm OCSP client a number of certificate to only verify the response 5450109998Smarkm signature against. 5451109998Smarkm [Richard Levitte] 5452109998Smarkm 5453109998Smarkm *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 5454109998Smarkm handle the new API. Currently only ECB, CBC modes supported. Add new 5455109998Smarkm AES OIDs. 5456109998Smarkm 5457109998Smarkm Add TLS AES ciphersuites as described in RFC3268, "Advanced 5458109998Smarkm Encryption Standard (AES) Ciphersuites for Transport Layer 5459109998Smarkm Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 5460109998Smarkm not enabled by default and were not part of the "ALL" ciphersuite 5461109998Smarkm alias because they were not yet official; they could be 5462109998Smarkm explicitly requested by specifying the "AESdraft" ciphersuite 5463109998Smarkm group alias. In the final release of OpenSSL 0.9.7, the group 5464109998Smarkm alias is called "AES" and is part of "ALL".) 5465109998Smarkm [Ben Laurie, Steve Henson, Bodo Moeller] 5466109998Smarkm 5467109998Smarkm *) New function OCSP_copy_nonce() to copy nonce value (if present) from 5468109998Smarkm request to response. 5469109998Smarkm [Steve Henson] 5470109998Smarkm 5471109998Smarkm *) Functions for OCSP responders. OCSP_request_onereq_count(), 5472109998Smarkm OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 5473109998Smarkm extract information from a certificate request. OCSP_response_create() 5474109998Smarkm creates a response and optionally adds a basic response structure. 5475109998Smarkm OCSP_basic_add1_status() adds a complete single response to a basic 5476109998Smarkm response and returns the OCSP_SINGLERESP structure just added (to allow 5477109998Smarkm extensions to be included for example). OCSP_basic_add1_cert() adds a 5478109998Smarkm certificate to a basic response and OCSP_basic_sign() signs a basic 5479109998Smarkm response with various flags. New helper functions ASN1_TIME_check() 5480109998Smarkm (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 5481109998Smarkm (converts ASN1_TIME to GeneralizedTime). 5482109998Smarkm [Steve Henson] 5483109998Smarkm 5484109998Smarkm *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 5485109998Smarkm in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 5486109998Smarkm structure from a certificate. X509_pubkey_digest() digests the public_key 5487109998Smarkm contents: this is used in various key identifiers. 5488109998Smarkm [Steve Henson] 5489109998Smarkm 5490109998Smarkm *) Make sk_sort() tolerate a NULL argument. 5491109998Smarkm [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 5492109998Smarkm 5493109998Smarkm *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 5494109998Smarkm passed by the function are trusted implicitly. If any of them signed the 5495109998Smarkm response then it is assumed to be valid and is not verified. 5496109998Smarkm [Steve Henson] 5497109998Smarkm 5498109998Smarkm *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 5499109998Smarkm to data. This was previously part of the PKCS7 ASN1 code. This 5500109998Smarkm was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 5501109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 5502109998Smarkm <support@securenetterm.com>] 5503109998Smarkm 5504109998Smarkm *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 5505109998Smarkm routines: without these tracing memory leaks is very painful. 5506109998Smarkm Fix leaks in PKCS12 and PKCS7 routines. 5507109998Smarkm [Steve Henson] 5508109998Smarkm 5509109998Smarkm *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 5510109998Smarkm Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 5511109998Smarkm effectively meant GeneralizedTime would never be used. Now it 5512109998Smarkm is initialised to -1 but X509_time_adj() now has to check the value 5513109998Smarkm and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 5514109998Smarkm V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 5515109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 5516109998Smarkm <support@securenetterm.com>] 5517109998Smarkm 5518109998Smarkm *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 5519109998Smarkm result in a zero length in the ASN1_INTEGER structure which was 5520109998Smarkm not consistent with the structure when d2i_ASN1_INTEGER() was used 5521109998Smarkm and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 5522109998Smarkm to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 5523109998Smarkm where it did not print out a minus for negative ASN1_INTEGER. 5524109998Smarkm [Steve Henson] 5525109998Smarkm 5526109998Smarkm *) Add summary printout to ocsp utility. The various functions which 5527109998Smarkm convert status values to strings have been renamed to: 5528109998Smarkm OCSP_response_status_str(), OCSP_cert_status_str() and 5529109998Smarkm OCSP_crl_reason_str() and are no longer static. New options 5530109998Smarkm to verify nonce values and to disable verification. OCSP response 5531109998Smarkm printout format cleaned up. 5532109998Smarkm [Steve Henson] 5533109998Smarkm 5534109998Smarkm *) Add additional OCSP certificate checks. These are those specified 5535109998Smarkm in RFC2560. This consists of two separate checks: the CA of the 5536109998Smarkm certificate being checked must either be the OCSP signer certificate 5537109998Smarkm or the issuer of the OCSP signer certificate. In the latter case the 5538109998Smarkm OCSP signer certificate must contain the OCSP signing extended key 5539109998Smarkm usage. This check is performed by attempting to match the OCSP 5540109998Smarkm signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 5541109998Smarkm in the OCSP_CERTID structures of the response. 5542109998Smarkm [Steve Henson] 5543109998Smarkm 5544109998Smarkm *) Initial OCSP certificate verification added to OCSP_basic_verify() 5545109998Smarkm and related routines. This uses the standard OpenSSL certificate 5546109998Smarkm verify routines to perform initial checks (just CA validity) and 5547109998Smarkm to obtain the certificate chain. Then additional checks will be 5548109998Smarkm performed on the chain. Currently the root CA is checked to see 5549109998Smarkm if it is explicitly trusted for OCSP signing. This is used to set 5550109998Smarkm a root CA as a global signing root: that is any certificate that 5551109998Smarkm chains to that CA is an acceptable OCSP signing certificate. 5552109998Smarkm [Steve Henson] 5553109998Smarkm 5554109998Smarkm *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 5555109998Smarkm extensions from a separate configuration file. 5556109998Smarkm As when reading extensions from the main configuration file, 5557109998Smarkm the '-extensions ...' option may be used for specifying the 5558109998Smarkm section to use. 5559109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 5560109998Smarkm 5561109998Smarkm *) New OCSP utility. Allows OCSP requests to be generated or 5562109998Smarkm read. The request can be sent to a responder and the output 5563109998Smarkm parsed, outputed or printed in text form. Not complete yet: 5564109998Smarkm still needs to check the OCSP response validity. 5565109998Smarkm [Steve Henson] 5566109998Smarkm 5567109998Smarkm *) New subcommands for 'openssl ca': 5568109998Smarkm 'openssl ca -status <serial>' prints the status of the cert with 5569109998Smarkm the given serial number (according to the index file). 5570109998Smarkm 'openssl ca -updatedb' updates the expiry status of certificates 5571109998Smarkm in the index file. 5572109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 5573109998Smarkm 5574109998Smarkm *) New '-newreq-nodes' command option to CA.pl. This is like 5575109998Smarkm '-newreq', but calls 'openssl req' with the '-nodes' option 5576109998Smarkm so that the resulting key is not encrypted. 5577109998Smarkm [Damien Miller <djm@mindrot.org>] 5578109998Smarkm 5579109998Smarkm *) New configuration for the GNU Hurd. 5580109998Smarkm [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 5581109998Smarkm 5582109998Smarkm *) Initial code to implement OCSP basic response verify. This 5583109998Smarkm is currently incomplete. Currently just finds the signer's 5584109998Smarkm certificate and verifies the signature on the response. 5585109998Smarkm [Steve Henson] 5586109998Smarkm 5587109998Smarkm *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 5588109998Smarkm value of OPENSSLDIR. This is available via the new '-d' option 5589109998Smarkm to 'openssl version', and is also included in 'openssl version -a'. 5590109998Smarkm [Bodo Moeller] 5591109998Smarkm 5592109998Smarkm *) Allowing defining memory allocation callbacks that will be given 5593109998Smarkm file name and line number information in additional arguments 5594109998Smarkm (a const char* and an int). The basic functionality remains, as 5595109998Smarkm well as the original possibility to just replace malloc(), 5596109998Smarkm realloc() and free() by functions that do not know about these 5597109998Smarkm additional arguments. To register and find out the current 5598109998Smarkm settings for extended allocation functions, the following 5599109998Smarkm functions are provided: 5600109998Smarkm 5601109998Smarkm CRYPTO_set_mem_ex_functions 5602109998Smarkm CRYPTO_set_locked_mem_ex_functions 5603109998Smarkm CRYPTO_get_mem_ex_functions 5604109998Smarkm CRYPTO_get_locked_mem_ex_functions 5605109998Smarkm 5606109998Smarkm These work the same way as CRYPTO_set_mem_functions and friends. 5607109998Smarkm CRYPTO_get_[locked_]mem_functions now writes 0 where such an 5608109998Smarkm extended allocation function is enabled. 5609109998Smarkm Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 5610109998Smarkm a conventional allocation function is enabled. 5611109998Smarkm [Richard Levitte, Bodo Moeller] 5612109998Smarkm 5613109998Smarkm *) Finish off removing the remaining LHASH function pointer casts. 5614109998Smarkm There should no longer be any prototype-casting required when using 5615109998Smarkm the LHASH abstraction, and any casts that remain are "bugs". See 5616109998Smarkm the callback types and macros at the head of lhash.h for details 5617109998Smarkm (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 5618109998Smarkm [Geoff Thorpe] 5619109998Smarkm 5620109998Smarkm *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 5621109998Smarkm If /dev/[u]random devices are not available or do not return enough 5622109998Smarkm entropy, EGD style sockets (served by EGD or PRNGD) will automatically 5623109998Smarkm be queried. 5624109998Smarkm The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 5625109998Smarkm /etc/entropy will be queried once each in this sequence, quering stops 5626109998Smarkm when enough entropy was collected without querying more sockets. 5627109998Smarkm [Lutz Jaenicke] 5628109998Smarkm 5629109998Smarkm *) Change the Unix RAND_poll() variant to be able to poll several 5630109998Smarkm random devices, as specified by DEVRANDOM, until a sufficient amount 5631109998Smarkm of data has been collected. We spend at most 10 ms on each file 5632109998Smarkm (select timeout) and read in non-blocking mode. DEVRANDOM now 5633109998Smarkm defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 5634109998Smarkm (previously it was just the string "/dev/urandom"), so on typical 5635109998Smarkm platforms the 10 ms delay will never occur. 5636109998Smarkm Also separate out the Unix variant to its own file, rand_unix.c. 5637109998Smarkm For VMS, there's a currently-empty rand_vms.c. 5638109998Smarkm [Richard Levitte] 5639109998Smarkm 5640109998Smarkm *) Move OCSP client related routines to ocsp_cl.c. These 5641109998Smarkm provide utility functions which an application needing 5642109998Smarkm to issue a request to an OCSP responder and analyse the 5643109998Smarkm response will typically need: as opposed to those which an 5644109998Smarkm OCSP responder itself would need which will be added later. 5645109998Smarkm 5646109998Smarkm OCSP_request_sign() signs an OCSP request with an API similar 5647109998Smarkm to PKCS7_sign(). OCSP_response_status() returns status of OCSP 5648109998Smarkm response. OCSP_response_get1_basic() extracts basic response 5649109998Smarkm from response. OCSP_resp_find_status(): finds and extracts status 5650109998Smarkm information from an OCSP_CERTID structure (which will be created 5651109998Smarkm when the request structure is built). These are built from lower 5652109998Smarkm level functions which work on OCSP_SINGLERESP structures but 5653109998Smarkm wont normally be used unless the application wishes to examine 5654109998Smarkm extensions in the OCSP response for example. 5655109998Smarkm 5656109998Smarkm Replace nonce routines with a pair of functions. 5657109998Smarkm OCSP_request_add1_nonce() adds a nonce value and optionally 5658109998Smarkm generates a random value. OCSP_check_nonce() checks the 5659109998Smarkm validity of the nonce in an OCSP response. 5660109998Smarkm [Steve Henson] 5661109998Smarkm 5662109998Smarkm *) Change function OCSP_request_add() to OCSP_request_add0_id(). 5663109998Smarkm This doesn't copy the supplied OCSP_CERTID and avoids the 5664109998Smarkm need to free up the newly created id. Change return type 5665109998Smarkm to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 5666109998Smarkm This can then be used to add extensions to the request. 5667109998Smarkm Deleted OCSP_request_new(), since most of its functionality 5668109998Smarkm is now in OCSP_REQUEST_new() (and the case insensitive name 5669109998Smarkm clash) apart from the ability to set the request name which 5670109998Smarkm will be added elsewhere. 5671109998Smarkm [Steve Henson] 5672109998Smarkm 5673109998Smarkm *) Update OCSP API. Remove obsolete extensions argument from 5674109998Smarkm various functions. Extensions are now handled using the new 5675109998Smarkm OCSP extension code. New simple OCSP HTTP function which 5676109998Smarkm can be used to send requests and parse the response. 5677109998Smarkm [Steve Henson] 5678109998Smarkm 5679109998Smarkm *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 5680109998Smarkm ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 5681109998Smarkm uses the special reorder version of SET OF to sort the attributes 5682109998Smarkm and reorder them to match the encoded order. This resolves a long 5683109998Smarkm standing problem: a verify on a PKCS7 structure just after signing 5684109998Smarkm it used to fail because the attribute order did not match the 5685109998Smarkm encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 5686109998Smarkm it uses the received order. This is necessary to tolerate some broken 5687109998Smarkm software that does not order SET OF. This is handled by encoding 5688109998Smarkm as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 5689109998Smarkm to produce the required SET OF. 5690109998Smarkm [Steve Henson] 5691109998Smarkm 5692109998Smarkm *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 5693109998Smarkm OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 5694109998Smarkm files to get correct declarations of the ASN.1 item variables. 5695109998Smarkm [Richard Levitte] 5696109998Smarkm 5697109998Smarkm *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 5698109998Smarkm PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 5699109998Smarkm asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 5700109998Smarkm NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 5701109998Smarkm New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 5702109998Smarkm ASN1_ITEM and no wrapper functions. 5703109998Smarkm [Steve Henson] 5704109998Smarkm 5705109998Smarkm *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 5706109998Smarkm replace the old function pointer based I/O routines. Change most of 5707109998Smarkm the *_d2i_bio() and *_d2i_fp() functions to use these. 5708109998Smarkm [Steve Henson] 5709109998Smarkm 5710109998Smarkm *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 5711109998Smarkm lines, recognice more "algorithms" that can be deselected, and make 5712109998Smarkm it complain about algorithm deselection that isn't recognised. 5713109998Smarkm [Richard Levitte] 5714109998Smarkm 5715109998Smarkm *) New ASN1 functions to handle dup, sign, verify, digest, pack and 5716109998Smarkm unpack operations in terms of ASN1_ITEM. Modify existing wrappers 5717109998Smarkm to use new functions. Add NO_ASN1_OLD which can be set to remove 5718109998Smarkm some old style ASN1 functions: this can be used to determine if old 5719109998Smarkm code will still work when these eventually go away. 5720109998Smarkm [Steve Henson] 5721109998Smarkm 5722109998Smarkm *) New extension functions for OCSP structures, these follow the 5723109998Smarkm same conventions as certificates and CRLs. 5724109998Smarkm [Steve Henson] 5725109998Smarkm 5726109998Smarkm *) New function X509V3_add1_i2d(). This automatically encodes and 5727109998Smarkm adds an extension. Its behaviour can be customised with various 5728109998Smarkm flags to append, replace or delete. Various wrappers added for 5729109998Smarkm certifcates and CRLs. 5730109998Smarkm [Steve Henson] 5731109998Smarkm 5732109998Smarkm *) Fix to avoid calling the underlying ASN1 print routine when 5733109998Smarkm an extension cannot be parsed. Correct a typo in the 5734109998Smarkm OCSP_SERVICELOC extension. Tidy up print OCSP format. 5735109998Smarkm [Steve Henson] 5736109998Smarkm 5737109998Smarkm *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 5738109998Smarkm entries for variables. 5739109998Smarkm [Steve Henson] 5740109998Smarkm 5741109998Smarkm *) Add functionality to apps/openssl.c for detecting locking 5742109998Smarkm problems: As the program is single-threaded, all we have 5743109998Smarkm to do is register a locking callback using an array for 5744109998Smarkm storing which locks are currently held by the program. 5745109998Smarkm [Bodo Moeller] 5746109998Smarkm 5747109998Smarkm *) Use a lock around the call to CRYPTO_get_ex_new_index() in 5748109998Smarkm SSL_get_ex_data_X509_STORE_idx(), which is used in 5749109998Smarkm ssl_verify_cert_chain() and thus can be called at any time 5750109998Smarkm during TLS/SSL handshakes so that thread-safety is essential. 5751109998Smarkm Unfortunately, the ex_data design is not at all suited 5752109998Smarkm for multi-threaded use, so it probably should be abolished. 5753109998Smarkm [Bodo Moeller] 5754109998Smarkm 5755109998Smarkm *) Added Broadcom "ubsec" ENGINE to OpenSSL. 5756109998Smarkm [Broadcom, tweaked and integrated by Geoff Thorpe] 5757109998Smarkm 5758109998Smarkm *) Move common extension printing code to new function 5759109998Smarkm X509V3_print_extensions(). Reorganise OCSP print routines and 5760109998Smarkm implement some needed OCSP ASN1 functions. Add OCSP extensions. 5761109998Smarkm [Steve Henson] 5762109998Smarkm 5763109998Smarkm *) New function X509_signature_print() to remove duplication in some 5764109998Smarkm print routines. 5765109998Smarkm [Steve Henson] 5766109998Smarkm 5767109998Smarkm *) Add a special meaning when SET OF and SEQUENCE OF flags are both 5768109998Smarkm set (this was treated exactly the same as SET OF previously). This 5769109998Smarkm is used to reorder the STACK representing the structure to match the 5770109998Smarkm encoding. This will be used to get round a problem where a PKCS7 5771109998Smarkm structure which was signed could not be verified because the STACK 5772109998Smarkm order did not reflect the encoded order. 5773109998Smarkm [Steve Henson] 5774109998Smarkm 5775109998Smarkm *) Reimplement the OCSP ASN1 module using the new code. 5776109998Smarkm [Steve Henson] 5777109998Smarkm 5778109998Smarkm *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 5779109998Smarkm for its ASN1 operations. The old style function pointers still exist 5780109998Smarkm for now but they will eventually go away. 5781109998Smarkm [Steve Henson] 5782109998Smarkm 5783109998Smarkm *) Merge in replacement ASN1 code from the ASN1 branch. This almost 5784109998Smarkm completely replaces the old ASN1 functionality with a table driven 5785109998Smarkm encoder and decoder which interprets an ASN1_ITEM structure describing 5786109998Smarkm the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 5787109998Smarkm largely maintained. Almost all of the old asn1_mac.h macro based ASN1 5788109998Smarkm has also been converted to the new form. 5789109998Smarkm [Steve Henson] 5790109998Smarkm 5791109998Smarkm *) Change BN_mod_exp_recp so that negative moduli are tolerated 5792109998Smarkm (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 5793109998Smarkm so that BN_mod_exp_mont and BN_mod_exp_mont_word work 5794109998Smarkm for negative moduli. 5795109998Smarkm [Bodo Moeller] 5796109998Smarkm 5797109998Smarkm *) Fix BN_uadd and BN_usub: Always return non-negative results instead 5798109998Smarkm of not touching the result's sign bit. 5799109998Smarkm [Bodo Moeller] 5800109998Smarkm 5801109998Smarkm *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 5802109998Smarkm set. 5803109998Smarkm [Bodo Moeller] 5804109998Smarkm 5805109998Smarkm *) Changed the LHASH code to use prototypes for callbacks, and created 5806109998Smarkm macros to declare and implement thin (optionally static) functions 5807109998Smarkm that provide type-safety and avoid function pointer casting for the 5808109998Smarkm type-specific callbacks. 5809109998Smarkm [Geoff Thorpe] 5810109998Smarkm 5811109998Smarkm *) Added Kerberos Cipher Suites to be used with TLS, as written in 5812109998Smarkm RFC 2712. 5813109998Smarkm [Veers Staats <staatsvr@asc.hpc.mil>, 5814109998Smarkm Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 5815109998Smarkm 5816109998Smarkm *) Reformat the FAQ so the different questions and answers can be divided 5817109998Smarkm in sections depending on the subject. 5818109998Smarkm [Richard Levitte] 5819109998Smarkm 5820109998Smarkm *) Have the zlib compression code load ZLIB.DLL dynamically under 5821109998Smarkm Windows. 5822109998Smarkm [Richard Levitte] 5823109998Smarkm 5824109998Smarkm *) New function BN_mod_sqrt for computing square roots modulo a prime 5825109998Smarkm (using the probabilistic Tonelli-Shanks algorithm unless 5826109998Smarkm p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 5827109998Smarkm be handled deterministically). 5828109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5829109998Smarkm 5830109998Smarkm *) Make BN_mod_inverse faster by explicitly handling small quotients 5831109998Smarkm in the Euclid loop. (Speed gain about 20% for small moduli [256 or 5832109998Smarkm 512 bits], about 30% for larger ones [1024 or 2048 bits].) 5833109998Smarkm [Bodo Moeller] 5834109998Smarkm 5835109998Smarkm *) New function BN_kronecker. 5836109998Smarkm [Bodo Moeller] 5837109998Smarkm 5838109998Smarkm *) Fix BN_gcd so that it works on negative inputs; the result is 5839109998Smarkm positive unless both parameters are zero. 5840109998Smarkm Previously something reasonably close to an infinite loop was 5841109998Smarkm possible because numbers could be growing instead of shrinking 5842109998Smarkm in the implementation of Euclid's algorithm. 5843109998Smarkm [Bodo Moeller] 5844109998Smarkm 5845109998Smarkm *) Fix BN_is_word() and BN_is_one() macros to take into account the 5846109998Smarkm sign of the number in question. 5847109998Smarkm 5848109998Smarkm Fix BN_is_word(a,w) to work correctly for w == 0. 5849109998Smarkm 5850109998Smarkm The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 5851109998Smarkm because its test if the absolute value of 'a' equals 'w'. 5852109998Smarkm Note that BN_abs_is_word does *not* handle w == 0 reliably; 5853109998Smarkm it exists mostly for use in the implementations of BN_is_zero(), 5854109998Smarkm BN_is_one(), and BN_is_word(). 5855109998Smarkm [Bodo Moeller] 5856109998Smarkm 5857109998Smarkm *) New function BN_swap. 5858109998Smarkm [Bodo Moeller] 5859109998Smarkm 5860109998Smarkm *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 5861109998Smarkm the exponentiation functions are more likely to produce reasonable 5862109998Smarkm results on negative inputs. 5863109998Smarkm [Bodo Moeller] 5864109998Smarkm 5865109998Smarkm *) Change BN_mod_mul so that the result is always non-negative. 5866109998Smarkm Previously, it could be negative if one of the factors was negative; 5867109998Smarkm I don't think anyone really wanted that behaviour. 5868109998Smarkm [Bodo Moeller] 5869109998Smarkm 5870109998Smarkm *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 5871109998Smarkm (except for exponentiation, which stays in crypto/bn/bn_exp.c, 5872109998Smarkm and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 5873109998Smarkm and add new functions: 5874109998Smarkm 5875109998Smarkm BN_nnmod 5876109998Smarkm BN_mod_sqr 5877109998Smarkm BN_mod_add 5878109998Smarkm BN_mod_add_quick 5879109998Smarkm BN_mod_sub 5880109998Smarkm BN_mod_sub_quick 5881109998Smarkm BN_mod_lshift1 5882109998Smarkm BN_mod_lshift1_quick 5883109998Smarkm BN_mod_lshift 5884109998Smarkm BN_mod_lshift_quick 5885109998Smarkm 5886109998Smarkm These functions always generate non-negative results. 5887109998Smarkm 5888109998Smarkm BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 5889109998Smarkm such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 5890109998Smarkm 5891109998Smarkm BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 5892109998Smarkm BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 5893109998Smarkm be reduced modulo m. 5894109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5895109998Smarkm 5896111147Snectar#if 0 5897111147Snectar The following entry accidentily appeared in the CHANGES file 5898111147Snectar distributed with OpenSSL 0.9.7. The modifications described in 5899111147Snectar it do *not* apply to OpenSSL 0.9.7. 5900111147Snectar 5901109998Smarkm *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 5902109998Smarkm was actually never needed) and in BN_mul(). The removal in BN_mul() 5903109998Smarkm required a small change in bn_mul_part_recursive() and the addition 5904109998Smarkm of the functions bn_cmp_part_words(), bn_sub_part_words() and 5905109998Smarkm bn_add_part_words(), which do the same thing as bn_cmp_words(), 5906109998Smarkm bn_sub_words() and bn_add_words() except they take arrays with 5907109998Smarkm differing sizes. 5908109998Smarkm [Richard Levitte] 5909111147Snectar#endif 5910109998Smarkm 5911109998Smarkm *) In 'openssl passwd', verify passwords read from the terminal 5912109998Smarkm unless the '-salt' option is used (which usually means that 5913109998Smarkm verification would just waste user's time since the resulting 5914109998Smarkm hash is going to be compared with some given password hash) 5915109998Smarkm or the new '-noverify' option is used. 5916109998Smarkm 5917109998Smarkm This is an incompatible change, but it does not affect 5918109998Smarkm non-interactive use of 'openssl passwd' (passwords on the command 5919109998Smarkm line, '-stdin' option, '-in ...' option) and thus should not 5920109998Smarkm cause any problems. 5921109998Smarkm [Bodo Moeller] 5922109998Smarkm 5923109998Smarkm *) Remove all references to RSAref, since there's no more need for it. 5924109998Smarkm [Richard Levitte] 5925109998Smarkm 5926109998Smarkm *) Make DSO load along a path given through an environment variable 5927109998Smarkm (SHLIB_PATH) with shl_load(). 5928109998Smarkm [Richard Levitte] 5929109998Smarkm 5930109998Smarkm *) Constify the ENGINE code as a result of BIGNUM constification. 5931109998Smarkm Also constify the RSA code and most things related to it. In a 5932109998Smarkm few places, most notable in the depth of the ASN.1 code, ugly 5933109998Smarkm casts back to non-const were required (to be solved at a later 5934109998Smarkm time) 5935109998Smarkm [Richard Levitte] 5936109998Smarkm 5937109998Smarkm *) Make it so the openssl application has all engines loaded by default. 5938109998Smarkm [Richard Levitte] 5939109998Smarkm 5940109998Smarkm *) Constify the BIGNUM routines a little more. 5941109998Smarkm [Richard Levitte] 5942109998Smarkm 5943109998Smarkm *) Add the following functions: 5944109998Smarkm 5945109998Smarkm ENGINE_load_cswift() 5946109998Smarkm ENGINE_load_chil() 5947109998Smarkm ENGINE_load_atalla() 5948109998Smarkm ENGINE_load_nuron() 5949109998Smarkm ENGINE_load_builtin_engines() 5950109998Smarkm 5951109998Smarkm That way, an application can itself choose if external engines that 5952109998Smarkm are built-in in OpenSSL shall ever be used or not. The benefit is 5953109998Smarkm that applications won't have to be linked with libdl or other dso 5954109998Smarkm libraries unless it's really needed. 5955109998Smarkm 5956109998Smarkm Changed 'openssl engine' to load all engines on demand. 5957109998Smarkm Changed the engine header files to avoid the duplication of some 5958109998Smarkm declarations (they differed!). 5959109998Smarkm [Richard Levitte] 5960109998Smarkm 5961109998Smarkm *) 'openssl engine' can now list capabilities. 5962109998Smarkm [Richard Levitte] 5963109998Smarkm 5964109998Smarkm *) Better error reporting in 'openssl engine'. 5965109998Smarkm [Richard Levitte] 5966109998Smarkm 5967109998Smarkm *) Never call load_dh_param(NULL) in s_server. 5968109998Smarkm [Bodo Moeller] 5969109998Smarkm 5970109998Smarkm *) Add engine application. It can currently list engines by name and 5971109998Smarkm identity, and test if they are actually available. 5972109998Smarkm [Richard Levitte] 5973109998Smarkm 5974109998Smarkm *) Improve RPM specification file by forcing symbolic linking and making 5975109998Smarkm sure the installed documentation is also owned by root.root. 5976109998Smarkm [Damien Miller <djm@mindrot.org>] 5977109998Smarkm 5978109998Smarkm *) Give the OpenSSL applications more possibilities to make use of 5979109998Smarkm keys (public as well as private) handled by engines. 5980109998Smarkm [Richard Levitte] 5981109998Smarkm 5982109998Smarkm *) Add OCSP code that comes from CertCo. 5983109998Smarkm [Richard Levitte] 5984109998Smarkm 5985109998Smarkm *) Add VMS support for the Rijndael code. 5986109998Smarkm [Richard Levitte] 5987109998Smarkm 5988109998Smarkm *) Added untested support for Nuron crypto accelerator. 5989109998Smarkm [Ben Laurie] 5990109998Smarkm 5991109998Smarkm *) Add support for external cryptographic devices. This code was 5992109998Smarkm previously distributed separately as the "engine" branch. 5993109998Smarkm [Geoff Thorpe, Richard Levitte] 5994109998Smarkm 5995109998Smarkm *) Rework the filename-translation in the DSO code. It is now possible to 5996109998Smarkm have far greater control over how a "name" is turned into a filename 5997109998Smarkm depending on the operating environment and any oddities about the 5998109998Smarkm different shared library filenames on each system. 5999109998Smarkm [Geoff Thorpe] 6000109998Smarkm 6001109998Smarkm *) Support threads on FreeBSD-elf in Configure. 6002109998Smarkm [Richard Levitte] 6003109998Smarkm 6004109998Smarkm *) Fix for SHA1 assembly problem with MASM: it produces 6005109998Smarkm warnings about corrupt line number information when assembling 6006109998Smarkm with debugging information. This is caused by the overlapping 6007109998Smarkm of two sections. 6008109998Smarkm [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 6009109998Smarkm 6010109998Smarkm *) NCONF changes. 6011109998Smarkm NCONF_get_number() has no error checking at all. As a replacement, 6012109998Smarkm NCONF_get_number_e() is defined (_e for "error checking") and is 6013109998Smarkm promoted strongly. The old NCONF_get_number is kept around for 6014109998Smarkm binary backward compatibility. 6015109998Smarkm Make it possible for methods to load from something other than a BIO, 6016109998Smarkm by providing a function pointer that is given a name instead of a BIO. 6017109998Smarkm For example, this could be used to load configuration data from an 6018109998Smarkm LDAP server. 6019109998Smarkm [Richard Levitte] 6020109998Smarkm 6021109998Smarkm *) Fix for non blocking accept BIOs. Added new I/O special reason 6022109998Smarkm BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 6023109998Smarkm with non blocking I/O was not possible because no retry code was 6024109998Smarkm implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 6025109998Smarkm this case. 6026109998Smarkm [Steve Henson] 6027109998Smarkm 6028109998Smarkm *) Added the beginnings of Rijndael support. 6029109998Smarkm [Ben Laurie] 6030109998Smarkm 6031109998Smarkm *) Fix for bug in DirectoryString mask setting. Add support for 6032109998Smarkm X509_NAME_print_ex() in 'req' and X509_print_ex() function 6033109998Smarkm to allow certificate printing to more controllable, additional 6034109998Smarkm 'certopt' option to 'x509' to allow new printing options to be 6035109998Smarkm set. 6036109998Smarkm [Steve Henson] 6037109998Smarkm 6038109998Smarkm *) Clean old EAY MD5 hack from e_os.h. 6039109998Smarkm [Richard Levitte] 6040109998Smarkm 6041142425Snectar Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 6042142425Snectar 6043142425Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 6044160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 6045142425Snectar [Joe Orton, Steve Henson] 6046142425Snectar 6047142425Snectar Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 6048142425Snectar 6049142425Snectar *) Fix additional bug revealed by the NISCC test suite: 6050142425Snectar 6051142425Snectar Stop bug triggering large recursion when presented with 6052160814Ssimon certain ASN.1 tags (CVE-2003-0851) 6053142425Snectar [Steve Henson] 6054142425Snectar 6055120631Snectar Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 6056120631Snectar 6057120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 6058120631Snectar 6059120631Snectar Stop out of bounds reads in the ASN1 code when presented with 6060160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 6061120631Snectar 6062120631Snectar If verify callback ignores invalid public key errors don't try to check 6063120631Snectar certificate signature with the NULL public key. 6064120631Snectar 6065120631Snectar [Steve Henson] 6066120631Snectar 6067120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 6068120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 6069120631Snectar specifications. 6070120631Snectar [Steve Henson] 6071120631Snectar 6072120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 6073120631Snectar extra data after the compression methods not only for TLS 1.0 6074120631Snectar but also for SSL 3.0 (as required by the specification). 6075120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 6076120631Snectar 6077120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 6078120631Snectar when it's 512 *bits* long, not 512 bytes. 6079120631Snectar [Richard Levitte] 6080120631Snectar 6081120631Snectar Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 6082120631Snectar 6083120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 6084120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 6085120631Snectar a protocol version number mismatch like a decryption error 6086120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 6087120631Snectar [Bodo Moeller] 6088120631Snectar 6089120631Snectar *) Turn on RSA blinding by default in the default implementation 6090120631Snectar to avoid a timing attack. Applications that don't want it can call 6091120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 6092120631Snectar They would be ill-advised to do so in most cases. 6093120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 6094120631Snectar 6095120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 6096120631Snectar seeded (in this case, the secret RSA exponent is abused as 6097120631Snectar an unpredictable seed -- if it is not unpredictable, there 6098120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 6099120631Snectar by remembering the creator's thread ID in rsa->blinding and 6100120631Snectar having all other threads use local one-time blinding factors 6101120631Snectar (this requires more computation than sharing rsa->blinding, but 6102120631Snectar avoids excessive locking; and if an RSA object is not shared 6103120631Snectar between threads, blinding will still be very fast). 6104120631Snectar [Bodo Moeller] 6105120631Snectar 6106111147Snectar Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 6107111147Snectar 6108111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 6109111147Snectar via timing by performing a MAC computation even if incorrrect 6110111147Snectar block cipher padding has been found. This is a countermeasure 6111111147Snectar against active attacks where the attacker has to distinguish 6112160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 6113111147Snectar 6114111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 6115111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 6116111147Snectar Martin Vuagnoux (EPFL, Ilion)] 6117111147Snectar 6118109998Smarkm Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 6119109998Smarkm 6120109998Smarkm *) New function OPENSSL_cleanse(), which is used to cleanse a section of 6121109998Smarkm memory from it's contents. This is done with a counter that will 6122109998Smarkm place alternating values in each byte. This can be used to solve 6123109998Smarkm two issues: 1) the removal of calls to memset() by highly optimizing 6124109998Smarkm compilers, and 2) cleansing with other values than 0, since those can 6125109998Smarkm be read through on certain media, for example a swap space on disk. 6126109998Smarkm [Geoff Thorpe] 6127109998Smarkm 6128109998Smarkm *) Bugfix: client side session caching did not work with external caching, 6129109998Smarkm because the session->cipher setting was not restored when reloading 6130109998Smarkm from the external cache. This problem was masked, when 6131109998Smarkm SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 6132109998Smarkm (Found by Steve Haslam <steve@araqnid.ddts.net>.) 6133109998Smarkm [Lutz Jaenicke] 6134109998Smarkm 6135109998Smarkm *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 6136109998Smarkm length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 6137109998Smarkm [Zeev Lieber <zeev-l@yahoo.com>] 6138109998Smarkm 6139109998Smarkm *) Undo an undocumented change introduced in 0.9.6e which caused 6140109998Smarkm repeated calls to OpenSSL_add_all_ciphers() and 6141109998Smarkm OpenSSL_add_all_digests() to be ignored, even after calling 6142109998Smarkm EVP_cleanup(). 6143109998Smarkm [Richard Levitte] 6144109998Smarkm 6145109998Smarkm *) Change the default configuration reader to deal with last line not 6146109998Smarkm being properly terminated. 6147109998Smarkm [Richard Levitte] 6148109998Smarkm 6149109998Smarkm *) Change X509_NAME_cmp() so it applies the special rules on handling 6150109998Smarkm DN values that are of type PrintableString, as well as RDNs of type 6151109998Smarkm emailAddress where the value has the type ia5String. 6152109998Smarkm [stefank@valicert.com via Richard Levitte] 6153109998Smarkm 6154109998Smarkm *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 6155109998Smarkm the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 6156109998Smarkm doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 6157109998Smarkm the bitwise-OR of the two for use by the majority of applications 6158109998Smarkm wanting this behaviour, and update the docs. The documented 6159109998Smarkm behaviour and actual behaviour were inconsistent and had been 6160109998Smarkm changing anyway, so this is more a bug-fix than a behavioural 6161109998Smarkm change. 6162109998Smarkm [Geoff Thorpe, diagnosed by Nadav Har'El] 6163109998Smarkm 6164109998Smarkm *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 6165109998Smarkm (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 6166109998Smarkm [Bodo Moeller] 6167109998Smarkm 6168109998Smarkm *) Fix initialization code race conditions in 6169109998Smarkm SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 6170109998Smarkm SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 6171109998Smarkm SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 6172109998Smarkm TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 6173109998Smarkm ssl2_get_cipher_by_char(), 6174109998Smarkm ssl3_get_cipher_by_char(). 6175109998Smarkm [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 6176109998Smarkm 6177109998Smarkm *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 6178109998Smarkm the cached sessions are flushed, as the remove_cb() might use ex_data 6179109998Smarkm contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 6180109998Smarkm (see [openssl.org #212]). 6181109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 6182109998Smarkm 6183109998Smarkm *) Fix typo in OBJ_txt2obj which incorrectly passed the content 6184109998Smarkm length, instead of the encoding length to d2i_ASN1_OBJECT. 6185109998Smarkm [Steve Henson] 6186109998Smarkm 6187101618Snectar Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 6188101618Snectar 6189101618Snectar *) [In 0.9.6g-engine release:] 6190101618Snectar Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 6191101618Snectar [Lynn Gazis <lgazis@rainbow.com>] 6192101618Snectar 6193101613Snectar Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 6194101613Snectar 6195101613Snectar *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 6196101613Snectar and get fix the header length calculation. 6197101613Snectar [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 6198101613Snectar Alon Kantor <alonk@checkpoint.com> (and others), 6199101613Snectar Steve Henson] 6200101613Snectar 6201101613Snectar *) Use proper error handling instead of 'assertions' in buffer 6202101613Snectar overflow checks added in 0.9.6e. This prevents DoS (the 6203101613Snectar assertions could call abort()). 6204101613Snectar [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 6205101613Snectar 6206100936Snectar Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 6207100936Snectar 6208109998Smarkm *) Add various sanity checks to asn1_get_length() to reject 6209109998Smarkm the ASN1 length bytes if they exceed sizeof(long), will appear 6210109998Smarkm negative or the content length exceeds the length of the 6211109998Smarkm supplied buffer. 6212109998Smarkm [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6213109998Smarkm 6214100936Snectar *) Fix cipher selection routines: ciphers without encryption had no flags 6215100936Snectar for the cipher strength set and where therefore not handled correctly 6216100936Snectar by the selection routines (PR #130). 6217100936Snectar [Lutz Jaenicke] 6218100936Snectar 6219100936Snectar *) Fix EVP_dsa_sha macro. 6220100936Snectar [Nils Larsch] 6221100936Snectar 6222100936Snectar *) New option 6223100936Snectar SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 6224100936Snectar for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 6225100936Snectar that was added in OpenSSL 0.9.6d. 6226100936Snectar 6227100936Snectar As the countermeasure turned out to be incompatible with some 6228100936Snectar broken SSL implementations, the new option is part of SSL_OP_ALL. 6229100936Snectar SSL_OP_ALL is usually employed when compatibility with weird SSL 6230100936Snectar implementations is desired (e.g. '-bugs' option to 's_client' and 6231100936Snectar 's_server'), so the new option is automatically set in many 6232100936Snectar applications. 6233100936Snectar [Bodo Moeller] 6234100936Snectar 6235100936Snectar *) Changes in security patch: 6236100936Snectar 6237100936Snectar Changes marked "(CHATS)" were sponsored by the Defense Advanced 6238100936Snectar Research Projects Agency (DARPA) and Air Force Research Laboratory, 6239100936Snectar Air Force Materiel Command, USAF, under agreement number 6240100936Snectar F30602-01-2-0537. 6241100936Snectar 6242100936Snectar *) Add various sanity checks to asn1_get_length() to reject 6243100936Snectar the ASN1 length bytes if they exceed sizeof(long), will appear 6244100936Snectar negative or the content length exceeds the length of the 6245160814Ssimon supplied buffer. (CVE-2002-0659) 6246100936Snectar [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6247100936Snectar 6248100936Snectar *) Assertions for various potential buffer overflows, not known to 6249100936Snectar happen in practice. 6250100936Snectar [Ben Laurie (CHATS)] 6251100936Snectar 6252100936Snectar *) Various temporary buffers to hold ASCII versions of integers were 6253160814Ssimon too small for 64 bit platforms. (CVE-2002-0655) 6254100936Snectar [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 6255100936Snectar 6256100936Snectar *) Remote buffer overflow in SSL3 protocol - an attacker could 6257160814Ssimon supply an oversized session ID to a client. (CVE-2002-0656) 6258100936Snectar [Ben Laurie (CHATS)] 6259100936Snectar 6260100936Snectar *) Remote buffer overflow in SSL2 protocol - an attacker could 6261160814Ssimon supply an oversized client master key. (CVE-2002-0656) 6262100936Snectar [Ben Laurie (CHATS)] 6263100936Snectar 6264100928Snectar Changes between 0.9.6c and 0.9.6d [9 May 2002] 6265100928Snectar 6266100928Snectar *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 6267100928Snectar encoded as NULL) with id-dsa-with-sha1. 6268100928Snectar [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 6269100928Snectar 6270100928Snectar *) Check various X509_...() return values in apps/req.c. 6271100928Snectar [Nils Larsch <nla@trustcenter.de>] 6272100928Snectar 6273100928Snectar *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 6274100928Snectar an end-of-file condition would erronously be flagged, when the CRLF 6275100928Snectar was just at the end of a processed block. The bug was discovered when 6276100928Snectar processing data through a buffering memory BIO handing the data to a 6277100928Snectar BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 6278100928Snectar <ptsekov@syntrex.com> and Nedelcho Stanev. 6279100928Snectar [Lutz Jaenicke] 6280100928Snectar 6281100928Snectar *) Implement a countermeasure against a vulnerability recently found 6282100928Snectar in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 6283100928Snectar before application data chunks to avoid the use of known IVs 6284100928Snectar with data potentially chosen by the attacker. 6285100928Snectar [Bodo Moeller] 6286100928Snectar 6287100928Snectar *) Fix length checks in ssl3_get_client_hello(). 6288100928Snectar [Bodo Moeller] 6289100928Snectar 6290100928Snectar *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 6291100928Snectar to prevent ssl3_read_internal() from incorrectly assuming that 6292100928Snectar ssl3_read_bytes() found application data while handshake 6293100928Snectar processing was enabled when in fact s->s3->in_read_app_data was 6294100928Snectar merely automatically cleared during the initial handshake. 6295100928Snectar [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 6296100928Snectar 6297100928Snectar *) Fix object definitions for Private and Enterprise: they were not 6298100928Snectar recognized in their shortname (=lowercase) representation. Extend 6299100928Snectar obj_dat.pl to issue an error when using undefined keywords instead 6300100928Snectar of silently ignoring the problem (Svenning Sorensen 6301100928Snectar <sss@sss.dnsalias.net>). 6302100928Snectar [Lutz Jaenicke] 6303100928Snectar 6304100928Snectar *) Fix DH_generate_parameters() so that it works for 'non-standard' 6305100928Snectar generators, i.e. generators other than 2 and 5. (Previously, the 6306100928Snectar code did not properly initialise the 'add' and 'rem' values to 6307100928Snectar BN_generate_prime().) 6308100928Snectar 6309100928Snectar In the new general case, we do not insist that 'generator' is 6310100928Snectar actually a primitive root: This requirement is rather pointless; 6311100928Snectar a generator of the order-q subgroup is just as good, if not 6312100928Snectar better. 6313100928Snectar [Bodo Moeller] 6314100928Snectar 6315100928Snectar *) Map new X509 verification errors to alerts. Discovered and submitted by 6316100928Snectar Tom Wu <tom@arcot.com>. 6317100928Snectar [Lutz Jaenicke] 6318100928Snectar 6319100928Snectar *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 6320100928Snectar returning non-zero before the data has been completely received 6321100928Snectar when using non-blocking I/O. 6322100928Snectar [Bodo Moeller; problem pointed out by John Hughes] 6323100928Snectar 6324100928Snectar *) Some of the ciphers missed the strength entry (SSL_LOW etc). 6325100928Snectar [Ben Laurie, Lutz Jaenicke] 6326100928Snectar 6327100928Snectar *) Fix bug in SSL_clear(): bad sessions were not removed (found by 6328100928Snectar Yoram Zahavi <YoramZ@gilian.com>). 6329100928Snectar [Lutz Jaenicke] 6330100928Snectar 6331100928Snectar *) Add information about CygWin 1.3 and on, and preserve proper 6332100928Snectar configuration for the versions before that. 6333100928Snectar [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 6334100928Snectar 6335100928Snectar *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 6336100928Snectar check whether we deal with a copy of a session and do not delete from 6337100928Snectar the cache in this case. Problem reported by "Izhar Shoshani Levi" 6338100928Snectar <izhar@checkpoint.com>. 6339100928Snectar [Lutz Jaenicke] 6340100928Snectar 6341100928Snectar *) Do not store session data into the internal session cache, if it 6342100928Snectar is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 6343100928Snectar flag is set). Proposed by Aslam <aslam@funk.com>. 6344100928Snectar [Lutz Jaenicke] 6345100928Snectar 6346100928Snectar *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 6347100928Snectar value is 0. 6348100928Snectar [Richard Levitte] 6349100928Snectar 6350109998Smarkm *) [In 0.9.6d-engine release:] 6351109998Smarkm Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 6352100928Snectar [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 6353100928Snectar 6354100928Snectar *) Add the configuration target linux-s390x. 6355100928Snectar [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 6356100928Snectar 6357100928Snectar *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 6358100928Snectar ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 6359100928Snectar variable as an indication that a ClientHello message has been 6360100928Snectar received. As the flag value will be lost between multiple 6361100928Snectar invocations of ssl3_accept when using non-blocking I/O, the 6362100928Snectar function may not be aware that a handshake has actually taken 6363100928Snectar place, thus preventing a new session from being added to the 6364100928Snectar session cache. 6365100928Snectar 6366100928Snectar To avoid this problem, we now set s->new_session to 2 instead of 6367100928Snectar using a local variable. 6368100928Snectar [Lutz Jaenicke, Bodo Moeller] 6369100928Snectar 6370100928Snectar *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 6371100928Snectar if the SSL_R_LENGTH_MISMATCH error is detected. 6372100928Snectar [Geoff Thorpe, Bodo Moeller] 6373100928Snectar 6374100928Snectar *) New 'shared_ldflag' column in Configure platform table. 6375100928Snectar [Richard Levitte] 6376100928Snectar 6377100928Snectar *) Fix EVP_CIPHER_mode macro. 6378100928Snectar ["Dan S. Camper" <dan@bti.net>] 6379100928Snectar 6380100928Snectar *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 6381100928Snectar type, we must throw them away by setting rr->length to 0. 6382100928Snectar [D P Chang <dpc@qualys.com>] 6383100928Snectar 638489837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 638589837Skris 638689837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 638789837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 638889837Skris worked incorrectly for those cases where range = 10..._2 and 638989837Skris 3*range is two bits longer than range.) 639089837Skris [Bodo Moeller] 639189837Skris 639289837Skris *) Only add signing time to PKCS7 structures if it is not already 639389837Skris present. 639489837Skris [Steve Henson] 639589837Skris 639689837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 639789837Skris OBJ_ld_ce should be OBJ_id_ce. 639889837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 639989837Skris incorrect (cf. RFC 3039). 640089837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 640189837Skris 640289837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 640389837Skris returns early because it has nothing to do. 640489837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 640589837Skris 640689837Skris *) [In 0.9.6c-engine release:] 640789837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 640889837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 640989837Skris 641089837Skris *) [In 0.9.6c-engine release:] 641189837Skris Add support for Cryptographic Appliance's keyserver technology. 641289837Skris (Use engine 'keyclient') 641389837Skris [Cryptographic Appliances and Geoff Thorpe] 641489837Skris 641589837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 641689837Skris is called via tools/c89.sh because arguments have to be 641789837Skris rearranged (all '-L' options must appear before the first object 641889837Skris modules). 641989837Skris [Richard Shapiro <rshapiro@abinitio.com>] 642089837Skris 642189837Skris *) [In 0.9.6c-engine release:] 642289837Skris Add support for Broadcom crypto accelerator cards, backported 642389837Skris from 0.9.7. 642489837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 642589837Skris 642689837Skris *) [In 0.9.6c-engine release:] 642789837Skris Add support for SureWare crypto accelerator cards from 642889837Skris Baltimore Technologies. (Use engine 'sureware') 642989837Skris [Baltimore Technologies and Mark Cox] 643089837Skris 643189837Skris *) [In 0.9.6c-engine release:] 643289837Skris Add support for crypto accelerator cards from Accelerated 643389837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 643489837Skris [AEP Inc. and Mark Cox] 643589837Skris 643689837Skris *) Add a configuration entry for gcc on UnixWare. 643789837Skris [Gary Benson <gbenson@redhat.com>] 643889837Skris 643989837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 644089837Skris messages are stored in a single piece (fixed-length part and 644189837Skris variable-length part combined) and fix various bugs found on the way. 644289837Skris [Bodo Moeller] 644389837Skris 644489837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 644589837Skris instead. BIO_gethostbyname() does not know what timeouts are 644689837Skris appropriate, so entries would stay in cache even when they have 644789837Skris become invalid. 644889837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 644989837Skris 645089837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 645189837Skris faced with a pathologically small ClientHello fragment that does 645289837Skris not contain client_version: Instead of aborting with an error, 645389837Skris simply choose the highest available protocol version (i.e., 645489837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 645589837Skris messages are never sent like this, but this change gives us 645689837Skris strictly correct behaviour at least for TLS. 645789837Skris [Bodo Moeller] 645889837Skris 645989837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 646089837Skris never resets s->method to s->ctx->method when called from within 646189837Skris one of the SSL handshake functions. 646289837Skris [Bodo Moeller; problem pointed out by Niko Baric] 646389837Skris 646489837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 646589837Skris (sent using the client's version number) if client_version is 646689837Skris smaller than the protocol version in use. Also change 646789837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 646889837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 646989837Skris the client will at least see that alert. 647089837Skris [Bodo Moeller] 647189837Skris 647289837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 647389837Skris correctly. 647489837Skris [Bodo Moeller] 647589837Skris 647689837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 647789837Skris client receives HelloRequest while in a handshake. 647889837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 647989837Skris 648089837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 648189837Skris should end in 'break', not 'goto end' which circuments various 648289837Skris cleanups done in state SSL_ST_OK. But session related stuff 648389837Skris must be disabled for SSL_ST_OK in the case that we just sent a 648489837Skris HelloRequest. 648589837Skris 648689837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 648789837Skris before just sending a HelloRequest. 648889837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 648989837Skris 649089837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 649189837Skris reveal whether illegal block cipher padding was found or a MAC 649289837Skris verification error occured. (Neither SSLerr() codes nor alerts 649389837Skris are directly visible to potential attackers, but the information 649489837Skris may leak via logfiles.) 649589837Skris 649689837Skris Similar changes are not required for the SSL 2.0 implementation 649789837Skris because the number of padding bytes is sent in clear for SSL 2.0, 649889837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 649989837Skris failed to verify that the purported number of padding bytes is in 650089837Skris the legal range. 650189837Skris [Bodo Moeller] 650289837Skris 650389837Skris *) Add OpenUNIX-8 support including shared libraries 650489837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 650589837Skris [Lutz Jaenicke] 650689837Skris 650789837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 650889837Skris 'wristwatch attack' using huge encoding parameters (cf. 650989837Skris James H. Manger's CRYPTO 2001 paper). Note that the 651089837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 651189837Skris encoding parameters and hence was not vulnerable. 651289837Skris [Bodo Moeller] 651389837Skris 651489837Skris *) BN_sqr() bug fix. 651589837Skris [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 651689837Skris 651789837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 651889837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 651989837Skris followed by modular reduction. 652089837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 652189837Skris 652289837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 652389837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 652489837Skris [Bodo Moeller] 652589837Skris 652689837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 652789837Skris This function was broken, as the check for a new client hello message 652889837Skris to handle SGC did not allow these large messages. 652989837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 653089837Skris [Lutz Jaenicke] 653189837Skris 653289837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 653389837Skris [Lutz Jaenicke] 653489837Skris 653589837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 653689837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 653789837Skris [Lutz Jaenicke] 653889837Skris 653989837Skris *) Rework the configuration and shared library support for Tru64 Unix. 654089837Skris The configuration part makes use of modern compiler features and 654189837Skris still retains old compiler behavior for those that run older versions 654289837Skris of the OS. The shared library support part includes a variant that 654389837Skris uses the RPATH feature, and is available through the special 654489837Skris configuration target "alpha-cc-rpath", which will never be selected 654589837Skris automatically. 654689837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 654789837Skris 654889837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 654989837Skris with the same message size as in ssl3_get_certificate_request(). 655089837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 655189837Skris messages might inadvertently be reject as too long. 655289837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 655389837Skris 655489837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 655589837Skris [Andy Polyakov] 655689837Skris 655789837Skris *) Modified SSL library such that the verify_callback that has been set 655889837Skris specificly for an SSL object with SSL_set_verify() is actually being 655989837Skris used. Before the change, a verify_callback set with this function was 656089837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 656189837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 656289837Skris to allow the necessary settings. 656389837Skris [Lutz Jaenicke] 656489837Skris 656589837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 656689837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 656789837Skris done automatically (in contradiction to the requirements of the C 656889837Skris standard). This made problems when used from OpenSSH. 656989837Skris [Lutz Jaenicke] 657089837Skris 657189837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 657289837Skris dh->length and always used 657389837Skris 657489837Skris BN_rand_range(priv_key, dh->p). 657589837Skris 657689837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 657789837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 657889837Skris dh->length (recommended exponent length) is much smaller than the 657989837Skris length of dh->p. We could use BN_rand_range() if the order of 658089837Skris the subgroup was stored in the DH structure, but we only have 658189837Skris dh->length. 658289837Skris 658389837Skris So switch back to 658489837Skris 658589837Skris BN_rand(priv_key, l, ...) 658689837Skris 658789837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 658889837Skris otherwise. 658989837Skris [Bodo Moeller] 659089837Skris 659189837Skris *) In 659289837Skris 659389837Skris RSA_eay_public_encrypt 659489837Skris RSA_eay_private_decrypt 659589837Skris RSA_eay_private_encrypt (signing) 659689837Skris RSA_eay_public_decrypt (signature verification) 659789837Skris 659889837Skris (default implementations for RSA_public_encrypt, 659989837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 660089837Skris always reject numbers >= n. 660189837Skris [Bodo Moeller] 660289837Skris 660389837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 660489837Skris to synchronize access to 'locking_thread'. This is necessary on 660589837Skris systems where access to 'locking_thread' (an 'unsigned long' 660689837Skris variable) is not atomic. 660789837Skris [Bodo Moeller] 660889837Skris 660989837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 661089837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 661189837Skris a race condition if 0 is a valid thread ID. 661289837Skris [Travis Vitek <vitek@roguewave.com>] 661389837Skris 661489837Skris *) Add support for shared libraries under Irix. 661589837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 661689837Skris 661789837Skris *) Add configuration option to build on Linux on both big-endian and 661889837Skris little-endian MIPS. 661989837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 662089837Skris 662189837Skris *) Add the possibility to create shared libraries on HP-UX. 662289837Skris [Richard Levitte] 662389837Skris 662479998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 662579998Skris 662679998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 662779998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 662879998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 662979998Skris PRNG state recovery was possible based on the output of 663079998Skris one PRNG request appropriately sized to gain knowledge on 663179998Skris 'md' followed by enough consecutive 1-byte PRNG requests 663279998Skris to traverse all of 'state'. 663379998Skris 663479998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 663579998Skris during PRNG output generation, hash all of the previous 663679998Skris 'md_local' value, not just the half used for PRNG output. 663779998Skris 663879998Skris 2. Make the number of bytes from 'state' included into the hash 663979998Skris independent from the number of PRNG bytes requested. 664079998Skris 664179998Skris The first measure alone would be sufficient to avoid 664279998Skris Markku-Juhani's attack. (Actually it had never occurred 664379998Skris to me that the half of 'md_local' used for chaining was the 664479998Skris half from which PRNG output bytes were taken -- I had always 664579998Skris assumed that the secret half would be used.) The second 664679998Skris measure makes sure that additional data from 'state' is never 664779998Skris mixed into 'md_local' in small portions; this heuristically 664879998Skris further strengthens the PRNG. 664979998Skris [Bodo Moeller] 665079998Skris 665179998Skris *) Fix crypto/bn/asm/mips3.s. 665279998Skris [Andy Polyakov] 665379998Skris 665479998Skris *) When only the key is given to "enc", the IV is undefined. Print out 665579998Skris an error message in this case. 665679998Skris [Lutz Jaenicke] 665779998Skris 665879998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 665979998Skris [Steve Henson] 666079998Skris 666179998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 666279998Skris positive and less than q. 666379998Skris [Bodo Moeller] 666479998Skris 666579998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 666679998Skris used: it isn't thread safe and the add_lock_callback should handle 666779998Skris that itself. 666879998Skris [Paul Rose <Paul.Rose@bridge.com>] 666979998Skris 667079998Skris *) Verify that incoming data obeys the block size in 667179998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 667279998Skris [Bodo Moeller] 667379998Skris 667479998Skris *) Fix OAEP check. 667579998Skris [Ulf M�ller, Bodo M�ller] 667679998Skris 667779998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 667889837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 667979998Skris when fixing the server behaviour for backwards-compatible 'client 668079998Skris hello' messages. (Note that the attack is impractical against 668179998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 668279998Skris means that the probability of guessing a valid ciphertext is 668379998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 668479998Skris paper.) 668579998Skris 668679998Skris Before 0.9.5, the countermeasure (hide the error by generating a 668779998Skris random 'decryption result') did not work properly because 668879998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 668979998Skris detect the supposedly ignored error. 669079998Skris 669179998Skris Both problems are now fixed. 669279998Skris [Bodo Moeller] 669379998Skris 669479998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 669579998Skris (previously it was 1024). 669679998Skris [Bodo Moeller] 669779998Skris 669879998Skris *) Fix for compatibility mode trust settings: ignore trust settings 669979998Skris unless some valid trust or reject settings are present. 670079998Skris [Steve Henson] 670179998Skris 670279998Skris *) Fix for blowfish EVP: its a variable length cipher. 670379998Skris [Steve Henson] 670479998Skris 670579998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 670679998Skris parameters in DSA public key structures and return an error in the 670779998Skris DSA routines if parameters are absent. 670879998Skris [Steve Henson] 670979998Skris 671079998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 671179998Skris in the current directory if neither $RANDFILE nor $HOME was set. 671279998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 671379998Skris caused some confusion to Windows users who haven't defined $HOME. 671479998Skris Thus RAND_file_name() is changed again: e_os.h can define a 671579998Skris DEFAULT_HOME, which will be used if $HOME is not set. 671679998Skris For Windows, we use "C:"; on other platforms, we still require 671779998Skris environment variables. 671879998Skris 671979998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 672079998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 672179998Skris having multiple threads call RAND_poll() concurrently. 672279998Skris [Bodo Moeller] 672379998Skris 672479998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 672579998Skris combination of a flag and a thread ID variable. 672679998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 672779998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 672889837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 672979998Skris that they do not hold after the first thread unsets add_do_not_lock). 673079998Skris [Bodo Moeller] 673179998Skris 673279998Skris *) Change bctest again: '-x' expressions are not available in all 673379998Skris versions of 'test'. 673479998Skris [Bodo Moeller] 673579998Skris 673676866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 673772613Skris 673876866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 673976866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 674076866Skris 674176866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 674276866Skris the default extension for executables, if any. Also, make the perl 674376866Skris scripts that use symlink() to test if it really exists and use "cp" 674476866Skris if it doesn't. All this made OpenSSL compilable and installable in 674576866Skris CygWin. 674676866Skris [Richard Levitte] 674776866Skris 674876866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 674976866Skris If SEQUENCE is length is indefinite just set c->slen to the total 675076866Skris amount of data available. 675176866Skris [Steve Henson, reported by shige@FreeBSD.org] 675276866Skris [This change does not apply to 0.9.7.] 675376866Skris 675476866Skris *) Change bctest to avoid here-documents inside command substitution 675576866Skris (workaround for FreeBSD /bin/sh bug). 675676866Skris For compatibility with Ultrix, avoid shell functions (introduced 675776866Skris in the bctest version that searches along $PATH). 675876866Skris [Bodo Moeller] 675976866Skris 676076866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 676176866Skris with des_encrypt() defined on some operating systems, like Solaris 676276866Skris and UnixWare. 676376866Skris [Richard Levitte] 676476866Skris 676576866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 676676866Skris On the Importance of Eliminating Errors in Cryptographic 676776866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 676876866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 676976866Skris [Ulf Moeller] 677076866Skris 677176866Skris *) MIPS assembler BIGNUM division bug fix. 677276866Skris [Andy Polyakov] 677376866Skris 677476866Skris *) Disabled incorrect Alpha assembler code. 677576866Skris [Richard Levitte] 677676866Skris 677776866Skris *) Fix PKCS#7 decode routines so they correctly update the length 677876866Skris after reading an EOC for the EXPLICIT tag. 677976866Skris [Steve Henson] 678076866Skris [This change does not apply to 0.9.7.] 678176866Skris 678276866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 678376866Skris if a 3DES key was generated with a 0 initial byte. Include 678476866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 678576866Skris (but broken) behaviour. 678676866Skris [Steve Henson] 678776866Skris 678876866Skris *) Enhance bctest to search for a working bc along $PATH and print 678976866Skris it when found. 679076866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 679176866Skris 679276866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 679376866Skris don't write to the wrong index in ERR_set_error_data. 679476866Skris [Bodo Moeller] 679576866Skris 679676866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 679776866Skris did not exist. 679876866Skris [Bodo Moeller] 679976866Skris 680076866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 680176866Skris [Jeremy Cooper <jeremy@baymoo.org>] 680276866Skris 680376866Skris *) Make it possible to reuse SSLv2 sessions. 680476866Skris [Richard Levitte] 680576866Skris 680676866Skris *) In copy_email() check for >= 0 as a return value for 680776866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 680876866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 680976866Skris 681076866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 681176866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 681276866Skris PKCS7_verify() fails with non detached data. 681376866Skris [Steve Henson] 681476866Skris 681576866Skris *) Don't use getenv in library functions when run as setuid/setgid. 681676866Skris New function OPENSSL_issetugid(). 681776866Skris [Ulf Moeller] 681876866Skris 681976866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 682076866Skris due to incorrect handling of multi-threading: 682176866Skris 682276866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 682376866Skris 682476866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 682576866Skris 682676866Skris 3. Count how many times MemCheck_off() has been called so that 682776866Skris nested use can be treated correctly. This also avoids 682876866Skris inband-signalling in the previous code (which relied on the 682976866Skris assumption that thread ID 0 is impossible). 683076866Skris [Bodo Moeller] 683176866Skris 683276866Skris *) Add "-rand" option also to s_client and s_server. 683376866Skris [Lutz Jaenicke] 683476866Skris 683576866Skris *) Fix CPU detection on Irix 6.x. 683676866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 683776866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 683876866Skris 683976866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 684076866Skris was empty. 684176866Skris [Steve Henson] 684276866Skris [This change does not apply to 0.9.7.] 684376866Skris 684476866Skris *) Use the cached encoding of an X509_NAME structure rather than 684576866Skris copying it. This is apparently the reason for the libsafe "errors" 684676866Skris but the code is actually correct. 684776866Skris [Steve Henson] 684876866Skris 684972613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 685072613Skris Bleichenbacher's DSA attack. 685176866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 685276866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 685376866Skris and leaves the highest bit random. 685476866Skris [Ulf Moeller, Bodo Moeller] 685572613Skris 685672613Skris *) In the NCONF_...-based implementations for CONF_... queries 685772613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 685872613Skris a temporary CONF structure with the data component set to NULL 685972613Skris (which gives segmentation faults in lh_retrieve). 686072613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 686172613Skris CONF_get_number (which may use environment variables) and directly 686272613Skris return NULL from CONF_get_section. 686372613Skris [Bodo Moeller] 686472613Skris 686572613Skris *) Fix potential buffer overrun for EBCDIC. 686672613Skris [Ulf Moeller] 686772613Skris 686872613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 686972613Skris keyUsage if basicConstraints absent for a CA. 687072613Skris [Steve Henson] 687172613Skris 687272613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 687372613Skris is more generally accepted (no spaces before the semicolon), since 687472613Skris some programs can't parse those values properly otherwise. Also make 687572613Skris sure BIO's that break lines after each write do not create invalid 687672613Skris headers. 687772613Skris [Richard Levitte] 687872613Skris 687972613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 688072613Skris macros previously used would not encode an empty SEQUENCE OF 688172613Skris and break the signature. 688272613Skris [Steve Henson] 688376866Skris [This change does not apply to 0.9.7.] 688472613Skris 688572613Skris *) Zero the premaster secret after deriving the master secret in 688672613Skris DH ciphersuites. 688772613Skris [Steve Henson] 688872613Skris 688972613Skris *) Add some EVP_add_digest_alias registrations (as found in 689072613Skris OpenSSL_add_all_digests()) to SSL_library_init() 689172613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 689272613Skris compatibility with peers using X.509 certificates 689372613Skris with unconventional AlgorithmIdentifier OIDs. 689472613Skris [Bodo Moeller] 689572613Skris 689672613Skris *) Fix for Irix with NO_ASM. 689772613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 689872613Skris 689972613Skris *) ./config script fixes. 690072613Skris [Ulf Moeller, Richard Levitte] 690172613Skris 690272613Skris *) Fix 'openssl passwd -1'. 690372613Skris [Bodo Moeller] 690472613Skris 690572613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 690672613Skris terminated strings whose length is passed in the passlen 690772613Skris parameter, for example from PEM callbacks. This was done 690872613Skris by adding an extra length parameter to asc2uni(). 690972613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 691072613Skris 691172613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 691272613Skris call failed, free the DSA structure. 691372613Skris [Bodo Moeller] 691472613Skris 691572613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 691672613Skris These are present in some PKCS#12 files. 691772613Skris [Steve Henson] 691872613Skris 691972613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 692072613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 692172613Skris when writing a 32767 byte record. 692272613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 692372613Skris 692472613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 692572613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 692672613Skris 692772613Skris (RSA objects have a reference count access to which is protected 692872613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 692972613Skris so they are meant to be shared between threads.) 693072613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 693172613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 693272613Skris 693372613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 693472613Skris [Bodo Moeller] 693572613Skris 693676866Skris *) Use better test patterns in bntest. 693776866Skris [Ulf M�ller] 693876866Skris 693972613Skris *) rand_win.c fix for Borland C. 694072613Skris [Ulf M�ller] 694172613Skris 694272613Skris *) BN_rshift bugfix for n == 0. 694372613Skris [Bodo Moeller] 694472613Skris 694576866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 694676866Skris so that 'make test' does not abort just because 'bc' is broken. 694776866Skris [Bodo Moeller] 694876866Skris 694972613Skris *) Store verify_result within SSL_SESSION also for client side to 695072613Skris avoid potential security hole. (Re-used sessions on the client side 695172613Skris always resulted in verify_result==X509_V_OK, not using the original 695272613Skris result of the server certificate verification.) 695372613Skris [Lutz Jaenicke] 695472613Skris 695572613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 695672613Skris SSL3_RT_APPLICATION_DATA, return 0. 695772613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 695872613Skris [Bodo Moeller] 695972613Skris 696072613Skris *) Fix SSL_peek: 696172613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 696272613Skris releases, have been re-implemented by renaming the previous 696372613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 696472613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 696572613Skris to them. The new ssl[23]_{read,peek} functions are calls to 696672613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 696772613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 696872613Skris does the actual work for ssl3_read_internal. 696972613Skris [Bodo Moeller] 697072613Skris 697176866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 697276866Skris the method-specific "init()" handler. Also clean up ex_data after 697376866Skris calling the method-specific "finish()" handler. Previously, this was 697476866Skris happening the other way round. 697576866Skris [Geoff Thorpe] 697676866Skris 697772613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 697872613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 697972613Skris [Bodo Moeller] 698072613Skris 698176866Skris *) Make sure that shared libraries get the internal name engine with 698276866Skris the full version number and not just 0. This should mark the 698376866Skris shared libraries as not backward compatible. Of course, this should 698476866Skris be changed again when we can guarantee backward binary compatibility. 698576866Skris [Richard Levitte] 698676866Skris 698772613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 698872613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 698972613Skris 699076866Skris *) Rework the system to generate shared libraries: 699176866Skris 699276866Skris - Make note of the expected extension for the shared libraries and 699376866Skris if there is a need for symbolic links from for example libcrypto.so.0 699476866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 699576866Skris that. 699676866Skris 699776866Skris - Make as few rebuilds of the shared libraries as possible. 699876866Skris 699976866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 700076866Skris 700176866Skris - When installing, install the shared libraries separately from the 700276866Skris static ones. 700376866Skris [Richard Levitte] 700476866Skris 700572613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 700672613Skris 700772613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 700872613Skris and not in SSL_clear because the latter is also used by the 700972613Skris accept/connect functions; previously, the settings made by 701072613Skris SSL_set_read_ahead would be lost during the handshake. 701172613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 701272613Skris 701372613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 701472613Skris Previously, it would create entries for disableed algorithms no 701572613Skris matter what. 701672613Skris [Richard Levitte] 701772613Skris 701872613Skris *) Added several new manual pages for SSL_* function. 701972613Skris [Lutz Jaenicke] 702072613Skris 702168651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 702268651Skris 702368651Skris *) In ssl23_get_client_hello, generate an error message when faced 702468651Skris with an initial SSL 3.0/TLS record that is too small to contain the 702568651Skris first two bytes of the ClientHello message, i.e. client_version. 702668651Skris (Note that this is a pathologic case that probably has never happened 702768651Skris in real life.) The previous approach was to use the version number 702868651Skris from the record header as a substitute; but our protocol choice 702968651Skris should not depend on that one because it is not authenticated 703068651Skris by the Finished messages. 703168651Skris [Bodo Moeller] 703268651Skris 703368651Skris *) More robust randomness gathering functions for Windows. 703468651Skris [Jeffrey Altman <jaltman@columbia.edu>] 703568651Skris 703668651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 703768651Skris not set then we don't setup the error code for issuer check errors 703868651Skris to avoid possibly overwriting other errors which the callback does 703968651Skris handle. If an application does set the flag then we assume it knows 704068651Skris what it is doing and can handle the new informational codes 704168651Skris appropriately. 704268651Skris [Steve Henson] 704368651Skris 704468651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 704568651Skris a general "ANY" type, as such it should be able to decode anything 704668651Skris including tagged types. However it didn't check the class so it would 704768651Skris wrongly interpret tagged types in the same way as their universal 704868651Skris counterpart and unknown types were just rejected. Changed so that the 704968651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 705068651Skris that is the encoding is stored intact. There is also a new type 705168651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 705268651Skris case we have no idea what the actual type is so we just lump them all 705368651Skris together. 705468651Skris [Steve Henson] 705568651Skris 705668651Skris *) On VMS, stdout may very well lead to a file that is written to 705768651Skris in a record-oriented fashion. That means that every write() will 705868651Skris write a separate record, which will be read separately by the 705968651Skris programs trying to read from it. This can be very confusing. 706068651Skris 706168651Skris The solution is to put a BIO filter in the way that will buffer 706268651Skris text until a linefeed is reached, and then write everything a 706368651Skris line at a time, so every record written will be an actual line, 706468651Skris not chunks of lines and not (usually doesn't happen, but I've 706568651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 706668651Skris the answer. 706768651Skris 706868651Skris Currently, it's a VMS-only method, because that's where it has 706968651Skris been tested well enough. 707068651Skris [Richard Levitte] 707168651Skris 707268651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 707368651Skris it can return incorrect results. 707468651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 707568651Skris but it was in 0.9.6-beta[12].) 707668651Skris [Bodo Moeller] 707768651Skris 707868651Skris *) Disable the check for content being present when verifying detached 707968651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 708068651Skris include zero length content when signing messages. 708168651Skris [Steve Henson] 708268651Skris 708368651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 708468651Skris BIO_ctrl (for BIO pairs). 708568651Skris [Bodo M�ller] 708668651Skris 708768651Skris *) Add DSO method for VMS. 708868651Skris [Richard Levitte] 708968651Skris 709068651Skris *) Bug fix: Montgomery multiplication could produce results with the 709168651Skris wrong sign. 709268651Skris [Ulf M�ller] 709368651Skris 709468651Skris *) Add RPM specification openssl.spec and modify it to build three 709568651Skris packages. The default package contains applications, application 709668651Skris documentation and run-time libraries. The devel package contains 709768651Skris include files, static libraries and function documentation. The 709868651Skris doc package contains the contents of the doc directory. The original 709968651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 710068651Skris [Richard Levitte] 710168651Skris 710268651Skris *) Add a large number of documentation files for many SSL routines. 710368651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 710468651Skris 710568651Skris *) Add a configuration entry for Sony News 4. 710668651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 710768651Skris 710868651Skris *) Don't set the two most significant bits to one when generating a 710968651Skris random number < q in the DSA library. 711068651Skris [Ulf M�ller] 711168651Skris 711268651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 711368651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 711468651Skris the underlying transport is blocking) if a handshake took place. 711568651Skris (The default behaviour is needed by applications such as s_client 711668651Skris and s_server that use select() to determine when to use SSL_read; 711768651Skris but for applications that know in advance when to expect data, it 711868651Skris just makes things more complicated.) 711968651Skris [Bodo Moeller] 712068651Skris 712168651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 712268651Skris from EGD. 712368651Skris [Ben Laurie] 712468651Skris 712568651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 712668651Skris work better on such systems. 712768651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 712868651Skris 712968651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 713068651Skris Update PKCS12_parse() so it copies the friendlyName and the 713168651Skris keyid to the certificates aux info. 713268651Skris [Steve Henson] 713368651Skris 713468651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 713568651Skris if there was more than one signature. 713668651Skris [Sven Uszpelkat <su@celocom.de>] 713768651Skris 713868651Skris *) Major change in util/mkdef.pl to include extra information 713968651Skris about each symbol, as well as presentig variables as well 714068651Skris as functions. This change means that there's n more need 714168651Skris to rebuild the .num files when some algorithms are excluded. 714268651Skris [Richard Levitte] 714368651Skris 714468651Skris *) Allow the verify time to be set by an application, 714568651Skris rather than always using the current time. 714668651Skris [Steve Henson] 714768651Skris 714868651Skris *) Phase 2 verify code reorganisation. The certificate 714968651Skris verify code now looks up an issuer certificate by a 715068651Skris number of criteria: subject name, authority key id 715168651Skris and key usage. It also verifies self signed certificates 715268651Skris by the same criteria. The main comparison function is 715368651Skris X509_check_issued() which performs these checks. 715468651Skris 715568651Skris Lot of changes were necessary in order to support this 715668651Skris without completely rewriting the lookup code. 715768651Skris 715868651Skris Authority and subject key identifier are now cached. 715968651Skris 716068651Skris The LHASH 'certs' is X509_STORE has now been replaced 716168651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 716268651Skris LHASH can't store or retrieve multiple objects with 716368651Skris the same hash value. 716468651Skris 716568651Skris As a result various functions (which were all internal 716668651Skris use only) have changed to handle the new X509_STORE 716768651Skris structure. This will break anything that messed round 716868651Skris with X509_STORE internally. 716968651Skris 717068651Skris The functions X509_STORE_add_cert() now checks for an 717168651Skris exact match, rather than just subject name. 717268651Skris 717368651Skris The X509_STORE API doesn't directly support the retrieval 717468651Skris of multiple certificates matching a given criteria, however 717568651Skris this can be worked round by performing a lookup first 717668651Skris (which will fill the cache with candidate certificates) 717768651Skris and then examining the cache for matches. This is probably 717868651Skris the best we can do without throwing out X509_LOOKUP 717968651Skris entirely (maybe later...). 718068651Skris 718168651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 718268651Skris 718368651Skris All certificate lookup operations now go via a get_issuer() 718468651Skris callback. Although this currently uses an X509_STORE it 718568651Skris can be replaced by custom lookups. This is a simple way 718668651Skris to bypass the X509_STORE hackery necessary to make this 718768651Skris work and makes it possible to use more efficient techniques 718868651Skris in future. A very simple version which uses a simple 718968651Skris STACK for its trusted certificate store is also provided 719068651Skris using X509_STORE_CTX_trusted_stack(). 719168651Skris 719268651Skris The verify_cb() and verify() callbacks now have equivalents 719368651Skris in the X509_STORE_CTX structure. 719468651Skris 719568651Skris X509_STORE_CTX also has a 'flags' field which can be used 719668651Skris to customise the verify behaviour. 719768651Skris [Steve Henson] 719868651Skris 719968651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 720068651Skris excludes S/MIME capabilities. 720168651Skris [Steve Henson] 720268651Skris 720368651Skris *) When a certificate request is read in keep a copy of the 720468651Skris original encoding of the signed data and use it when outputing 720568651Skris again. Signatures then use the original encoding rather than 720668651Skris a decoded, encoded version which may cause problems if the 720768651Skris request is improperly encoded. 720868651Skris [Steve Henson] 720968651Skris 721068651Skris *) For consistency with other BIO_puts implementations, call 721168651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 721268651Skris BIO_write(b, ...). 721368651Skris 721468651Skris In BIO_puts, increment b->num_write as in BIO_write. 721568651Skris [Peter.Sylvester@EdelWeb.fr] 721668651Skris 721768651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 721868651Skris BN_zero, we may not return a BIGNUM with an array consisting of 721968651Skris words set to zero.) 722068651Skris [Bodo Moeller] 722168651Skris 722268651Skris *) Avoid calling abort() from within the library when problems are 722368651Skris detected, except if preprocessor symbols have been defined 722468651Skris (such as REF_CHECK, BN_DEBUG etc.). 722568651Skris [Bodo Moeller] 722668651Skris 722768651Skris *) New openssl application 'rsautl'. This utility can be 722868651Skris used for low level RSA operations. DER public key 722968651Skris BIO/fp routines also added. 723068651Skris [Steve Henson] 723168651Skris 723268651Skris *) New Configure entry and patches for compiling on QNX 4. 723368651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 723468651Skris 723568651Skris *) A demo state-machine implementation was sponsored by 723668651Skris Nuron (http://www.nuron.com/) and is now available in 723768651Skris demos/state_machine. 723868651Skris [Ben Laurie] 723968651Skris 724068651Skris *) New options added to the 'dgst' utility for signature 724168651Skris generation and verification. 724268651Skris [Steve Henson] 724368651Skris 724468651Skris *) Unrecognized PKCS#7 content types are now handled via a 724568651Skris catch all ASN1_TYPE structure. This allows unsupported 724668651Skris types to be stored as a "blob" and an application can 724768651Skris encode and decode it manually. 724868651Skris [Steve Henson] 724968651Skris 725068651Skris *) Fix various signed/unsigned issues to make a_strex.c 725168651Skris compile under VC++. 725268651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 725368651Skris 725468651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 725568651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 725668651Skris if passed a NULL BN and its argument was negative. 725768651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 725868651Skris 725968651Skris *) Modification to PKCS#7 encoding routines to output definite 726068651Skris length encoding. Since currently the whole structures are in 726168651Skris memory there's not real point in using indefinite length 726268651Skris constructed encoding. However if OpenSSL is compiled with 726368651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 726468651Skris [Steve Henson] 726568651Skris 726668651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 726768651Skris [Richard Levitte] 726868651Skris 726968651Skris *) Added more prefixes to parse for in the the strings written 727068651Skris through a logging bio, to cover all the levels that are available 727168651Skris through syslog. The prefixes are now: 727268651Skris 727368651Skris PANIC, EMERG, EMR => LOG_EMERG 727468651Skris ALERT, ALR => LOG_ALERT 727568651Skris CRIT, CRI => LOG_CRIT 727668651Skris ERROR, ERR => LOG_ERR 727768651Skris WARNING, WARN, WAR => LOG_WARNING 727868651Skris NOTICE, NOTE, NOT => LOG_NOTICE 727968651Skris INFO, INF => LOG_INFO 728068651Skris DEBUG, DBG => LOG_DEBUG 728168651Skris 728268651Skris and as before, if none of those prefixes are present at the 728368651Skris beginning of the string, LOG_ERR is chosen. 728468651Skris 728568651Skris On Win32, the LOG_* levels are mapped according to this: 728668651Skris 728768651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 728868651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 728968651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 729068651Skris 729168651Skris [Richard Levitte] 729268651Skris 729368651Skris *) Made it possible to reconfigure with just the configuration 729468651Skris argument "reconf" or "reconfigure". The command line arguments 729568651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 729668651Skris and are retrieved from there when reconfiguring. 729768651Skris [Richard Levitte] 729868651Skris 729968651Skris *) MD4 implemented. 730068651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 730168651Skris 730268651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 730368651Skris [Richard Levitte] 730468651Skris 730568651Skris *) The obj_dat.pl script was messing up the sorting of object 730668651Skris names. The reason was that it compared the quoted version 730768651Skris of strings as a result "OCSP" > "OCSP Signing" because 730868651Skris " > SPACE. Changed script to store unquoted versions of 730968651Skris names and add quotes on output. It was also omitting some 731068651Skris names from the lookup table if they were given a default 731168651Skris value (that is if SN is missing it is given the same 731268651Skris value as LN and vice versa), these are now added on the 731368651Skris grounds that if an object has a name we should be able to 731468651Skris look it up. Finally added warning output when duplicate 731568651Skris short or long names are found. 731668651Skris [Steve Henson] 731768651Skris 731868651Skris *) Changes needed for Tandem NSK. 731968651Skris [Scott Uroff <scott@xypro.com>] 732068651Skris 732168651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 732268651Skris RSA_padding_check_SSLv23(), special padding was never detected 732368651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 732468651Skris version rollback attacks was not effective. 732568651Skris 732668651Skris In s23_clnt.c, don't use special rollback-attack detection padding 732768651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 732868651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 732968651Skris SSL 2.0 is the only protocol enabled in the server. 733068651Skris [Bodo Moeller] 733168651Skris 733268651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 733368651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 733468651Skris BIO_dump_indent() are added. 733568651Skris [Richard Levitte] 733668651Skris 733768651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 733868651Skris these print out strings and name structures based on various 733968651Skris flags including RFC2253 support and proper handling of 734068651Skris multibyte characters. Added options to the 'x509' utility 734168651Skris to allow the various flags to be set. 734268651Skris [Steve Henson] 734368651Skris 734468651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 734568651Skris Also change the functions X509_cmp_current_time() and 734668651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 734768651Skris this will enable certificates using GeneralizedTime in validity 734868651Skris dates to be checked. 734968651Skris [Steve Henson] 735068651Skris 735168651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 735268651Skris negative public key encodings) on by default, 735368651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 735468651Skris [Steve Henson] 735568651Skris 735668651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 735768651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 735868651Skris the encoding can be trivially obtained from the structure. 735968651Skris [Steve Henson] 736068651Skris 736168651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 736268651Skris not read locks (CRYPTO_r_[un]lock). 736368651Skris [Bodo Moeller] 736468651Skris 736568651Skris *) A first attempt at creating official support for shared 736668651Skris libraries through configuration. I've kept it so the 736768651Skris default is static libraries only, and the OpenSSL programs 736868651Skris are always statically linked for now, but there are 736968651Skris preparations for dynamic linking in place. 737089837Skris This has been tested on Linux and Tru64. 737168651Skris [Richard Levitte] 737268651Skris 737368651Skris *) Randomness polling function for Win9x, as described in: 737468651Skris Peter Gutmann, Software Generation of Practically Strong 737568651Skris Random Numbers. 737668651Skris [Ulf M�ller] 737768651Skris 737868651Skris *) Fix so PRNG is seeded in req if using an already existing 737968651Skris DSA key. 738068651Skris [Steve Henson] 738168651Skris 738268651Skris *) New options to smime application. -inform and -outform 738368651Skris allow alternative formats for the S/MIME message including 738468651Skris PEM and DER. The -content option allows the content to be 738568651Skris specified separately. This should allow things like Netscape 738668651Skris form signing output easier to verify. 738768651Skris [Steve Henson] 738868651Skris 738968651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 739068651Skris [Steve Henson] 739168651Skris 739268651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 739368651Skris STRING types. These convert content octets to and from the 739468651Skris underlying type. The actual tag and length octets are 739568651Skris already assumed to have been read in and checked. These 739668651Skris are needed because all other string types have virtually 739768651Skris identical handling apart from the tag. By having versions 739868651Skris of the ASN1 functions that just operate on content octets 739968651Skris IMPLICIT tagging can be handled properly. It also allows 740068651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 740168651Skris and ASN1_INTEGER are identical apart from the tag. 740268651Skris [Steve Henson] 740368651Skris 740468651Skris *) Change the handling of OID objects as follows: 740568651Skris 740668651Skris - New object identifiers are inserted in objects.txt, following 740768651Skris the syntax given in objects.README. 740868651Skris - objects.pl is used to process obj_mac.num and create a new 740968651Skris obj_mac.h. 741068651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 741168651Skris obj_mac.h. 741268651Skris 741368651Skris This is currently kind of a hack, and the perl code in objects.pl 741468651Skris isn't very elegant, but it works as I intended. The simplest way 741568651Skris to check that it worked correctly is to look in obj_dat.h and 741668651Skris check the array nid_objs and make sure the objects haven't moved 741768651Skris around (this is important!). Additions are OK, as well as 741868651Skris consistent name changes. 741968651Skris [Richard Levitte] 742068651Skris 742168651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 742268651Skris [Bodo Moeller] 742368651Skris 742468651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 742568651Skris The given file adds to whatever has already been seeded into the 742668651Skris random pool through the RANDFILE configuration file option or 742768651Skris environment variable, or the default random state file. 742868651Skris [Richard Levitte] 742968651Skris 743068651Skris *) mkstack.pl now sorts each macro group into lexical order. 743168651Skris Previously the output order depended on the order the files 743268651Skris appeared in the directory, resulting in needless rewriting 743368651Skris of safestack.h . 743468651Skris [Steve Henson] 743568651Skris 743668651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 743768651Skris work arounds for the VC++ problem that it treats func() as 743868651Skris func(void). Also stripped out the parts of mkdef.pl that 743968651Skris added extra typesafe functions: these no longer exist. 744068651Skris [Steve Henson] 744168651Skris 744268651Skris *) Reorganisation of the stack code. The macros are now all 744368651Skris collected in safestack.h . Each macro is defined in terms of 744468651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 744568651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 744668651Skris this has the advantage of retaining type safety without the 744768651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 744868651Skris then the non typesafe macros are used instead. Also modified the 744968651Skris mkstack.pl script to handle the new form. Needs testing to see 745068651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 745168651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 745268651Skris and PKCS12_STACK_OF. 745368651Skris [Steve Henson] 745468651Skris 745568651Skris *) When some versions of IIS use the 'NET' form of private key the 745668651Skris key derivation algorithm is different. Normally MD5(password) is 745768651Skris used as a 128 bit RC4 key. In the modified case 745868651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 745968651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 746068651Skris as the old Netscape_RSA functions except they have an additional 746168651Skris 'sgckey' parameter which uses the modified algorithm. Also added 746268651Skris an -sgckey command line option to the rsa utility. Thanks to 746368651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 746468651Skris algorithm to openssl-dev. 746568651Skris [Steve Henson] 746668651Skris 746768651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 746868651Skris invalid expansion on some systems (SCO 5.0.5 for example). 746968651Skris Corrected to 'c.kname'. 747068651Skris [Phillip Porch <root@theporch.com>] 747168651Skris 747268651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 747368651Skris a STACK of email addresses from a certificate or request, these look 747468651Skris in the subject name and the subject alternative name extensions and 747568651Skris omit any duplicate addresses. 747668651Skris [Steve Henson] 747768651Skris 747868651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 747968651Skris This makes DSA verification about 2 % faster. 748068651Skris [Bodo Moeller] 748168651Skris 748268651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 748368651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 748468651Skris plus overhead for 1024 bit moduli). 748568651Skris This makes exponentiations about 0.5 % faster for 1024 bit 748668651Skris exponents (as measured by "openssl speed rsa2048"). 748768651Skris [Bodo Moeller] 748868651Skris 748968651Skris *) Rename memory handling macros to avoid conflicts with other 749068651Skris software: 749168651Skris Malloc => OPENSSL_malloc 749268651Skris Malloc_locked => OPENSSL_malloc_locked 749368651Skris Realloc => OPENSSL_realloc 749468651Skris Free => OPENSSL_free 749568651Skris [Richard Levitte] 749668651Skris 749768651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 749868651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 749968651Skris [Bodo Moeller] 750068651Skris 750168651Skris *) CygWin32 support. 750268651Skris [John Jarvie <jjarvie@newsguy.com>] 750368651Skris 750468651Skris *) The type-safe stack code has been rejigged. It is now only compiled 750568651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 750668651Skris by default all type-specific stack functions are "#define"d back to 750768651Skris standard stack functions. This results in more streamlined output 750868651Skris but retains the type-safety checking possibilities of the original 750968651Skris approach. 751068651Skris [Geoff Thorpe] 751168651Skris 751268651Skris *) The STACK code has been cleaned up, and certain type declarations 751368651Skris that didn't make a lot of sense have been brought in line. This has 751468651Skris also involved a cleanup of sorts in safestack.h to more correctly 751568651Skris map type-safe stack functions onto their plain stack counterparts. 751668651Skris This work has also resulted in a variety of "const"ifications of 751768651Skris lots of the code, especially "_cmp" operations which should normally 751868651Skris be prototyped with "const" parameters anyway. 751968651Skris [Geoff Thorpe] 752068651Skris 752168651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 752268651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 752368651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 752468651Skris where all of 'md' is used each time the PRNG is used, but 'state' 752568651Skris is used only indexed by a cyclic counter. As entropy may not be 752668651Skris well distributed from the beginning, 'md' is important as a 752768651Skris chaining variable. However, the output function chains only half 752868651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 752968651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 753068651Skris in all of 'state' being rewritten, with the new values depending 753168651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 753268651Skris [Bodo Moeller] 753368651Skris 753468651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 753568651Skris the handshake is continued after ssl_verify_cert_chain(); 753668651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 753768651Skris can lead to 'unexplainable' connection aborts later. 753868651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 753968651Skris 754068651Skris *) Major EVP API cipher revision. 754168651Skris Add hooks for extra EVP features. This allows various cipher 754268651Skris parameters to be set in the EVP interface. Support added for variable 754368651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 754468651Skris setting of RC2 and RC5 parameters. 754568651Skris 754668651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 754768651Skris ciphers. 754868651Skris 754968651Skris Remove lots of duplicated code from the EVP library. For example *every* 755068651Skris cipher init() function handles the 'iv' in the same way according to the 755168651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 755268651Skris for CFB and OFB modes they zero ctx->num. 755368651Skris 755468651Skris New functionality allows removal of S/MIME code RC2 hack. 755568651Skris 755668651Skris Most of the routines have the same form and so can be declared in terms 755768651Skris of macros. 755868651Skris 755968651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 756068651Skris all individual ciphers. If the cipher wants to handle IVs or keys 756168651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 756268651Skris flags. 756368651Skris 756468651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 756568651Skris value: although software versions of the algorithms cannot fail 756668651Skris any installed hardware versions can. 756768651Skris [Steve Henson] 756868651Skris 756968651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 757068651Skris this option is set, tolerate broken clients that send the negotiated 757168651Skris protocol version number instead of the requested protocol version 757268651Skris number. 757368651Skris [Bodo Moeller] 757468651Skris 757568651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 757668651Skris i.e. non-zero for export ciphersuites, zero otherwise. 757768651Skris Previous versions had this flag inverted, inconsistent with 757868651Skris rsa_tmp_cb (..._TMP_RSA_CB). 757968651Skris [Bodo Moeller; problem reported by Amit Chopra] 758068651Skris 758168651Skris *) Add missing DSA library text string. Work around for some IIS 758268651Skris key files with invalid SEQUENCE encoding. 758368651Skris [Steve Henson] 758468651Skris 758568651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 758668651Skris and so on that are implemented in OpenSSL. 758768651Skris [Richard Levitte] 758868651Skris 758968651Skris *) Enhance c_rehash script. Old version would mishandle certificates 759068651Skris with the same subject name hash and wouldn't handle CRLs at all. 759168651Skris Added -fingerprint option to crl utility, to support new c_rehash 759268651Skris features. 759368651Skris [Steve Henson] 759468651Skris 759568651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 759668651Skris [Ulf M�ller] 759768651Skris 759868651Skris *) Fix for SSL server purpose checking. Server checking was 759968651Skris rejecting certificates which had extended key usage present 760068651Skris but no ssl client purpose. 760168651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 760268651Skris 760368651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 760468651Skris is a little unclear about how a blank password is handled. 760568651Skris Since the password in encoded as a BMPString with terminating 760668651Skris double NULL a zero length password would end up as just the 760768651Skris double NULL. However no password at all is different and is 760868651Skris handled differently in the PKCS#12 key generation code. NS 760968651Skris treats a blank password as zero length. MSIE treats it as no 761068651Skris password on export: but it will try both on import. We now do 761168651Skris the same: PKCS12_parse() tries zero length and no password if 761268651Skris the password is set to "" or NULL (NULL is now a valid password: 761368651Skris it wasn't before) as does the pkcs12 application. 761468651Skris [Steve Henson] 761568651Skris 761668651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 761768651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 761868651Skris be obtained from the error queue. 761968651Skris [Bodo Moeller] 762068651Skris 762168651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 762268651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 762368651Skris accordingly to avoid race conditions (this is necessary because 762468651Skris thread_hash is no longer constant once set). 762568651Skris [Bodo Moeller] 762668651Skris 762768651Skris *) Bugfix for linux-elf makefile.one. 762868651Skris [Ulf M�ller] 762968651Skris 763068651Skris *) RSA_get_default_method() will now cause a default 763168651Skris RSA_METHOD to be chosen if one doesn't exist already. 763268651Skris Previously this was only set during a call to RSA_new() 763368651Skris or RSA_new_method(NULL) meaning it was possible for 763468651Skris RSA_get_default_method() to return NULL. 763568651Skris [Geoff Thorpe] 763668651Skris 763768651Skris *) Added native name translation to the existing DSO code 763868651Skris that will convert (if the flag to do so is set) filenames 763968651Skris that are sufficiently small and have no path information 764068651Skris into a canonical native form. Eg. "blah" converted to 764168651Skris "libblah.so" or "blah.dll" etc. 764268651Skris [Geoff Thorpe] 764368651Skris 764468651Skris *) New function ERR_error_string_n(e, buf, len) which is like 764568651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 764668651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 764768651Skris may not be NULL. 764868651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 764968651Skris 765068651Skris *) CONF library reworked to become more general. A new CONF 765168651Skris configuration file reader "class" is implemented as well as a 765268651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 765368651Skris old CONF_* functions are still there, but are reimplemented to 765468651Skris work in terms of the new functions. Also, a set of functions 765568651Skris to handle the internal storage of the configuration data is 765668651Skris provided to make it easier to write new configuration file 765768651Skris reader "classes" (I can definitely see something reading a 765868651Skris configuration file in XML format, for example), called _CONF_*, 765968651Skris or "the configuration storage API"... 766068651Skris 766168651Skris The new configuration file reading functions are: 766268651Skris 766368651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 766468651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 766568651Skris 766668651Skris NCONF_default, NCONF_WIN32 766768651Skris 766868651Skris NCONF_dump_fp, NCONF_dump_bio 766968651Skris 767068651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 767168651Skris NCONF_new creates a new CONF object. This works in the same way 767268651Skris as other interfaces in OpenSSL, like the BIO interface. 767368651Skris NCONF_dump_* dump the internal storage of the configuration file, 767468651Skris which is useful for debugging. All other functions take the same 767568651Skris arguments as the old CONF_* functions wth the exception of the 767668651Skris first that must be a `CONF *' instead of a `LHASH *'. 767768651Skris 767868651Skris To make it easer to use the new classes with the old CONF_* functions, 767968651Skris the function CONF_set_default_method is provided. 768068651Skris [Richard Levitte] 768168651Skris 768268651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 768368651Skris mentioned in the documentation but had not been implemented. 768468651Skris (This option is not yet really useful because even the additional 768568651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 768668651Skris [Bodo Moeller] 768768651Skris 768868651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 768968651Skris OpenSSL-based applications) load shared libraries and bind to 769068651Skris them in a portable way. 769168651Skris [Geoff Thorpe, with contributions from Richard Levitte] 769268651Skris 769359191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 769459191Skris 769559191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 769659191Skris 769759191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 769859191Skris (the default implementation of RAND_status). 769959191Skris 770059191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 770159191Skris to '-clrext' (= clear extensions), as intended and documented. 770259191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 770359191Skris <attili@amaxo.com>] 770459191Skris 770559191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 770659191Skris was larger than the MD block size. 770759191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 770859191Skris 770959191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 771059191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 771159191Skris using the passed key: if the passed key was a private key the result 771259191Skris of X509_print(), for example, would be to print out all the private key 771359191Skris components. 771459191Skris [Steve Henson] 771559191Skris 771659191Skris *) des_quad_cksum() byte order bug fix. 771759191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 771859191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 771959191Skris 772059191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 772159191Skris discouraged. 772259191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 772359191Skris 772459191Skris *) For easily testing in shell scripts whether some command 772559191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 772659191Skris returns with exit code 0 iff no command of the given name is available. 772759191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 772859191Skris the output goes to stdout and nothing is printed to stderr. 772959191Skris Additional arguments are always ignored. 773059191Skris 773159191Skris Since for each cipher there is a command of the same name, 773259191Skris the 'no-cipher' compilation switches can be tested this way. 773359191Skris 773459191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 773559191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 773659191Skris [Bodo Moeller] 773759191Skris 773859191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 773959191Skris [Bodo Moeller] 774059191Skris 774159191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 774259191Skris is set; it will be thrown away anyway because each handshake creates 774359191Skris its own key. 774459191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 774559191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 774659191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 774759191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 774859191Skris [Bodo Moeller] 774959191Skris 775059191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 775159191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 775259191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 775359191Skris does not suppress any output. 775459191Skris [Richard Levitte] 775559191Skris 775659191Skris *) Add compatibility options to the purpose and trust code. The 775759191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 775859191Skris accepts a certificate or CA, this was the previous behaviour, 775959191Skris with all the associated security issues. 776059191Skris 776159191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 776259191Skris automatically trust self signed roots in certificate store. A 776359191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 776459191Skris a purpose has no associated trust setting and it should instead 776559191Skris use the value in the default purpose. 776659191Skris [Steve Henson] 776759191Skris 776859191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 776959191Skris and fix a memory leak. 777059191Skris [Steve Henson] 777159191Skris 777259191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 777359191Skris reason strings from the previous version of the .c file, as 777459191Skris the default to have only downcase letters (and digits) in 777559191Skris automatically generated reasons codes is not always appropriate. 777659191Skris [Bodo Moeller] 777759191Skris 777859191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 777959191Skris using strerror. Previously, ERR_reason_error_string() returned 778059191Skris library names as reason strings for SYSerr; but SYSerr is a special 778159191Skris case where small numbers are errno values, not library numbers. 778259191Skris [Bodo Moeller] 778359191Skris 778459191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 778559191Skris converts DSA parameters into DH parameters. (When creating parameters, 778659191Skris DSA_generate_parameters is used.) 778759191Skris [Bodo Moeller] 778859191Skris 778959191Skris *) Include 'length' (recommended exponent length) in C code generated 779059191Skris by 'openssl dhparam -C'. 779159191Skris [Bodo Moeller] 779259191Skris 779359191Skris *) The second argument to set_label in perlasm was already being used 779459191Skris so couldn't be used as a "file scope" flag. Moved to third argument 779559191Skris which was free. 779659191Skris [Steve Henson] 779759191Skris 779859191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 779959191Skris instead of RAND_bytes for encryption IVs and salts. 780059191Skris [Bodo Moeller] 780159191Skris 780259191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 780359191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 780459191Skris RAND_set_rand_method would be impossible. 780559191Skris [Bodo Moeller] 780659191Skris 780759191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 780859191Skris number generation fails. 780959191Skris [Bodo Moeller] 781059191Skris 781159191Skris *) New 'rand' application for creating pseudo-random output. 781259191Skris [Bodo Moeller] 781359191Skris 781459191Skris *) Added configuration support for Linux/IA64 781559191Skris [Rolf Haberrecker <rolf@suse.de>] 781659191Skris 781759191Skris *) Assembler module support for Mingw32. 781859191Skris [Ulf M�ller] 781959191Skris 782059191Skris *) Shared library support for HPUX (in shlib/). 782159191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 782259191Skris 782359191Skris *) Shared library support for Solaris gcc. 782459191Skris [Lutz Behnke <behnke@trustcenter.de>] 782559191Skris 782659191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 782759191Skris 782859191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 782959191Skris were added manually and by SMIME_crlf_copy(). 783059191Skris [Steve Henson] 783159191Skris 783259191Skris *) In bntest.c don't call BN_rand with zero bits argument. 783359191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 783459191Skris 783559191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 783659191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 783759191Skris [Ulf M�ller] 783859191Skris 783959191Skris *) Add an optional second argument to the set_label() in the perl 784059191Skris assembly language builder. If this argument exists and is set 784159191Skris to 1 it signals that the assembler should use a symbol whose 784259191Skris scope is the entire file, not just the current function. This 784359191Skris is needed with MASM which uses the format label:: for this scope. 784459191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 784559191Skris 784659191Skris *) Change the ASN1 types so they are typedefs by default. Before 784759191Skris almost all types were #define'd to ASN1_STRING which was causing 784859191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 784959191Skris for example. 785059191Skris [Steve Henson] 785159191Skris 785259191Skris *) Change names of new functions to the new get1/get0 naming 785359191Skris convention: After 'get1', the caller owns a reference count 785459191Skris and has to call ..._free; 'get0' returns a pointer to some 785559191Skris data structure without incrementing reference counters. 785659191Skris (Some of the existing 'get' functions increment a reference 785759191Skris counter, some don't.) 785859191Skris Similarly, 'set1' and 'add1' functions increase reference 785959191Skris counters or duplicate objects. 786059191Skris [Steve Henson] 786159191Skris 786259191Skris *) Allow for the possibility of temp RSA key generation failure: 786359191Skris the code used to assume it always worked and crashed on failure. 786459191Skris [Steve Henson] 786559191Skris 786659191Skris *) Fix potential buffer overrun problem in BIO_printf(). 786759191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 786859191Skris pointed out by David Sacerdote <das33@cornell.edu>] 786959191Skris 787059191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 787159191Skris RAND_egd() and RAND_status(). In the command line application, 787259191Skris the EGD socket can be specified like a seed file using RANDFILE 787359191Skris or -rand. 787459191Skris [Ulf M�ller] 787559191Skris 787659191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 787759191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 787859191Skris [Steve Henson] 787959191Skris 788059191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 788159191Skris list to exclude them. This means that no special compilation option 788259191Skris is needed to use anonymous DH: it just needs to be included in the 788359191Skris cipher list. 788459191Skris [Steve Henson] 788559191Skris 788659191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 788759191Skris EVP_MD_type. The old functionality is available in a new macro called 788859191Skris EVP_MD_md(). Change code that uses it and update docs. 788959191Skris [Steve Henson] 789059191Skris 789159191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 789259191Skris where the 'void *' argument is replaced by a function pointer argument. 789359191Skris Previously 'void *' was abused to point to functions, which works on 789459191Skris many platforms, but is not correct. As these functions are usually 789559191Skris called by macros defined in OpenSSL header files, most source code 789659191Skris should work without changes. 789759191Skris [Richard Levitte] 789859191Skris 789959191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 790059191Skris sections with information on -D... compiler switches used for 790159191Skris compiling the library so that applications can see them. To enable 790259191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 790359191Skris must be defined. E.g., 790459191Skris #define OPENSSL_ALGORITHM_DEFINES 790559191Skris #include <openssl/opensslconf.h> 790659191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 790759191Skris [Richard Levitte, Ulf and Bodo M�ller] 790859191Skris 790959191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 791059191Skris record layer. 791159191Skris [Bodo Moeller] 791259191Skris 791359191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 791459191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 791559191Skris the required ASN1 format: arbitrary types determined by an OID. 791659191Skris [Steve Henson] 791759191Skris 791859191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 791959191Skris argument to 'req'. This is not because the function is newer or 792059191Skris better than others it just uses the work 'NEW' in the certificate 792159191Skris request header lines. Some software needs this. 792259191Skris [Steve Henson] 792359191Skris 792459191Skris *) Reorganise password command line arguments: now passwords can be 792559191Skris obtained from various sources. Delete the PEM_cb function and make 792659191Skris it the default behaviour: i.e. if the callback is NULL and the 792759191Skris usrdata argument is not NULL interpret it as a null terminated pass 792859191Skris phrase. If usrdata and the callback are NULL then the pass phrase 792959191Skris is prompted for as usual. 793059191Skris [Steve Henson] 793159191Skris 793259191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 793359191Skris the support is automatically enabled. The resulting binaries will 793459191Skris autodetect the card and use it if present. 793559191Skris [Ben Laurie and Compaq Inc.] 793659191Skris 793759191Skris *) Work around for Netscape hang bug. This sends certificate request 793859191Skris and server done in one record. Since this is perfectly legal in the 793959191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 794059191Skris the bugs/SSLv3 entry for more info. 794159191Skris [Steve Henson] 794259191Skris 794359191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 794459191Skris [Andy Polyakov] 794559191Skris 794659191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 794759191Skris of seed file. 794859191Skris [Steve Henson] 794959191Skris 795059191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 795159191Skris [Bodo Moeller] 795259191Skris 795359191Skris *) Add command line password options to the remaining applications. 795459191Skris [Steve Henson] 795559191Skris 795659191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 795759191Skris bits. 795859191Skris [Ulf M�ller] 795959191Skris 796059191Skris *) More tests in bntest.c, and changed test_bn output. 796159191Skris [Ulf M�ller] 796259191Skris 796359191Skris *) ./config recognizes MacOS X now. 796459191Skris [Andy Polyakov] 796559191Skris 796659191Skris *) Bug fix for BN_div() when the first words of num and divsor are 796759191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 796859191Skris [Ulf M�ller] 796959191Skris 797059191Skris *) Add support for various broken PKCS#8 formats, and command line 797159191Skris options to produce them. 797259191Skris [Steve Henson] 797359191Skris 797459191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 797559191Skris get temporary BIGNUMs from a BN_CTX. 797659191Skris [Ulf M�ller] 797759191Skris 797859191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 797959191Skris for p == 0. 798059191Skris [Ulf M�ller] 798159191Skris 798259191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 798359191Skris include a #define from the old name to the new. The original intent 798459191Skris was that statically linked binaries could for example just call 798559191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 798659191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 798759191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 798859191Skris one would link with the other. They are now in separate source files. 798959191Skris [Steve Henson] 799059191Skris 799159191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 799259191Skris [Steve Henson] 799359191Skris 799459191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 799559191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 799659191Skris loop, our standard modexp algorithms are faster). 799759191Skris [Bodo Moeller] 799859191Skris 799959191Skris *) Support for the EBCDIC character set completed. 800059191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 800159191Skris 800259191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 800359191Skris use void * instead of char * in lhash. 800459191Skris [Ulf M�ller] 800559191Skris 800659191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 800759191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 800859191Skris this the server could overwrite ephemeral keys that the client 800959191Skris has already seen). 801059191Skris [Bodo Moeller] 801159191Skris 801259191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 801359191Skris using 50 iterations of the Rabin-Miller test. 801459191Skris 801559191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 801659191Skris iterations of the Rabin-Miller test as required by the appendix 801759191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 801859191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 801959191Skris generation becomes much faster. 802059191Skris 802159191Skris This implies a change for the callback functions in DSA_is_prime 802259191Skris and DSA_generate_parameters: The callback function is called once 802359191Skris for each positive witness in the Rabin-Miller test, not just 802459191Skris occasionally in the inner loop; and the parameters to the 802559191Skris callback function now provide an iteration count for the outer 802659191Skris loop rather than for the current invocation of the inner loop. 802759191Skris DSA_generate_parameters additionally can call the callback 802859191Skris function with an 'iteration count' of -1, meaning that a 802959191Skris candidate has passed the trial division test (when q is generated 803059191Skris from an application-provided seed, trial division is skipped). 803159191Skris [Bodo Moeller] 803259191Skris 803359191Skris *) New function BN_is_prime_fasttest that optionally does trial 803459191Skris division before starting the Rabin-Miller test and has 803559191Skris an additional BN_CTX * argument (whereas BN_is_prime always 803659191Skris has to allocate at least one BN_CTX). 803759191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 803859191Skris trial division stage. 803959191Skris [Bodo Moeller] 804059191Skris 804159191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 804259191Skris as ASN1_TIME. 804359191Skris [Steve Henson] 804459191Skris 804559191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 804659191Skris [Steve Henson] 804759191Skris 804859191Skris *) New function BN_pseudo_rand(). 804959191Skris [Ulf M�ller] 805059191Skris 805159191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 805259191Skris bignum version of BN_from_montgomery() with the working code from 805359191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 805459191Skris the comments. 805559191Skris [Ulf M�ller] 805659191Skris 805759191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 805859191Skris made it impossible to use the same SSL_SESSION data structure in 805959191Skris SSL2 clients in multiple threads. 806059191Skris [Bodo Moeller] 806159191Skris 806259191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 806359191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 806459191Skris to seed the PRNG (previously an explicit byte count was required). 806559191Skris [Ulf M�ller, Bodo M�ller] 806659191Skris 806759191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 806859191Skris used (char *) instead of (void *) and had casts all over the place. 806959191Skris [Steve Henson] 807059191Skris 807159191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 807259191Skris [Ulf M�ller] 807359191Skris 807459191Skris *) Retain source code compatibility for BN_prime_checks macro: 807559191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 807659191Skris BN_prime_checks_for_size to determine the appropriate number of 807759191Skris Rabin-Miller iterations. 807859191Skris [Ulf M�ller] 807959191Skris 808059191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 808159191Skris DH_CHECK_P_NOT_SAFE_PRIME. 808259191Skris (Check if this is true? OpenPGP calls them "strong".) 808359191Skris [Ulf M�ller] 808459191Skris 808559191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 808659191Skris "dhparam". The old programs are retained for now but will handle DH keys 808759191Skris (instead of parameters) in future. 808859191Skris [Steve Henson] 808959191Skris 809059191Skris *) Make the ciphers, s_server and s_client programs check the return values 809159191Skris when a new cipher list is set. 809259191Skris [Steve Henson] 809359191Skris 809459191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 809559191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 809659191Skris wrong. 809759191Skris 809859191Skris The syntax for the cipher sorting has been extended to support sorting by 809959191Skris cipher-strength (using the strength_bits hard coded in the tables). 810059191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 810159191Skris 810259191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 810359191Skris string with an "undefined" symbol (neither command nor alphanumeric 810459191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 810559191Skris an error is flagged. 810659191Skris 810759191Skris Due to the strength-sorting extension, the code of the 810859191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 810959191Skris the readability was also increased :-) 811059191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 811159191Skris 811259191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 811359191Skris for the first serial number and places 2 in the serial number file. This 811459191Skris avoids problems when the root CA is created with serial number zero and 811559191Skris the first user certificate has the same issuer name and serial number 811659191Skris as the root CA. 811759191Skris [Steve Henson] 811859191Skris 811959191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 812059191Skris the new code. Add documentation for this stuff. 812159191Skris [Steve Henson] 812259191Skris 812359191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 812459191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 812559191Skris structures and behave in an analagous way to the X509v3 functions: 812659191Skris they shouldn't be called directly but wrapper functions should be used 812759191Skris instead. 812859191Skris 812959191Skris So we also now have some wrapper functions that call the X509at functions 813059191Skris when passed certificate requests. (TO DO: similar things can be done with 813159191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 813259191Skris things. Some of these need some d2i or i2d and print functionality 813359191Skris because they handle more complex structures.) 813459191Skris [Steve Henson] 813559191Skris 813659191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 813759191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 813859191Skris NO_RSA in ssl/s2*.c. 813959191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 814059191Skris 814159191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 814259191Skris has a return value which indicates the quality of the random data 814359191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 814459191Skris error queue. New function RAND_pseudo_bytes() generates output that is 814559191Skris guaranteed to be unique but not unpredictable. RAND_add is like 814659191Skris RAND_seed, but takes an extra argument for an entropy estimate 814759191Skris (RAND_seed always assumes full entropy). 814859191Skris [Ulf M�ller] 814959191Skris 815059191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 815159191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 815259191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 815359191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 815459191Skris false-positive rate of at most 2^-80 for random input. 815559191Skris [Bodo Moeller] 815659191Skris 815759191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 815859191Skris [Bodo Moeller] 815959191Skris 816059191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 816159191Skris in the 0.9.5 release), this returns the chain 816259191Skris from an X509_CTX structure with a dup of the stack and all 816359191Skris the X509 reference counts upped: so the stack will exist 816459191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 816559191Skris to use this. 816659191Skris 816759191Skris Also make SSL_SESSION_print() print out the verify return 816859191Skris code. 816959191Skris [Steve Henson] 817059191Skris 817159191Skris *) Add manpage for the pkcs12 command. Also change the default 817259191Skris behaviour so MAC iteration counts are used unless the new 817359191Skris -nomaciter option is used. This improves file security and 817459191Skris only older versions of MSIE (4.0 for example) need it. 817559191Skris [Steve Henson] 817659191Skris 817759191Skris *) Honor the no-xxx Configure options when creating .DEF files. 817859191Skris [Ulf M�ller] 817959191Skris 818059191Skris *) Add PKCS#10 attributes to field table: challengePassword, 818159191Skris unstructuredName and unstructuredAddress. These are taken from 818259191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 818359191Skris international characters are used. 818459191Skris 818559191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 818659191Skris based on strings. Remove the 'loc' parameter when adding 818759191Skris attributes because these will be a SET OF encoding which is sorted 818859191Skris in ASN1 order. 818959191Skris [Steve Henson] 819059191Skris 819159191Skris *) Initial changes to the 'req' utility to allow request generation 819259191Skris automation. This will allow an application to just generate a template 819359191Skris file containing all the field values and have req construct the 819459191Skris request. 819559191Skris 819659191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 819759191Skris used all over the place including certificate requests and PKCS#7 819859191Skris structures. They are currently handled manually where necessary with 819959191Skris some primitive wrappers for PKCS#7. The new functions behave in a 820059191Skris manner analogous to the X509 extension functions: they allow 820159191Skris attributes to be looked up by NID and added. 820259191Skris 820359191Skris Later something similar to the X509V3 code would be desirable to 820459191Skris automatically handle the encoding, decoding and printing of the 820559191Skris more complex types. The string types like challengePassword can 820659191Skris be handled by the string table functions. 820759191Skris 820859191Skris Also modified the multi byte string table handling. Now there is 820959191Skris a 'global mask' which masks out certain types. The table itself 821059191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 821159191Skris is useful when for example there is only one permissible type 821259191Skris (as in countryName) and using the mask might result in no valid 821359191Skris types at all. 821459191Skris [Steve Henson] 821559191Skris 821659191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 821759191Skris SSL_get_peer_finished to allow applications to obtain the latest 821859191Skris Finished messages sent to the peer or expected from the peer, 821959191Skris respectively. (SSL_get_peer_finished is usually the Finished message 822059191Skris actually received from the peer, otherwise the protocol will be aborted.) 822159191Skris 822259191Skris As the Finished message are message digests of the complete handshake 822359191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 822459191Skris be used for external authentication procedures when the authentication 822559191Skris provided by SSL/TLS is not desired or is not enough. 822659191Skris [Bodo Moeller] 822759191Skris 822859191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 822959191Skris the host supports BWX extension and if Compaq C is present on the 823059191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 823159191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 823259191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 823359191Skris SHA1. 823459191Skris [Andy Polyakov] 823559191Skris 823659191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 823759191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 823859191Skris weak crypto and after checking the certificate is SGC a second one 823959191Skris with strong crypto. MS SGC stops the first handshake after receiving 824059191Skris the server certificate message and sends a second client hello. Since 824159191Skris a server will typically do all the time consuming operations before 824259191Skris expecting any further messages from the client (server key exchange 824359191Skris is the most expensive) there is little difference between the two. 824459191Skris 824559191Skris To get OpenSSL to support MS SGC we have to permit a second client 824659191Skris hello message after we have sent server done. In addition we have to 824759191Skris reset the MAC if we do get this second client hello. 824859191Skris [Steve Henson] 824959191Skris 825059191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 825159191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 825259191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 825359191Skris format DER encoded private key. Newer code should use PKCS#8 format which 825459191Skris has the key type encoded in the ASN1 structure. Added DER private key 825559191Skris support to pkcs8 application. 825659191Skris [Steve Henson] 825759191Skris 825859191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 825959191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 826059191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 826159191Skris is set, we interpret this as a request to violate the specification 826259191Skris (the worst that can happen is a handshake failure, and 'correct' 826359191Skris behaviour would result in a handshake failure anyway). 826459191Skris [Bodo Moeller] 826559191Skris 826659191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 826759191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 826859191Skris concurrently obtain them from an external cache). 826959191Skris The internal cache can handle only one SSL_SESSION with a given ID, 827059191Skris so if there's a conflict, we now throw out the old one to achieve 827159191Skris consistency. 827259191Skris [Bodo Moeller] 827359191Skris 827459191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 827559191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 827659191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 827759191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 827859191Skris example. 827959191Skris [Steve Henson] 828059191Skris 828159191Skris *) Simplify the trust setting structure and code. Now we just have 828259191Skris two sequences of OIDs for trusted and rejected settings. These will 828359191Skris typically have values the same as the extended key usage extension 828459191Skris and any application specific purposes. 828559191Skris 828659191Skris The trust checking code now has a default behaviour: it will just 828759191Skris check for an object with the same NID as the passed id. Functions can 828859191Skris be provided to override either the default behaviour or the behaviour 828959191Skris for a given id. SSL client, server and email already have functions 829059191Skris in place for compatibility: they check the NID and also return "trusted" 829159191Skris if the certificate is self signed. 829259191Skris [Steve Henson] 829359191Skris 829459191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 829559191Skris traditional format into an EVP_PKEY structure. 829659191Skris [Steve Henson] 829759191Skris 829859191Skris *) Add a password callback function PEM_cb() which either prompts for 829959191Skris a password if usr_data is NULL or otherwise assumes it is a null 830059191Skris terminated password. Allow passwords to be passed on command line 830159191Skris environment or config files in a few more utilities. 830259191Skris [Steve Henson] 830359191Skris 830459191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 830559191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 830659191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 830759191Skris Update documentation. 830859191Skris [Steve Henson] 830959191Skris 831059191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 831159191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 831259191Skris and produce an error if it couldn't. For compatibility we also have 831359191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 831459191Skris don't allocate anything because they don't need to. 831559191Skris [Steve Henson] 831659191Skris 831759191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 831859191Skris for details. 831959191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 832059191Skris 832159191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 832259191Skris possibly others as well. The purpose is to make an interface that 832359191Skris provide hooks so anyone can build a separate set of allocation and 832459191Skris deallocation routines to be used by OpenSSL, for example memory 832559191Skris pool implementations, or something else, which was previously hard 832659191Skris since Malloc(), Realloc() and Free() were defined as macros having 832759191Skris the values malloc, realloc and free, respectively (except for Win32 832859191Skris compilations). The same is provided for memory debugging code. 832959191Skris OpenSSL already comes with functionality to find memory leaks, but 833059191Skris this gives people a chance to debug other memory problems. 833159191Skris 833259191Skris With these changes, a new set of functions and macros have appeared: 833359191Skris 833468651Skris CRYPTO_set_mem_debug_functions() [F] 833568651Skris CRYPTO_get_mem_debug_functions() [F] 833668651Skris CRYPTO_dbg_set_options() [F] 833768651Skris CRYPTO_dbg_get_options() [F] 833868651Skris CRYPTO_malloc_debug_init() [M] 833959191Skris 834059191Skris The memory debug functions are NULL by default, unless the library 834159191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 834259191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 834359191Skris gives the standard debugging functions that come with OpenSSL) or 834459191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 834559191Skris provided by the library user) must be used. When the standard 834659191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 834759191Skris request additional information: 834859191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 834959191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 835059191Skris 835159191Skris Also, things like CRYPTO_set_mem_functions will always give the 835259191Skris expected result (the new set of functions is used for allocation 835359191Skris and deallocation) at all times, regardless of platform and compiler 835459191Skris options. 835559191Skris 835659191Skris To finish it up, some functions that were never use in any other 835759191Skris way than through macros have a new API and new semantic: 835859191Skris 835959191Skris CRYPTO_dbg_malloc() 836059191Skris CRYPTO_dbg_realloc() 836159191Skris CRYPTO_dbg_free() 836259191Skris 836359191Skris All macros of value have retained their old syntax. 836459191Skris [Richard Levitte and Bodo Moeller] 836559191Skris 836659191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 836759191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 836859191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 836959191Skris algorithm. 837059191Skris [Steve Henson] 837159191Skris 837259191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 837359191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 837459191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 837559191Skris 837659191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 837759191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 837859191Skris functionality to handle multipart/signed properly) and a utility 837959191Skris called 'smime' to call all this stuff. This is based on code I 838059191Skris originally wrote for Celo who have kindly allowed it to be 838159191Skris included in OpenSSL. 838259191Skris [Steve Henson] 838359191Skris 838459191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 838559191Skris des_set_key (aka des_key_sched). Global variable des_check_key 838659191Skris decides which of these is called by des_set_key; this way 838759191Skris des_check_key behaves as it always did, but applications and 838859191Skris the library itself, which was buggy for des_check_key == 1, 838959191Skris have a cleaner way to pick the version they need. 839059191Skris [Bodo Moeller] 839159191Skris 839259191Skris *) New function PKCS12_newpass() which changes the password of a 839359191Skris PKCS12 structure. 839459191Skris [Steve Henson] 839559191Skris 839659191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 839759191Skris dynamic mix. In both cases the ids can be used as an index into the 839859191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 839959191Skris functions so they accept a list of the field values and the 840059191Skris application doesn't need to directly manipulate the X509_TRUST 840159191Skris structure. 840259191Skris [Steve Henson] 840359191Skris 840459191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 840559191Skris need initialising. 840659191Skris [Steve Henson] 840759191Skris 840859191Skris *) Modify the way the V3 extension code looks up extensions. This now 840959191Skris works in a similar way to the object code: we have some "standard" 841059191Skris extensions in a static table which is searched with OBJ_bsearch() 841159191Skris and the application can add dynamic ones if needed. The file 841259191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 841359191Skris updated whenever a new extension is added to the core code and kept 841459191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 841559191Skris this. New extensions are not added too often so this file can readily 841659191Skris be maintained manually. 841759191Skris 841859191Skris There are two big advantages in doing things this way. The extensions 841959191Skris can be looked up immediately and no longer need to be "added" using 842059191Skris X509V3_add_standard_extensions(): this function now does nothing. 842159191Skris [Side note: I get *lots* of email saying the extension code doesn't 842259191Skris work because people forget to call this function] 842359191Skris Also no dynamic allocation is done unless new extensions are added: 842459191Skris so if we don't add custom extensions there is no need to call 842559191Skris X509V3_EXT_cleanup(). 842659191Skris [Steve Henson] 842759191Skris 842859191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 842959191Skris magic header, so unsalted files fail gracefully instead of just decrypting 843059191Skris to garbage. This is because not salting is a big security hole, so people 843159191Skris should be discouraged from doing it. 843259191Skris [Ben Laurie] 843359191Skris 843459191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 843559191Skris digest to be passed on the command line but it only used this 843659191Skris parameter when signing a certificate. Modified so all relevant 843759191Skris operations are affected by the digest parameter including the 843859191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 843959191Skris DSA key was used because it didn't fix the digest. 844059191Skris [Steve Henson] 844159191Skris 844259191Skris *) Initial certificate chain verify code. Currently tests the untrusted 844359191Skris certificates for consistency with the verify purpose (which is set 844459191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 844559191Skris 844659191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 844759191Skris this is because it will reject chains with invalid extensions whereas 844859191Skris every previous version of OpenSSL and SSLeay made no checks at all. 844959191Skris 845059191Skris Trust code: checks the root CA for the relevant trust settings. Trust 845159191Skris settings have an initial value consistent with the verify purpose: e.g. 845259191Skris if the verify purpose is for SSL client use it expects the CA to be 845359191Skris trusted for SSL client use. However the default value can be changed to 845459191Skris permit custom trust settings: one example of this would be to only trust 845559191Skris certificates from a specific "secure" set of CAs. 845659191Skris 845759191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 845859191Skris which should be used for version portability: especially since the 845959191Skris verify structure is likely to change more often now. 846059191Skris 846159191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 846259191Skris to set them. If not set then assume SSL clients will verify SSL servers 846359191Skris and vice versa. 846459191Skris 846559191Skris Two new options to the verify program: -untrusted allows a set of 846659191Skris untrusted certificates to be passed in and -purpose which sets the 846759191Skris intended purpose of the certificate. If a purpose is set then the 846859191Skris new chain verify code is used to check extension consistency. 846959191Skris [Steve Henson] 847059191Skris 847159191Skris *) Support for the authority information access extension. 847259191Skris [Steve Henson] 847359191Skris 847459191Skris *) Modify RSA and DSA PEM read routines to transparently handle 847559191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 847659191Skris public keys in a format compatible with certificate 847759191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 847859191Skris functions called *_PublicKey_* which used various odd formats so 847959191Skris these are retained for compatibility: however the DSA variants were 848059191Skris never in a public release so they have been deleted. Changed dsa/rsa 848159191Skris utilities to handle the new format: note no releases ever handled public 848259191Skris keys so we should be OK. 848359191Skris 848459191Skris The primary motivation for this change is to avoid the same fiasco 848559191Skris that dogs private keys: there are several incompatible private key 848659191Skris formats some of which are standard and some OpenSSL specific and 848759191Skris require various evil hacks to allow partial transparent handling and 848859191Skris even then it doesn't work with DER formats. Given the option anything 848959191Skris other than PKCS#8 should be dumped: but the other formats have to 849059191Skris stay in the name of compatibility. 849159191Skris 849259191Skris With public keys and the benefit of hindsight one standard format 849359191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 849459191Skris it clearly returns an error if you try to read the wrong kind of key. 849559191Skris 849659191Skris Added a -pubkey option to the 'x509' utility to output the public key. 849759191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 849859191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 849959191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 850059191Skris that do the same as the EVP_PKEY_assign_*() except they up the 850159191Skris reference count of the added key (they don't "swallow" the 850259191Skris supplied key). 850359191Skris [Steve Henson] 850459191Skris 850559191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 850659191Skris CRLs would fail if the file contained no certificates or no CRLs: 850759191Skris added a new function to read in both types and return the number 850859191Skris read: this means that if none are read it will be an error. The 850959191Skris DER versions of the certificate and CRL reader would always fail 851059191Skris because it isn't possible to mix certificates and CRLs in DER format 851159191Skris without choking one or the other routine. Changed this to just read 851259191Skris a certificate: this is the best we can do. Also modified the code 851359191Skris in apps/verify.c to take notice of return codes: it was previously 851459191Skris attempting to read in certificates from NULL pointers and ignoring 851559191Skris any errors: this is one reason why the cert and CRL reader seemed 851659191Skris to work. It doesn't check return codes from the default certificate 851759191Skris routines: these may well fail if the certificates aren't installed. 851859191Skris [Steve Henson] 851959191Skris 852059191Skris *) Code to support otherName option in GeneralName. 852159191Skris [Steve Henson] 852259191Skris 852359191Skris *) First update to verify code. Change the verify utility 852459191Skris so it warns if it is passed a self signed certificate: 852559191Skris for consistency with the normal behaviour. X509_verify 852659191Skris has been modified to it will now verify a self signed 852759191Skris certificate if *exactly* the same certificate appears 852859191Skris in the store: it was previously impossible to trust a 852959191Skris single self signed certificate. This means that: 853059191Skris openssl verify ss.pem 853159191Skris now gives a warning about a self signed certificate but 853259191Skris openssl verify -CAfile ss.pem ss.pem 853359191Skris is OK. 853459191Skris [Steve Henson] 853559191Skris 853659191Skris *) For servers, store verify_result in SSL_SESSION data structure 853759191Skris (and add it to external session representation). 853859191Skris This is needed when client certificate verifications fails, 853959191Skris but an application-provided verification callback (set by 854059191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 854159191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 854259191Skris but returns 1): When the session is reused, we have to set 854359191Skris ssl->verify_result to the appropriate error code to avoid 854459191Skris security holes. 854559191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 854659191Skris 854759191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 854859191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 854959191Skris didn't contain any existing data because it was being created. 855059191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 855159191Skris 855259191Skris *) Add a salt to the key derivation routines in enc.c. This 855359191Skris forms the first 8 bytes of the encrypted file. Also add a 855459191Skris -S option to allow a salt to be input on the command line. 855559191Skris [Steve Henson] 855659191Skris 855759191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 855859191Skris to compare two certificates. We do this by working out the SHA1 855959191Skris hash and comparing that. X509_cmp() will be needed by the trust 856059191Skris code. 856159191Skris [Steve Henson] 856259191Skris 856359191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 856459191Skris the reference count in the SSL_SESSION returned. 856559191Skris [Geoff Thorpe <geoff@eu.c2.net>] 856659191Skris 856759191Skris *) Fix for 'req': it was adding a null to request attributes. 856859191Skris Also change the X509_LOOKUP and X509_INFO code to handle 856959191Skris certificate auxiliary information. 857059191Skris [Steve Henson] 857159191Skris 857259191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 857359191Skris the 'enc' command. 857459191Skris [Steve Henson] 857559191Skris 857659191Skris *) Add the possibility to add extra information to the memory leak 857759191Skris detecting output, to form tracebacks, showing from where each 857859191Skris allocation was originated: CRYPTO_push_info("constant string") adds 857959191Skris the string plus current file name and line number to a per-thread 858059191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 858159191Skris is like calling CYRPTO_pop_info() until the stack is empty. 858259191Skris Also updated memory leak detection code to be multi-thread-safe. 858359191Skris [Richard Levitte] 858459191Skris 858559191Skris *) Add options -text and -noout to pkcs7 utility and delete the 858659191Skris encryption options which never did anything. Update docs. 858759191Skris [Steve Henson] 858859191Skris 858959191Skris *) Add options to some of the utilities to allow the pass phrase 859059191Skris to be included on either the command line (not recommended on 859159191Skris OSes like Unix) or read from the environment. Update the 859259191Skris manpages and fix a few bugs. 859359191Skris [Steve Henson] 859459191Skris 859559191Skris *) Add a few manpages for some of the openssl commands. 859659191Skris [Steve Henson] 859759191Skris 859859191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 859959191Skris leaking and not finding already revoked certificates. 860059191Skris [Steve Henson] 860159191Skris 860259191Skris *) Extensive changes to support certificate auxiliary information. 860359191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 860459191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 860559191Skris can still read in a certificate file in the usual way but it 860659191Skris will also read in any additional "auxiliary information". By 860759191Skris doing things this way a fair degree of compatibility can be 860859191Skris retained: existing certificates can have this information added 860959191Skris using the new 'x509' options. 861059191Skris 861159191Skris Current auxiliary information includes an "alias" and some trust 861259191Skris settings. The trust settings will ultimately be used in enhanced 861359191Skris certificate chain verification routines: currently a certificate 861459191Skris can only be trusted if it is self signed and then it is trusted 861559191Skris for all purposes. 861659191Skris [Steve Henson] 861759191Skris 861859191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 861959191Skris The problem was that one of the replacement routines had not been working 862059191Skris since SSLeay releases. For now the offending routine has been replaced 862159191Skris with non-optimised assembler. Even so, this now gives around 95% 862259191Skris performance improvement for 1024 bit RSA signs. 862359191Skris [Mark Cox] 862459191Skris 862559191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 862659191Skris handling. Most clients have the effective key size in bits equal to 862759191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 862859191Skris A few however don't do this and instead use the size of the decrypted key 862959191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 863059191Skris the effective key length. In this case the effective key length can still 863159191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 863259191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 863359191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 863459191Skris the key length and effective key length are equal. 863559191Skris [Steve Henson] 863659191Skris 863759191Skris *) Add a bunch of functions that should simplify the creation of 863859191Skris X509_NAME structures. Now you should be able to do: 863959191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 864059191Skris and have it automatically work out the correct field type and fill in 864159191Skris the structures. The more adventurous can try: 864259191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 864359191Skris and it will (hopefully) work out the correct multibyte encoding. 864459191Skris [Steve Henson] 864559191Skris 864659191Skris *) Change the 'req' utility to use the new field handling and multibyte 864759191Skris copy routines. Before the DN field creation was handled in an ad hoc 864859191Skris way in req, ca, and x509 which was rather broken and didn't support 864959191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 865059191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 865159191Skris using the dirstring_type option. See the new comment in the default 865259191Skris openssl.cnf for more info. 865359191Skris [Steve Henson] 865459191Skris 865559191Skris *) Make crypto/rand/md_rand.c more robust: 865659191Skris - Assure unique random numbers after fork(). 865759191Skris - Make sure that concurrent threads access the global counter and 865859191Skris md serializably so that we never lose entropy in them 865959191Skris or use exactly the same state in multiple threads. 866059191Skris Access to the large state is not always serializable because 866159191Skris the additional locking could be a performance killer, and 866259191Skris md should be large enough anyway. 866359191Skris [Bodo Moeller] 866459191Skris 866559191Skris *) New file apps/app_rand.c with commonly needed functionality 866659191Skris for handling the random seed file. 866759191Skris 866859191Skris Use the random seed file in some applications that previously did not: 866959191Skris ca, 867059191Skris dsaparam -genkey (which also ignored its '-rand' option), 867159191Skris s_client, 867259191Skris s_server, 867359191Skris x509 (when signing). 867459191Skris Except on systems with /dev/urandom, it is crucial to have a random 867559191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 867659191Skris for RSA signatures we could do without one. 867759191Skris 867859191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 867959191Skris of each file listed in the '-rand' option. The function as previously 868059191Skris found in genrsa is now in app_rand.c and is used by all programs 868159191Skris that support '-rand'. 868259191Skris [Bodo Moeller] 868359191Skris 868459191Skris *) In RAND_write_file, use mode 0600 for creating files; 868559191Skris don't just chmod when it may be too late. 868659191Skris [Bodo Moeller] 868759191Skris 868859191Skris *) Report an error from X509_STORE_load_locations 868959191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 869059191Skris [Bill Perry] 869159191Skris 869259191Skris *) New function ASN1_mbstring_copy() this copies a string in either 869359191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 869459191Skris into an ASN1_STRING type. A mask of permissible types is passed 869559191Skris and it chooses the "minimal" type to use or an error if not type 869659191Skris is suitable. 869759191Skris [Steve Henson] 869859191Skris 869959191Skris *) Add function equivalents to the various macros in asn1.h. The old 870059191Skris macros are retained with an M_ prefix. Code inside the library can 870159191Skris use the M_ macros. External code (including the openssl utility) 870259191Skris should *NOT* in order to be "shared library friendly". 870359191Skris [Steve Henson] 870459191Skris 870559191Skris *) Add various functions that can check a certificate's extensions 870659191Skris to see if it usable for various purposes such as SSL client, 870759191Skris server or S/MIME and CAs of these types. This is currently 870859191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 870959191Skris verification. Also added a -purpose flag to x509 utility to 871059191Skris print out all the purposes. 871159191Skris [Steve Henson] 871259191Skris 871359191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 871459191Skris functions. 871559191Skris [Steve Henson] 871659191Skris 871759191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 871859191Skris for, obtain and decode and extension and obtain its critical flag. 871959191Skris This allows all the necessary extension code to be handled in a 872059191Skris single function call. 872159191Skris [Steve Henson] 872259191Skris 872359191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 872459191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 872559191Skris [Andy Polyakov] 872659191Skris 872759191Skris *) New -noout option to asn1parse. This causes no output to be produced 872859191Skris its main use is when combined with -strparse and -out to extract data 872959191Skris from a file (which may not be in ASN.1 format). 873059191Skris [Steve Henson] 873159191Skris 873259191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 873359191Skris when producing the local key id. 873459191Skris [Richard Levitte <levitte@stacken.kth.se>] 873559191Skris 873659191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 873759191Skris stated explicitly. If it is not stated then it tries the first server 873859191Skris certificate file. The previous behaviour hard coded the filename 873959191Skris "server.pem". 874059191Skris [Steve Henson] 874159191Skris 874259191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 874359191Skris a public key to be input or output. For example: 874459191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 874559191Skris Also added necessary DSA public key functions to handle this. 874659191Skris [Steve Henson] 874759191Skris 874859191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 874959191Skris in the message. This was handled by allowing 875059191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 875159191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 875259191Skris 875359191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 875459191Skris to the end of the strings whereas this didn't. This would cause problems 875559191Skris if strings read with d2i_ASN1_bytes() were later modified. 875659191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 875759191Skris 875859191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 875959191Skris data and it contains EOF it will end up returning an error. This is 876059191Skris caused by input 46 bytes long. The cause is due to the way base64 876159191Skris BIOs find the start of base64 encoded data. They do this by trying a 876259191Skris trial decode on each line until they find one that works. When they 876359191Skris do a flag is set and it starts again knowing it can pass all the 876459191Skris data directly through the decoder. Unfortunately it doesn't reset 876559191Skris the context it uses. This means that if EOF is reached an attempt 876659191Skris is made to pass two EOFs through the context and this causes the 876759191Skris resulting error. This can also cause other problems as well. As is 876859191Skris usual with these problems it takes *ages* to find and the fix is 876959191Skris trivial: move one line. 877059191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 877159191Skris 877259191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 877359191Skris old code wouldn't work because it needed to select() on sockets and the 877459191Skris tty (for keypresses and to see if data could be written). Win32 only 877559191Skris supports select() on sockets so we select() with a 1s timeout on the 877659191Skris sockets and then see if any characters are waiting to be read, if none 877759191Skris are present then we retry, we also assume we can always write data to 877859191Skris the tty. This isn't nice because the code then blocks until we've 877959191Skris received a complete line of data and it is effectively polling the 878059191Skris keyboard at 1s intervals: however it's quite a bit better than not 878159191Skris working at all :-) A dedicated Windows application might handle this 878259191Skris with an event loop for example. 878359191Skris [Steve Henson] 878459191Skris 878559191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 878659191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 878759191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 878859191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 878959191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 879059191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 879159191Skris This necessitated the support of an extra signature type NID_md5_sha1 879259191Skris for SSL signatures and modifications to the SSL library to use it instead 879359191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 879459191Skris [Steve Henson] 879559191Skris 879659191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 879759191Skris will lookup a CRL issuers certificate and verify the signature in a 879859191Skris similar way to the verify program. Tidy up the crl program so it 879959191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 880059191Skris less strict. It will now permit CRL extensions even if it is not 880159191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 880259191Skris [Steve Henson] 880359191Skris 880459191Skris *) Initialize all non-automatic variables each time one of the openssl 880559191Skris sub-programs is started (this is necessary as they may be started 880659191Skris multiple times from the "OpenSSL>" prompt). 880759191Skris [Lennart Bang, Bodo Moeller] 880859191Skris 880959191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 881059191Skris removing all other RSA functionality (this is what NO_RSA does). This 881159191Skris is so (for example) those in the US can disable those operations covered 881259191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 881359191Skris key generation. 881459191Skris [Steve Henson] 881559191Skris 881659191Skris *) Non-copying interface to BIO pairs. 881759191Skris (still largely untested) 881859191Skris [Bodo Moeller] 881959191Skris 882059191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 882159191Skris ASCII string. This was handled independently in various places before. 882259191Skris [Steve Henson] 882359191Skris 882459191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 882559191Skris UTF8 strings a character at a time. 882659191Skris [Steve Henson] 882759191Skris 882859191Skris *) Use client_version from client hello to select the protocol 882959191Skris (s23_srvr.c) and for RSA client key exchange verification 883059191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 883159191Skris [Bodo Moeller] 883259191Skris 883359191Skris *) Add various utility functions to handle SPKACs, these were previously 883459191Skris handled by poking round in the structure internals. Added new function 883559191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 883659191Skris print, verify and generate SPKACs. Based on an original idea from 883759191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 883859191Skris [Steve Henson] 883959191Skris 884059191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 884159191Skris [Andy Polyakov] 884259191Skris 884359191Skris *) Allow the config file extension section to be overwritten on the 884459191Skris command line. Based on an original idea from Massimiliano Pala 884559191Skris <madwolf@comune.modena.it>. The new option is called -extensions 884659191Skris and can be applied to ca, req and x509. Also -reqexts to override 884759191Skris the request extensions in req and -crlexts to override the crl extensions 884859191Skris in ca. 884959191Skris [Steve Henson] 885059191Skris 885159191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 885259191Skris the same field multiple times by preceding it by "XXXX." for example: 885359191Skris 1.OU="Unit name 1" 885459191Skris 2.OU="Unit name 2" 885559191Skris this is the same syntax as used in the req config file. 885659191Skris [Steve Henson] 885759191Skris 885859191Skris *) Allow certificate extensions to be added to certificate requests. These 885959191Skris are specified in a 'req_extensions' option of the req section of the 886059191Skris config file. They can be printed out with the -text option to req but 886159191Skris are otherwise ignored at present. 886259191Skris [Steve Henson] 886359191Skris 886459191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 886559191Skris data read consists of only the final block it would not decrypted because 886659191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 886759191Skris A misplaced 'break' also meant the decrypted final block might not be 886859191Skris copied until the next read. 886959191Skris [Steve Henson] 887059191Skris 887159191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 887259191Skris a few extra parameters to the DH structure: these will be useful if 887359191Skris for example we want the value of 'q' or implement X9.42 DH. 887459191Skris [Steve Henson] 887559191Skris 887659191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 887759191Skris provides hooks that allow the default DSA functions or functions on a 887859191Skris "per key" basis to be replaced. This allows hardware acceleration and 887959191Skris hardware key storage to be handled without major modification to the 888059191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 888159191Skris associated functions. 888259191Skris [Steve Henson] 888359191Skris 888459191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 888559191Skris as "read only": it can't be written to and the buffer it points to will 888659191Skris not be freed. Reading from a read only BIO is much more efficient than 888759191Skris a normal memory BIO. This was added because there are several times when 888859191Skris an area of memory needs to be read from a BIO. The previous method was 888959191Skris to create a memory BIO and write the data to it, this results in two 889059191Skris copies of the data and an O(n^2) reading algorithm. There is a new 889159191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 889259191Skris an area of memory. Also modified the PKCS#7 routines to use read only 889359191Skris memory BIOs. 889459191Skris [Steve Henson] 889559191Skris 889659191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 889759191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 889859191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 889959191Skris but a retry condition occured while trying to read the rest. 890059191Skris [Bodo Moeller] 890159191Skris 890259191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 890359191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 890459191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 890559191Skris the encrypted data type: this is a more sensible place to put it and it 890659191Skris allows the PKCS#12 code to be tidied up that duplicated this 890759191Skris functionality. 890859191Skris [Steve Henson] 890959191Skris 891059191Skris *) Changed obj_dat.pl script so it takes its input and output files on 891159191Skris the command line. This should avoid shell escape redirection problems 891259191Skris under Win32. 891359191Skris [Steve Henson] 891459191Skris 891559191Skris *) Initial support for certificate extension requests, these are included 891659191Skris in things like Xenroll certificate requests. Included functions to allow 891759191Skris extensions to be obtained and added. 891859191Skris [Steve Henson] 891959191Skris 892059191Skris *) -crlf option to s_client and s_server for sending newlines as 892159191Skris CRLF (as required by many protocols). 892259191Skris [Bodo Moeller] 892359191Skris 892455714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 892555714Skris 892655714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 892755714Skris [Ralf S. Engelschall] 892855714Skris 892955714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 893055714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 893155714Skris 893255714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 893355714Skris program. 893455714Skris [Steve Henson] 893555714Skris 893655714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 893755714Skris DH parameters/keys (q is lost during that conversion, but the resulting 893855714Skris DH parameters contain its length). 893955714Skris 894055714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 894155714Skris much faster than DH_generate_parameters (which creates parameters 894255714Skris where p = 2*q + 1), and also the smaller q makes DH computations 894355714Skris much more efficient (160-bit exponentiation instead of 1024-bit 894455714Skris exponentiation); so this provides a convenient way to support DHE 894555714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 894655714Skris utter importance to use 894755714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 894855714Skris or 894955714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 895055714Skris when such DH parameters are used, because otherwise small subgroup 895155714Skris attacks may become possible! 895255714Skris [Bodo Moeller] 895355714Skris 895455714Skris *) Avoid memory leak in i2d_DHparams. 895555714Skris [Bodo Moeller] 895655714Skris 895755714Skris *) Allow the -k option to be used more than once in the enc program: 895855714Skris this allows the same encrypted message to be read by multiple recipients. 895955714Skris [Steve Henson] 896055714Skris 896155714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 896255714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 896355714Skris it will always use the numerical form of the OID, even if it has a short 896455714Skris or long name. 896555714Skris [Steve Henson] 896655714Skris 896755714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 896855714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 896955714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 897055714Skris no private key components need be present and it might store extra data 897159191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 897259191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 897359191Skris private key operations. 897455714Skris [Steve Henson] 897555714Skris 897655714Skris *) Added support for SPARC Linux. 897755714Skris [Andy Polyakov] 897855714Skris 897955714Skris *) pem_password_cb function type incompatibly changed from 898055714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 898155714Skris to 898255714Skris ....(char *buf, int size, int rwflag, void *userdata); 898355714Skris so that applications can pass data to their callbacks: 898455714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 898555714Skris additional void * argument, which is just handed through whenever 898655714Skris the password callback is called. 898759191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 898855714Skris 898955714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 899055714Skris 899155714Skris Compatibility note: As many C implementations push function arguments 899255714Skris onto the stack in reverse order, the new library version is likely to 899355714Skris interoperate with programs that have been compiled with the old 899455714Skris pem_password_cb definition (PEM_whatever takes some data that 899555714Skris happens to be on the stack as its last argument, and the callback 899655714Skris just ignores this garbage); but there is no guarantee whatsoever that 899755714Skris this will work. 899855714Skris 899955714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 900055714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 900155714Skris problems not only on Windows, but also on some Unix platforms. 900255714Skris To avoid problematic command lines, these definitions are now in an 900355714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 900455714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 900555714Skris [Bodo Moeller] 900655714Skris 900755714Skris *) MIPS III/IV assembler module is reimplemented. 900855714Skris [Andy Polyakov] 900955714Skris 901055714Skris *) More DES library cleanups: remove references to srand/rand and 901155714Skris delete an unused file. 901255714Skris [Ulf M�ller] 901355714Skris 901455714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 901555714Skris since not many people have MASM (ml) and it can be hard to obtain. 901655714Skris This is currently experimental but it seems to work OK and pass all 901755714Skris the tests. Check out INSTALL.W32 for info. 901855714Skris [Steve Henson] 901955714Skris 902055714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 902155714Skris without temporary keys kept an extra copy of the server key, 902255714Skris and connections with temporary keys did not free everything in case 902355714Skris of an error. 902455714Skris [Bodo Moeller] 902555714Skris 902655714Skris *) New function RSA_check_key and new openssl rsa option -check 902755714Skris for verifying the consistency of RSA keys. 902855714Skris [Ulf Moeller, Bodo Moeller] 902955714Skris 903055714Skris *) Various changes to make Win32 compile work: 903155714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 903255714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 903355714Skris comparison" warnings. 903455714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 903555714Skris [Steve Henson] 903655714Skris 903755714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 903855714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 903955714Skris derived keys are printed to stderr. 904055714Skris [Steve Henson] 904155714Skris 904255714Skris *) Copy the flags in ASN1_STRING_dup(). 904355714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 904455714Skris 904555714Skris *) The x509 application mishandled signing requests containing DSA 904655714Skris keys when the signing key was also DSA and the parameters didn't match. 904755714Skris 904855714Skris It was supposed to omit the parameters when they matched the signing key: 904955714Skris the verifying software was then supposed to automatically use the CA's 905055714Skris parameters if they were absent from the end user certificate. 905155714Skris 905255714Skris Omitting parameters is no longer recommended. The test was also 905355714Skris the wrong way round! This was probably due to unusual behaviour in 905455714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 905555714Skris This meant that parameters were omitted when they *didn't* match and 905655714Skris the certificate was useless. Certificates signed with 'ca' didn't have 905755714Skris this bug. 905855714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 905955714Skris 906055714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 906155714Skris The interface is as follows: 906255714Skris Applications can use 906355714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 906455714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 906555714Skris "off" is now the default. 906655714Skris The library internally uses 906755714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 906855714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 906955714Skris to disable memory-checking temporarily. 907055714Skris 907155714Skris Some inconsistent states that previously were possible (and were 907255714Skris even the default) are now avoided. 907355714Skris 907455714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 907555714Skris with each memory chunk allocated; this is occasionally more helpful 907655714Skris than just having a counter. 907755714Skris 907855714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 907955714Skris 908055714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 908155714Skris extensions. 908255714Skris [Bodo Moeller] 908355714Skris 908455714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 908555714Skris which largely parallels "options", but is for changing API behaviour, 908655714Skris whereas "options" are about protocol behaviour. 908755714Skris Initial "mode" flags are: 908855714Skris 908955714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 909055714Skris a single record has been written. 909155714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 909255714Skris retries use the same buffer location. 909355714Skris (But all of the contents must be 909455714Skris copied!) 909555714Skris [Bodo Moeller] 909655714Skris 909779998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 909855714Skris worked. 909955714Skris 910055714Skris *) Fix problems with no-hmac etc. 910155714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 910255714Skris 910355714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 910455714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 910555714Skris to mess around with the internals of an RSA structure. 910655714Skris [Steve Henson] 910755714Skris 910855714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 910955714Skris Also really enable memory leak checks in openssl.c and in some 911055714Skris test programs. 911155714Skris [Chad C. Mulligan, Bodo Moeller] 911255714Skris 911355714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 911455714Skris up the length of negative integers. This has now been simplified to just 911555714Skris store the length when it is first determined and use it later, rather 911655714Skris than trying to keep track of where data is copied and updating it to 911755714Skris point to the end. 911855714Skris [Steve Henson, reported by Brien Wheeler 911955714Skris <bwheeler@authentica-security.com>] 912055714Skris 912155714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 912255714Skris of a PKCS#7 signature but with the signing certificate passed to the 912355714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 912455714Skris certificate is present in the PKCS#7 structure. This isn't always the 912555714Skris case: certificates can be omitted from a PKCS#7 structure and be 912655714Skris distributed by "out of band" means (such as a certificate database). 912755714Skris [Steve Henson] 912855714Skris 912955714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 913055714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 913155714Skris necessary function names. 913255714Skris [Steve Henson] 913355714Skris 913455714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 913555714Skris options set by Configure in the top level Makefile, and Configure 913655714Skris was not even able to write more than one option correctly. 913755714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 913855714Skris [Bodo Moeller] 913955714Skris 914055714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 914155714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 914255714Skris for example allow memory BIOs to contain config info. 914355714Skris [Steve Henson] 914455714Skris 914555714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 914655714Skris Whoever hopes to achieve shared-library compatibility across versions 914755714Skris must use this, not the compile-time macro. 914855714Skris (Exercise 0.9.4: Which is the minimum library version required by 914955714Skris such programs?) 915055714Skris Note: All this applies only to multi-threaded programs, others don't 915155714Skris need locks. 915255714Skris [Bodo Moeller] 915355714Skris 915455714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 915555714Skris through a BIO pair triggered the default case, i.e. 915655714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 915755714Skris [Bodo Moeller] 915855714Skris 915955714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 916055714Skris can use the SSL library even if none of the specific BIOs is 916155714Skris appropriate. 916255714Skris [Bodo Moeller] 916355714Skris 916455714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 916555714Skris for the encoded length. 916655714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 916755714Skris 916855714Skris *) Add initial documentation of the X509V3 functions. 916955714Skris [Steve Henson] 917055714Skris 917155714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 917255714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 917355714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 917455714Skris secure PKCS#8 private key format with a high iteration count. 917555714Skris [Steve Henson] 917655714Skris 917755714Skris *) Fix determination of Perl interpreter: A perl or perl5 917855714Skris _directory_ in $PATH was also accepted as the interpreter. 917955714Skris [Ralf S. Engelschall] 918055714Skris 918155714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 918255714Skris wrong with it but it was very old and did things like calling 918355714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 918455714Skris unusual formatting. 918555714Skris [Steve Henson] 918655714Skris 918755714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 918855714Skris to use the new extension code. 918955714Skris [Steve Henson] 919055714Skris 919155714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 919255714Skris with macros. This should make it easier to change their form, add extra 919355714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 919455714Skris constant. 919555714Skris [Steve Henson] 919655714Skris 919755714Skris *) Add to configuration table a new entry that can specify an alternative 919855714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 919955714Skris according to Mark Crispin <MRC@Panda.COM>. 920055714Skris [Bodo Moeller] 920155714Skris 920255714Skris#if 0 920355714Skris *) DES CBC did not update the IV. Weird. 920455714Skris [Ben Laurie] 920555714Skris#else 920655714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 920755714Skris Changing the behaviour of the former might break existing programs -- 920855714Skris where IV updating is needed, des_ncbc_encrypt can be used. 920955714Skris#endif 921055714Skris 921155714Skris *) When bntest is run from "make test" it drives bc to check its 921255714Skris calculations, as well as internally checking them. If an internal check 921355714Skris fails, it needs to cause bc to give a non-zero result or make test carries 921455714Skris on without noticing the failure. Fixed. 921555714Skris [Ben Laurie] 921655714Skris 921755714Skris *) DES library cleanups. 921855714Skris [Ulf M�ller] 921955714Skris 922055714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 922155714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 922255714Skris ciphers. NOTE: although the key derivation function has been verified 922355714Skris against some published test vectors it has not been extensively tested 922455714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 922555714Skris of v2.0. 922655714Skris [Steve Henson] 922755714Skris 922855714Skris *) Instead of "mkdir -p", which is not fully portable, use new 922955714Skris Perl script "util/mkdir-p.pl". 923055714Skris [Bodo Moeller] 923155714Skris 923255714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 923355714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 923455714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 923555714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 923655714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 923755714Skris underlying key generation function so it must do its own ASN1 parsing. 923855714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 923955714Skris 'parameter' argument instead of literal salt and iteration count values 924055714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 924155714Skris [Steve Henson] 924255714Skris 924355714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 924455714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 924555714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 924655714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 924755714Skris value was just used as a "magic string" and not used directly its 924855714Skris value doesn't matter. 924955714Skris [Steve Henson] 925055714Skris 925155714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 925255714Skris support mutable. 925355714Skris [Ben Laurie] 925455714Skris 925555714Skris *) "linux-sparc64" configuration (ultrapenguin). 925655714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 925755714Skris "linux-sparc" configuration. 925855714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 925955714Skris 926055714Skris *) config now generates no-xxx options for missing ciphers. 926155714Skris [Ulf M�ller] 926255714Skris 926355714Skris *) Support the EBCDIC character set (work in progress). 926455714Skris File ebcdic.c not yet included because it has a different license. 926555714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 926655714Skris 926755714Skris *) Support BS2000/OSD-POSIX. 926855714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 926955714Skris 927055714Skris *) Make callbacks for key generation use void * instead of char *. 927155714Skris [Ben Laurie] 927255714Skris 927355714Skris *) Make S/MIME samples compile (not yet tested). 927455714Skris [Ben Laurie] 927555714Skris 927655714Skris *) Additional typesafe stacks. 927755714Skris [Ben Laurie] 927855714Skris 927955714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 928055714Skris [Bodo Moeller] 928155714Skris 928255714Skris 928355714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 928455714Skris 928555714Skris *) New configuration variant "sco5-gcc". 928655714Skris 928755714Skris *) Updated some demos. 928855714Skris [Sean O Riordain, Wade Scholine] 928955714Skris 929055714Skris *) Add missing BIO_free at exit of pkcs12 application. 929155714Skris [Wu Zhigang] 929255714Skris 929355714Skris *) Fix memory leak in conf.c. 929455714Skris [Steve Henson] 929555714Skris 929655714Skris *) Updates for Win32 to assembler version of MD5. 929755714Skris [Steve Henson] 929855714Skris 929955714Skris *) Set #! path to perl in apps/der_chop to where we found it 930055714Skris instead of using a fixed path. 930155714Skris [Bodo Moeller] 930255714Skris 930355714Skris *) SHA library changes for irix64-mips4-cc. 930455714Skris [Andy Polyakov] 930555714Skris 930655714Skris *) Improvements for VMS support. 930755714Skris [Richard Levitte] 930855714Skris 930955714Skris 931055714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 931155714Skris 931255714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 931355714Skris This also avoids the problems with SC4.2 and unpatched SC5. 931455714Skris [Andy Polyakov <appro@fy.chalmers.se>] 931555714Skris 931655714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 931755714Skris These are required because of the typesafe stack would otherwise break 931855714Skris existing code. If old code used a structure member which used to be STACK 931955714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 932055714Skris sk_num or sk_value it would produce an error because the num, data members 932155714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 932255714Skris replaces the old method of assigning a value to sk_value 932355714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 932455714Skris that does this will no longer work (and should use sk_set instead) but 932555714Skris this could be regarded as a "questionable" behaviour anyway. 932655714Skris [Steve Henson] 932755714Skris 932855714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 932955714Skris correctly handle encrypted S/MIME data. 933055714Skris [Steve Henson] 933155714Skris 933255714Skris *) Change type of various DES function arguments from des_cblock 933355714Skris (which means, in function argument declarations, pointer to char) 933455714Skris to des_cblock * (meaning pointer to array with 8 char elements), 933555714Skris which allows the compiler to do more typechecking; it was like 933655714Skris that back in SSLeay, but with lots of ugly casts. 933755714Skris 933855714Skris Introduce new type const_des_cblock. 933955714Skris [Bodo Moeller] 934055714Skris 934155714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 934255714Skris problems: find RecipientInfo structure that matches recipient certificate 934355714Skris and initialise the ASN1 structures properly based on passed cipher. 934455714Skris [Steve Henson] 934555714Skris 934655714Skris *) Belatedly make the BN tests actually check the results. 934755714Skris [Ben Laurie] 934855714Skris 934955714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 935055714Skris to and from BNs: it was completely broken. New compilation option 935155714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 935255714Skris key elements as negative integers. 935355714Skris [Steve Henson] 935455714Skris 935555714Skris *) Reorganize and speed up MD5. 935655714Skris [Andy Polyakov <appro@fy.chalmers.se>] 935755714Skris 935855714Skris *) VMS support. 935955714Skris [Richard Levitte <richard@levitte.org>] 936055714Skris 936155714Skris *) New option -out to asn1parse to allow the parsed structure to be 936255714Skris output to a file. This is most useful when combined with the -strparse 936355714Skris option to examine the output of things like OCTET STRINGS. 936455714Skris [Steve Henson] 936555714Skris 936655714Skris *) Make SSL library a little more fool-proof by not requiring any longer 936755714Skris that SSL_set_{accept,connect}_state be called before 936855714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 936955714Skris in many applications because usually everything *appeared* to work as 937055714Skris intended anyway -- now it really works as intended). 937155714Skris [Bodo Moeller] 937255714Skris 937355714Skris *) Move openssl.cnf out of lib/. 937455714Skris [Ulf M�ller] 937555714Skris 937655714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 937755714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 937855714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 937955714Skris [Ralf S. Engelschall] 938055714Skris 938155714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 938255714Skris handle PKCS#7 enveloped data properly. 938355714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 938455714Skris 938555714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 938655714Skris copying pointers. The cert_st handling is changed by this in 938755714Skris various ways (and thus what used to be known as ctx->default_cert 938855714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 938955714Skris any longer when s->cert does not give us what we need). 939055714Skris ssl_cert_instantiate becomes obsolete by this change. 939155714Skris As soon as we've got the new code right (possibly it already is?), 939255714Skris we have solved a couple of bugs of the earlier code where s->cert 939355714Skris was used as if it could not have been shared with other SSL structures. 939455714Skris 939555714Skris Note that using the SSL API in certain dirty ways now will result 939655714Skris in different behaviour than observed with earlier library versions: 939755714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 939855714Skris does not influence s as it used to. 939955714Skris 940055714Skris In order to clean up things more thoroughly, inside SSL_SESSION 940155714Skris we don't use CERT any longer, but a new structure SESS_CERT 940255714Skris that holds per-session data (if available); currently, this is 940355714Skris the peer's certificate chain and, for clients, the server's certificate 940455714Skris and temporary key. CERT holds only those values that can have 940555714Skris meaningful defaults in an SSL_CTX. 940655714Skris [Bodo Moeller] 940755714Skris 940855714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 940955714Skris from the internal representation. Various PKCS#7 fixes: remove some 941055714Skris evil casts and set the enc_dig_alg field properly based on the signing 941155714Skris key type. 941255714Skris [Steve Henson] 941355714Skris 941455714Skris *) Allow PKCS#12 password to be set from the command line or the 941555714Skris environment. Let 'ca' get its config file name from the environment 941655714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 941755714Skris and 'x509'). 941855714Skris [Steve Henson] 941955714Skris 942055714Skris *) Allow certificate policies extension to use an IA5STRING for the 942155714Skris organization field. This is contrary to the PKIX definition but 942255714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 942355714Skris extension option. 942455714Skris [Steve Henson] 942555714Skris 942655714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 942755714Skris without disallowing inline assembler and the like for non-pedantic builds. 942855714Skris [Ben Laurie] 942955714Skris 943055714Skris *) Support Borland C++ builder. 943155714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 943255714Skris 943355714Skris *) Support Mingw32. 943455714Skris [Ulf M�ller] 943555714Skris 943655714Skris *) SHA-1 cleanups and performance enhancements. 943755714Skris [Andy Polyakov <appro@fy.chalmers.se>] 943855714Skris 943955714Skris *) Sparc v8plus assembler for the bignum library. 944055714Skris [Andy Polyakov <appro@fy.chalmers.se>] 944155714Skris 944255714Skris *) Accept any -xxx and +xxx compiler options in Configure. 944355714Skris [Ulf M�ller] 944455714Skris 944555714Skris *) Update HPUX configuration. 944655714Skris [Anonymous] 944755714Skris 944855714Skris *) Add missing sk_<type>_unshift() function to safestack.h 944955714Skris [Ralf S. Engelschall] 945055714Skris 945155714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 945255714Skris "extra_cert"s in addition to the certificate. (This makes sense 945355714Skris only for "PEM" format files, as chains as a whole are not 945455714Skris DER-encoded.) 945555714Skris [Bodo Moeller] 945655714Skris 945755714Skris *) Support verify_depth from the SSL API. 945855714Skris x509_vfy.c had what can be considered an off-by-one-error: 945955714Skris Its depth (which was not part of the external interface) 946055714Skris was actually counting the number of certificates in a chain; 946155714Skris now it really counts the depth. 946255714Skris [Bodo Moeller] 946355714Skris 946455714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 946555714Skris instead of X509err, which often resulted in confusing error 946655714Skris messages since the error codes are not globally unique 946755714Skris (e.g. an alleged error in ssl3_accept when a certificate 946855714Skris didn't match the private key). 946955714Skris 947055714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 947155714Skris value (so that you don't need SSL_set_session_id_context for each 947255714Skris connection using the SSL_CTX). 947355714Skris [Bodo Moeller] 947455714Skris 947555714Skris *) OAEP decoding bug fix. 947655714Skris [Ulf M�ller] 947755714Skris 947855714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 947955714Skris David Harris. 948055714Skris [Bodo Moeller] 948155714Skris 948255714Skris *) New Configure options "threads" and "no-threads". For systems 948355714Skris where the proper compiler options are known (currently Solaris 948455714Skris and Linux), "threads" is the default. 948555714Skris [Bodo Moeller] 948655714Skris 948755714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 948855714Skris [Bodo Moeller] 948955714Skris 949055714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 949155714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 949255714Skris such as /usr/local/bin. 949355714Skris [Bodo Moeller] 949455714Skris 949555714Skris *) "make linux-shared" to build shared libraries. 949655714Skris [Niels Poppe <niels@netbox.org>] 949755714Skris 949855714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 949955714Skris [Ulf M�ller] 950055714Skris 950155714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 950255714Skris extension adding in x509 utility. 950355714Skris [Steve Henson] 950455714Skris 950555714Skris *) Remove NOPROTO sections and error code comments. 950655714Skris [Ulf M�ller] 950755714Skris 950855714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 950955714Skris prototypes. 951055714Skris [Steve Henson] 951155714Skris 951255714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 951355714Skris [Ulf M�ller] 951455714Skris 951555714Skris *) Complete rewrite of the error code script(s). It is all now handled 951655714Skris by one script at the top level which handles error code gathering, 951755714Skris header rewriting and C source file generation. It should be much better 951855714Skris than the old method: it now uses a modified version of Ulf's parser to 951955714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 952055714Skris aren't needed for error creation any more) and do a better job of 952155714Skris translating function codes into names. The old 'ASN1 error code imbedded 952255714Skris in a comment' is no longer necessary and it doesn't use .err files which 952355714Skris have now been deleted. Also the error code call doesn't have to appear all 952455714Skris on one line (which resulted in some large lines...). 952555714Skris [Steve Henson] 952655714Skris 952755714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 952855714Skris [Bodo Moeller] 952955714Skris 953055714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 953155714Skris 0 (which usually indicates a closed connection), but continue reading. 953255714Skris [Bodo Moeller] 953355714Skris 953455714Skris *) Fix some race conditions. 953555714Skris [Bodo Moeller] 953655714Skris 953755714Skris *) Add support for CRL distribution points extension. Add Certificate 953855714Skris Policies and CRL distribution points documentation. 953955714Skris [Steve Henson] 954055714Skris 954155714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 954255714Skris [Ulf M�ller] 954355714Skris 954455714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 954555714Skris 8 of keying material. Merlin has also confirmed interop with this fix 954655714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 954755714Skris [Merlin Hughes <merlin@baltimore.ie>] 954855714Skris 954955714Skris *) Fix lots of warnings. 955055714Skris [Richard Levitte <levitte@stacken.kth.se>] 955155714Skris 955255714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 955355714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 955455714Skris [Richard Levitte <levitte@stacken.kth.se>] 955555714Skris 955655714Skris *) Fix problems with sizeof(long) == 8. 955755714Skris [Andy Polyakov <appro@fy.chalmers.se>] 955855714Skris 955955714Skris *) Change functions to ANSI C. 956055714Skris [Ulf M�ller] 956155714Skris 956255714Skris *) Fix typos in error codes. 956355714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 956455714Skris 956555714Skris *) Remove defunct assembler files from Configure. 956655714Skris [Ulf M�ller] 956755714Skris 956855714Skris *) SPARC v8 assembler BIGNUM implementation. 956955714Skris [Andy Polyakov <appro@fy.chalmers.se>] 957055714Skris 957155714Skris *) Support for Certificate Policies extension: both print and set. 957255714Skris Various additions to support the r2i method this uses. 957355714Skris [Steve Henson] 957455714Skris 957555714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 957655714Skris return a const string when you are expecting an allocated buffer. 957755714Skris [Ben Laurie] 957855714Skris 957955714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 958055714Skris types DirectoryString and DisplayText. 958155714Skris [Steve Henson] 958255714Skris 958355714Skris *) Add code to allow r2i extensions to access the configuration database, 958455714Skris add an LHASH database driver and add several ctx helper functions. 958555714Skris [Steve Henson] 958655714Skris 958755714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 958855714Skris fail when they extended the size of a BIGNUM. 958955714Skris [Steve Henson] 959055714Skris 959155714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 959255714Skris support typesafe stack. 959355714Skris [Steve Henson] 959455714Skris 959555714Skris *) Fix typo in SSL_[gs]et_options(). 959655714Skris [Nils Frostberg <nils@medcom.se>] 959755714Skris 959855714Skris *) Delete various functions and files that belonged to the (now obsolete) 959955714Skris old X509V3 handling code. 960055714Skris [Steve Henson] 960155714Skris 960255714Skris *) New Configure option "rsaref". 960355714Skris [Ulf M�ller] 960455714Skris 960555714Skris *) Don't auto-generate pem.h. 960655714Skris [Bodo Moeller] 960755714Skris 960855714Skris *) Introduce type-safe ASN.1 SETs. 960955714Skris [Ben Laurie] 961055714Skris 961155714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 961255714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 961355714Skris 961455714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 961555714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 961655714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 961755714Skris few STACKed things have been converted already. Feel free to convert more. 961855714Skris In the fullness of time, I'll do away with the STACK type altogether. 961955714Skris [Ben Laurie] 962055714Skris 962155714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 962255714Skris specified in <certfile> by updating the entry in the index.txt file. 962355714Skris This way one no longer has to edit the index.txt file manually for 962455714Skris revoking a certificate. The -revoke option does the gory details now. 962555714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 962655714Skris 962755714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 962855714Skris `-text' option at all and this way the `-noout -text' combination was 962955714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 963055714Skris [Ralf S. Engelschall] 963155714Skris 963255714Skris *) Make sure a corresponding plain text error message exists for the 963355714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 963455714Skris verify callback function determined that a certificate was revoked. 963555714Skris [Ralf S. Engelschall] 963655714Skris 963755714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 963855714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 963955714Skris all available cipers including rc5, which was forgotten until now. 964055714Skris In order to let the testing shell script know which algorithms 964155714Skris are available, a new (up to now undocumented) command 964255714Skris "openssl list-cipher-commands" is used. 964355714Skris [Bodo Moeller] 964455714Skris 964555714Skris *) Bugfix: s_client occasionally would sleep in select() when 964655714Skris it should have checked SSL_pending() first. 964755714Skris [Bodo Moeller] 964855714Skris 964955714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 965055714Skris the raw DSA values prior to ASN.1 encoding. 965155714Skris [Ulf M�ller] 965255714Skris 965355714Skris *) Tweaks to Configure 965455714Skris [Niels Poppe <niels@netbox.org>] 965555714Skris 965655714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 965755714Skris yet... 965855714Skris [Steve Henson] 965955714Skris 966055714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 966155714Skris [Ulf M�ller] 966255714Skris 966355714Skris *) New config option to avoid instructions that are illegal on the 80386. 966455714Skris The default code is faster, but requires at least a 486. 966555714Skris [Ulf M�ller] 966655714Skris 966755714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 966855714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 966955714Skris same as SSL2_VERSION anyway. 967055714Skris [Bodo Moeller] 967155714Skris 967255714Skris *) New "-showcerts" option for s_client. 967355714Skris [Bodo Moeller] 967455714Skris 967555714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 967655714Skris application. Various cleanups and fixes. 967755714Skris [Steve Henson] 967855714Skris 967955714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 968055714Skris modify error routines to work internally. Add error codes and PBE init 968155714Skris to library startup routines. 968255714Skris [Steve Henson] 968355714Skris 968455714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 968555714Skris packing functions to asn1 and evp. Changed function names and error 968655714Skris codes along the way. 968755714Skris [Steve Henson] 968855714Skris 968955714Skris *) PKCS12 integration: and so it begins... First of several patches to 969055714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 969155714Skris objects to objects.h 969255714Skris [Steve Henson] 969355714Skris 969455714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 969555714Skris and display support for Thawte strong extranet extension. 969655714Skris [Steve Henson] 969755714Skris 969855714Skris *) Add LinuxPPC support. 969955714Skris [Jeff Dubrule <igor@pobox.org>] 970055714Skris 970155714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 970255714Skris bn_div_words in alpha.s. 970355714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 970455714Skris 970555714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 970655714Skris OAEP isn't supported when OpenSSL is built with RSAref. 970755714Skris [Ulf Moeller <ulf@fitug.de>] 970855714Skris 970955714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 971055714Skris so they no longer are missing under -DNOPROTO. 971155714Skris [Soren S. Jorvang <soren@t.dk>] 971255714Skris 971355714Skris 971455714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 971555714Skris 971655714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 971755714Skris doesn't work when the session is reused. Coming soon! 971855714Skris [Ben Laurie] 971955714Skris 972055714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 972155714Skris context thus bypassing client cert protection! All software that uses 972255714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 972355714Skris allow session reuse! A fuller solution is in the works. 972455714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 972555714Skris 972655714Skris *) Some more source tree cleanups (removed obsolete files 972755714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 972855714Skris permission on "config" script to be executable) and a fix for the INSTALL 972955714Skris document. 973055714Skris [Ulf Moeller <ulf@fitug.de>] 973155714Skris 973255714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 973355714Skris Malloc, Free. 973455714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 973555714Skris 973655714Skris *) Make rsa_oaep_test return non-zero on error. 973755714Skris [Ulf Moeller <ulf@fitug.de>] 973855714Skris 973955714Skris *) Add support for native Solaris shared libraries. Configure 974055714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 974155714Skris if someone would make that last step automatic. 974255714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 974355714Skris 974455714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 974555714Skris [Ben Laurie] 974655714Skris 974755714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 974855714Skris except NULL ciphers". This means the default cipher list will no longer 974955714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 975055714Skris the string "DEFAULT:eNULL". 975155714Skris [Steve Henson] 975255714Skris 975355714Skris *) Fix to RSA private encryption routines: if p < q then it would 975455714Skris occasionally produce an invalid result. This will only happen with 975555714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 975655714Skris [Steve Henson] 975755714Skris 975855714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 975955714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 976055714Skris because this way one can also use an interpreter named `perl5' (which is 976155714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 976255714Skris installed as `perl'). 976355714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 976455714Skris 976555714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 976655714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 976755714Skris 976855714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 976955714Skris advapi32.lib to Win32 build and change the pem test comparision 977055714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 977155714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 977255714Skris and crypto/des/ede_cbcm_enc.c. 977355714Skris [Steve Henson] 977455714Skris 977555714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 977655714Skris [Ben Laurie] 977755714Skris 977855714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 977955714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 978055714Skris is horrible: I feel ill.... 978155714Skris [Steve Henson] 978255714Skris 978355714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 978455714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 978555714Skris sections: 10 functions were absent from non ANSI section and not exported 978655714Skris from Windows DLLs. Fixed up libeay.num for new functions. 978755714Skris [Steve Henson] 978855714Skris 978955714Skris *) Make `openssl version' output lines consistent. 979055714Skris [Ralf S. Engelschall] 979155714Skris 979255714Skris *) Fix Win32 symbol export lists for BIO functions: Added 979355714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 979455714Skris to ms/libeay{16,32}.def. 979555714Skris [Ralf S. Engelschall] 979655714Skris 979755714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 979855714Skris fine under Unix and passes some trivial tests I've now added. But the 979955714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 980055714Skris added to make sure no one expects that this stuff really works in the 980155714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 980255714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 980355714Skris openssl_bio.xs. 980455714Skris [Ralf S. Engelschall] 980555714Skris 980655714Skris *) Fix the generation of two part addresses in perl. 980755714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 980855714Skris 980955714Skris *) Add config entry for Linux on MIPS. 981055714Skris [John Tobey <jtobey@channel1.com>] 981155714Skris 981255714Skris *) Make links whenever Configure is run, unless we are on Windoze. 981355714Skris [Ben Laurie] 981455714Skris 981555714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 981655714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 981755714Skris in CRLs. 981855714Skris [Steve Henson] 981955714Skris 982055714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 982155714Skris other platforms details on the command line without having to patch the 982255714Skris Configure script everytime: One now can use ``perl Configure 982355714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 982455714Skris to them (seperated by colons). This is treated as there would be a static 982555714Skris pre-configured entry in Configure's %table under key <id> with value 982655714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 982755714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 982855714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 982955714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 983055714Skris [Ralf S. Engelschall] 983155714Skris 983255714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 983355714Skris [Ben Laurie] 983455714Skris 983555714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 983655714Skris on the `perl Configure ...' command line. This way one can compile 983755714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 983855714Skris for linking it into DSOs. 983955714Skris [Ralf S. Engelschall] 984055714Skris 984155714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 984255714Skris Fixed. 984355714Skris [Ben Laurie] 984455714Skris 984555714Skris *) Cleaned up the LICENSE document: The official contact for any license 984655714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 984755714Skris And add a paragraph about the dual-license situation to make sure people 984855714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 984955714Skris to the OpenSSL toolkit. 985055714Skris [Ralf S. Engelschall] 985155714Skris 985255714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 985355714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 985455714Skris Additonally cleaned up the `make links' target: Remove unnecessary 985555714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 985655714Skris to speed processing and no longer clutter the display with confusing 985755714Skris stuff. Instead only the actually done links are displayed. 985855714Skris [Ralf S. Engelschall] 985955714Skris 986055714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 986155714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 986255714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 986355714Skris encryption. 986455714Skris [Ben Laurie] 986555714Skris 986655714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 986755714Skris signed attributes when verifying signatures (this would break them), 986855714Skris the detached data encoding was wrong and public keys obtained using 986955714Skris X509_get_pubkey() weren't freed. 987055714Skris [Steve Henson] 987155714Skris 987255714Skris *) Add text documentation for the BUFFER functions. Also added a work around 987355714Skris to a Win95 console bug. This was triggered by the password read stuff: the 987455714Skris last character typed gets carried over to the next fread(). If you were 987555714Skris generating a new cert request using 'req' for example then the last 987655714Skris character of the passphrase would be CR which would then enter the first 987755714Skris field as blank. 987855714Skris [Steve Henson] 987955714Skris 988055714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 988155714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 988255714Skris button and can be used by applications based on OpenSSL to show the 988355714Skris relationship to the OpenSSL project. 988455714Skris [Ralf S. Engelschall] 988555714Skris 988655714Skris *) Remove confusing variables in function signatures in files 988755714Skris ssl/ssl_lib.c and ssl/ssl.h. 988855714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 988955714Skris 989055714Skris *) Don't install bss_file.c under PREFIX/include/ 989155714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 989255714Skris 989355714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 989455714Skris functions that return function pointers and has support for NT specific 989555714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 989655714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 989755714Skris unsigned to signed types: this was killing the Win32 compile. 989855714Skris [Steve Henson] 989955714Skris 990055714Skris *) Add new certificate file to stack functions, 990155714Skris SSL_add_dir_cert_subjects_to_stack() and 990255714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 990355714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 990455714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 990555714Skris This means that Apache-SSL and similar packages don't have to mess around 990655714Skris to add as many CAs as they want to the preferred list. 990755714Skris [Ben Laurie] 990855714Skris 990955714Skris *) Experiment with doxygen documentation. Currently only partially applied to 991055714Skris ssl/ssl_lib.c. 991155714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 991255714Skris openssl.doxy as the configuration file. 991355714Skris [Ben Laurie] 991455714Skris 991555714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 991655714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 991755714Skris 991855714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 991955714Skris compiled in by default: it has problems with large keys. 992055714Skris [Steve Henson] 992155714Skris 992255714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 992355714Skris DH private keys and/or callback functions which directly correspond to 992455714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 992555714Skris is needed for applications which have to configure certificates on a 992655714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 992755714Skris (e.g. s_server). 992855714Skris For the RSA certificate situation is makes no difference, but 992955714Skris for the DSA certificate situation this fixes the "no shared cipher" 993055714Skris problem where the OpenSSL cipher selection procedure failed because the 993155714Skris temporary keys were not overtaken from the context and the API provided 993255714Skris no way to reconfigure them. 993355714Skris The new functions now let applications reconfigure the stuff and they 993455714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 993555714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 993655714Skris non-public-API function ssl_cert_instantiate() is used as a helper 993755714Skris function and also to reduce code redundancy inside ssl_rsa.c. 993855714Skris [Ralf S. Engelschall] 993955714Skris 994055714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 994155714Skris area because they are useful for the DSA situation and should be 994255714Skris recognized by the users. 994355714Skris [Ralf S. Engelschall] 994455714Skris 994555714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 994655714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 994755714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 994855714Skris already masked variable. 994955714Skris [Richard Levitte <levitte@stacken.kth.se>] 995055714Skris 995155714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 995255714Skris [Richard Levitte <levitte@stacken.kth.se>] 995355714Skris 995455714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 995555714Skris from `int' to `unsigned int' because it's a length and initialized by 995655714Skris EVP_DigestFinal() which expects an `unsigned int *'. 995755714Skris [Richard Levitte <levitte@stacken.kth.se>] 995855714Skris 995955714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 996055714Skris script. Instead use the usual Shell->Perl transition trick. 996155714Skris [Ralf S. Engelschall] 996255714Skris 996355714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 996455714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 996555714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 996655714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 996755714Skris currently the public key is printed (a decision which was already done by 996855714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 996955714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 997055714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 997155714Skris now, too. 997255714Skris [Ralf S. Engelschall] 997355714Skris 997455714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 997555714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 997655714Skris [Arne Ansper <arne@ats.cyber.ee>] 997755714Skris 997855714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 997955714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 998055714Skris config file. 998155714Skris [Steve Henson] 998255714Skris 998355714Skris *) Add cool BIO that does syslog (or event log on NT). 998455714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 998555714Skris 998655714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 998755714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 998855714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 998955714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 999055714Skris [Ben Laurie] 999155714Skris 999255714Skris *) Add preliminary config info for new extension code. 999355714Skris [Steve Henson] 999455714Skris 999555714Skris *) Make RSA_NO_PADDING really use no padding. 999655714Skris [Ulf Moeller <ulf@fitug.de>] 999755714Skris 999855714Skris *) Generate errors when private/public key check is done. 999955714Skris [Ben Laurie] 1000055714Skris 1000155714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 1000255714Skris for some CRL extensions and new objects added. 1000355714Skris [Steve Henson] 1000455714Skris 1000555714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 1000655714Skris key usage extension and fuller support for authority key id. 1000755714Skris [Steve Henson] 1000855714Skris 1000955714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 1001055714Skris padding method for RSA, which is recommended for new applications in PKCS 1001155714Skris #1 v2.0 (RFC 2437, October 1998). 1001255714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 1001355714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 1001455714Skris against Bleichbacher's attack on RSA. 1001555714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 1001655714Skris Ben Laurie] 1001755714Skris 1001855714Skris *) Updates to the new SSL compression code 1001955714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 1002055714Skris 1002155714Skris *) Fix so that the version number in the master secret, when passed 1002255714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 1002355714Skris (because the server will not accept higher), that the version number 1002455714Skris is 0x03,0x01, not 0x03,0x00 1002555714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 1002655714Skris 1002755714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 1002855714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 1002955714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 1003055714Skris [Steve Henson] 1003155714Skris 1003255714Skris *) Support for RAW extensions where an arbitrary extension can be 1003355714Skris created by including its DER encoding. See apps/openssl.cnf for 1003455714Skris an example. 1003555714Skris [Steve Henson] 1003655714Skris 1003755714Skris *) Make sure latest Perl versions don't interpret some generated C array 1003855714Skris code as Perl array code in the crypto/err/err_genc.pl script. 1003955714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 1004055714Skris 1004155714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 1004255714Skris not many people have the assembler. Various Win32 compilation fixes and 1004355714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 1004455714Skris build instructions. 1004555714Skris [Steve Henson] 1004655714Skris 1004755714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 1004855714Skris file under Win32 and also build pem.h from pem.org. New script 1004955714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 1005055714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 1005155714Skris [Steve Henson] 1005255714Skris 1005355714Skris *) Major rework of DES function declarations, in the pursuit of correctness 1005455714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 1005555714Skris too. You may find this causes warnings in your code. Zapping your evil 1005655714Skris casts will probably fix them. Mostly. 1005755714Skris [Ben Laurie] 1005855714Skris 1005955714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 1006055714Skris obj_dat.pl. It considered a zero in an object definition to mean 1006155714Skris "end of object": none of the objects in objects.h have any zeros 1006255714Skris so it wasn't spotted. 1006355714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 1006455714Skris 1006555714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 1006655714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 1006755714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 1006855714Skris vectors if you have them. 1006955714Skris [Ben Laurie] 1007055714Skris 1007155714Skris *) Correct calculation of key length for export ciphers (too much space was 1007255714Skris allocated for null ciphers). This has not been tested! 1007355714Skris [Ben Laurie] 1007455714Skris 1007555714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 1007655714Skris message is now correct (it understands "crypto" and "ssl" on its 1007755714Skris command line). There is also now an "update" option. This will update 1007855714Skris the util/ssleay.num and util/libeay.num files with any new functions. 1007955714Skris If you do a: 1008055714Skris perl util/mkdef.pl crypto ssl update 1008155714Skris it will update them. 1008255714Skris [Steve Henson] 1008355714Skris 1008455714Skris *) Overhauled the Perl interface (perl/*): 1008555714Skris - ported BN stuff to OpenSSL's different BN library 1008655714Skris - made the perl/ source tree CVS-aware 1008755714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 1008855714Skris their history because I've copied them in the repository) 1008955714Skris - removed obsolete files (the test scripts will be replaced 1009055714Skris by better Test::Harness variants in the future) 1009155714Skris [Ralf S. Engelschall] 1009255714Skris 1009355714Skris *) First cut for a very conservative source tree cleanup: 1009455714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 1009555714Skris where we collect the old documents and readme texts. 1009655714Skris 2. remove the first part of files where I'm already sure that we no 1009755714Skris longer need them because of three reasons: either they are just temporary 1009855714Skris files which were left by Eric or they are preserved original files where 1009955714Skris I've verified that the diff is also available in the CVS via "cvs diff 1010055714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 1010155714Skris the crypto/md/ stuff). 1010255714Skris [Ralf S. Engelschall] 1010355714Skris 1010455714Skris *) More extension code. Incomplete support for subject and issuer alt 1010555714Skris name, issuer and authority key id. Change the i2v function parameters 1010655714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 1010755714Skris what that's for :-) Fix to ASN1 macro which messed up 1010855714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 1010955714Skris [Steve Henson] 1011055714Skris 1011155714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 1011255714Skris INTEGER code. 1011355714Skris [Steve Henson] 1011455714Skris 1011555714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 1011655714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 1011755714Skris 1011855714Skris *) Make sure `make rehash' target really finds the `openssl' program. 1011955714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 1012055714Skris 1012155714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 1012255714Skris like to hear about it if this slows down other processors. 1012355714Skris [Ben Laurie] 1012455714Skris 1012555714Skris *) Add CygWin32 platform information to Configure script. 1012655714Skris [Alan Batie <batie@aahz.jf.intel.com>] 1012755714Skris 1012855714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 1012955714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 1013055714Skris 1013155714Skris *) New program nseq to manipulate netscape certificate sequences 1013255714Skris [Steve Henson] 1013355714Skris 1013455714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 1013555714Skris few typos. 1013655714Skris [Steve Henson] 1013755714Skris 1013855714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 1013955714Skris but the BN code had some problems that would cause failures when 1014055714Skris doing certificate verification and some other functions. 1014155714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 1014255714Skris 1014355714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 1014455714Skris [Steve Henson] 1014555714Skris 1014655714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 1014755714Skris [Steve Henson] 1014855714Skris 1014955714Skris *) Add several PKIX and private extended key usage OIDs. 1015055714Skris [Steve Henson] 1015155714Skris 1015255714Skris *) Modify the 'ca' program to handle the new extension code. Modify 1015355714Skris openssl.cnf for new extension format, add comments. 1015455714Skris [Steve Henson] 1015555714Skris 1015655714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 1015755714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 1015855714Skris CA extensions. 1015955714Skris [Steve Henson] 1016055714Skris 1016155714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 1016255714Skris error code, add initial support to X509_print() and x509 application. 1016355714Skris [Steve Henson] 1016455714Skris 1016555714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 1016655714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 1016755714Skris stuff is currently isolated and isn't even compiled yet. 1016855714Skris [Steve Henson] 1016955714Skris 1017055714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 1017155714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 1017255714Skris Removed the versions check from X509 routines when loading extensions: 1017355714Skris this allows certain broken certificates that don't set the version 1017455714Skris properly to be processed. 1017555714Skris [Steve Henson] 1017655714Skris 1017755714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 1017855714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 1017955714Skris can still be regenerated with "make depend". 1018055714Skris [Ben Laurie] 1018155714Skris 1018255714Skris *) Spelling mistake in C version of CAST-128. 1018355714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 1018455714Skris 1018555714Skris *) Changes to the error generation code. The perl script err-code.pl 1018655714Skris now reads in the old error codes and retains the old numbers, only 1018755714Skris adding new ones if necessary. It also only changes the .err files if new 1018855714Skris codes are added. The makefiles have been modified to only insert errors 1018955714Skris when needed (to avoid needlessly modifying header files). This is done 1019055714Skris by only inserting errors if the .err file is newer than the auto generated 1019155714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 1019255714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 1019355714Skris or delete all the .err files. 1019455714Skris [Steve Henson] 1019555714Skris 1019655714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 1019755714Skris been fixed, but is untested. The assembler versions are also fixed, but 1019855714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 1019955714Skris to regenerate it if needed. 1020055714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 1020155714Skris Hagino <itojun@kame.net>] 1020255714Skris 1020355714Skris *) File was opened incorrectly in randfile.c. 1020455714Skris [Ulf M�ller <ulf@fitug.de>] 1020555714Skris 1020655714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 1020755714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 1020855714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 1020955714Skris al: it's just almost always a UTCTime. Note this patch adds new error 1021055714Skris codes so do a "make errors" if there are problems. 1021155714Skris [Steve Henson] 1021255714Skris 1021355714Skris *) Correct Linux 1 recognition in config. 1021455714Skris [Ulf M�ller <ulf@fitug.de>] 1021555714Skris 1021655714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 1021755714Skris [Anonymous <nobody@replay.com>] 1021855714Skris 1021955714Skris *) Generate an error if given an empty string as a cert directory. Also 1022055714Skris generate an error if handed NULL (previously returned 0 to indicate an 1022155714Skris error, but didn't set one). 1022255714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 1022355714Skris 1022455714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 1022555714Skris [Ben Laurie] 1022655714Skris 1022755714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 1022855714Skris parameters. This was causing a warning which killed off the Win32 compile. 1022955714Skris [Steve Henson] 1023055714Skris 1023155714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 1023255714Skris [Neil Costigan <neil.costigan@celocom.com>] 1023355714Skris 1023455714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 1023555714Skris based on a text string, looking up short and long names and finally 1023655714Skris "dot" format. The "dot" format stuff didn't work. Added new function 1023755714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 1023855714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 1023955714Skris OID is not part of the table. 1024055714Skris [Steve Henson] 1024155714Skris 1024255714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 1024355714Skris X509_LOOKUP_by_alias(). 1024455714Skris [Ben Laurie] 1024555714Skris 1024655714Skris *) Sort openssl functions by name. 1024755714Skris [Ben Laurie] 1024855714Skris 1024955714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 1025055714Skris encryption from sample DSA keys (in case anyone is interested the password 1025155714Skris was "1234"). 1025255714Skris [Steve Henson] 1025355714Skris 1025455714Skris *) Make _all_ *_free functions accept a NULL pointer. 1025555714Skris [Frans Heymans <fheymans@isaserver.be>] 1025655714Skris 1025755714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 1025855714Skris NULL pointers. 1025955714Skris [Anonymous <nobody@replay.com>] 1026055714Skris 1026155714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 1026255714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 1026355714Skris 1026455714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 1026555714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 1026655714Skris 1026755714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 1026855714Skris [Anonymous <nobody@replay.com>] 1026955714Skris 1027055714Skris *) Add prototype for temp key callback functions 1027155714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 1027255714Skris [Ben Laurie] 1027355714Skris 1027455714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 1027555714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 1027655714Skris [Steve Henson] 1027755714Skris 1027855714Skris *) X509_name_add_entry() freed the wrong thing after an error. 1027955714Skris [Arne Ansper <arne@ats.cyber.ee>] 1028055714Skris 1028155714Skris *) rsa_eay.c would attempt to free a NULL context. 1028255714Skris [Arne Ansper <arne@ats.cyber.ee>] 1028355714Skris 1028455714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 1028555714Skris [Arne Ansper <arne@ats.cyber.ee>] 1028655714Skris 1028755714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 1028855714Skris [Arne Ansper <arne@ats.cyber.ee>] 1028955714Skris 1029055714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 1029155714Skris in X509_STORE_new(), but document the fact that this variable is still 1029255714Skris unused in the certificate verification process. 1029355714Skris [Ralf S. Engelschall] 1029455714Skris 1029555714Skris *) Fix the various library and apps files to free up pkeys obtained from 1029655714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 1029755714Skris [Steve Henson] 1029855714Skris 1029955714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 1030055714Skris demos/maurice/example2.c work, amongst others, probably. 1030155714Skris [Steve Henson and Ben Laurie] 1030255714Skris 1030355714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 1030455714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 1030555714Skris are no longer created. This way we have a single and consistent command 1030655714Skris line interface `openssl <command>', similar to `cvs <command>'. 1030755714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 1030855714Skris 1030955714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 1031055714Skris BIT STRING wrapper always have zero unused bits. 1031155714Skris [Steve Henson] 1031255714Skris 1031355714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 1031455714Skris [Steve Henson] 1031555714Skris 1031655714Skris *) Make the top-level INSTALL documentation easier to understand. 1031755714Skris [Paul Sutton] 1031855714Skris 1031955714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 1032055714Skris make (including if user presses ^C) [Paul Sutton] 1032155714Skris 1032255714Skris *) Make Montgomery context stuff explicit in RSA data structure. 1032355714Skris [Ben Laurie] 1032455714Skris 1032555714Skris *) Fix build order of pem and err to allow for generated pem.h. 1032655714Skris [Ben Laurie] 1032755714Skris 1032855714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 1032955714Skris [Ben Laurie] 1033055714Skris 1033155714Skris *) Enhanced the err-ins.pl script so it makes the error library number 1033255714Skris global and can add a library name. This is needed for external ASN1 and 1033355714Skris other error libraries. 1033455714Skris [Steve Henson] 1033555714Skris 1033655714Skris *) Fixed sk_insert which never worked properly. 1033755714Skris [Steve Henson] 1033855714Skris 1033955714Skris *) Fix ASN1 macros so they can handle indefinite length construted 1034055714Skris EXPLICIT tags. Some non standard certificates use these: they can now 1034155714Skris be read in. 1034255714Skris [Steve Henson] 1034355714Skris 1034455714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 1034555714Skris into a single doc/ssleay.txt bundle. This way the information is still 1034655714Skris preserved but no longer messes up this directory. Now it's new room for 1034755714Skris the new set of documenation files. 1034855714Skris [Ralf S. Engelschall] 1034955714Skris 1035055714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 1035155714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 1035255714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 1035355714Skris number of arguments. 1035455714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 1035555714Skris 1035655714Skris *) Fix test data to work with the above. 1035755714Skris [Ben Laurie] 1035855714Skris 1035955714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 1036055714Skris was already fixed by Eric for 0.9.1 it seems. 1036155714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 1036255714Skris 1036355714Skris *) Autodetect FreeBSD3. 1036455714Skris [Ben Laurie] 1036555714Skris 1036655714Skris *) Fix various bugs in Configure. This affects the following platforms: 1036755714Skris nextstep 1036855714Skris ncr-scde 1036955714Skris unixware-2.0 1037055714Skris unixware-2.0-pentium 1037155714Skris sco5-cc. 1037255714Skris [Ben Laurie] 1037355714Skris 1037455714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 1037555714Skris before they are needed. 1037655714Skris [Ben Laurie] 1037755714Skris 1037855714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 1037955714Skris [Ben Laurie] 1038055714Skris 1038155714Skris 1038255714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 1038355714Skris 1038455714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 1038555714Skris changed SSLeay to OpenSSL in version strings. 1038655714Skris [Ralf S. Engelschall] 1038755714Skris 1038855714Skris *) Some fixups to the top-level documents. 1038955714Skris [Paul Sutton] 1039055714Skris 1039155714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 1039255714Skris because the symlink to include/ was missing. 1039355714Skris [Ralf S. Engelschall] 1039455714Skris 1039555714Skris *) Incorporated the popular no-RSA/DSA-only patches 1039655714Skris which allow to compile a RSA-free SSLeay. 1039755714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 1039855714Skris 1039955714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 1040055714Skris when "ssleay" is still not found. 1040155714Skris [Ralf S. Engelschall] 1040255714Skris 1040355714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 1040455714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 1040555714Skris 1040655714Skris *) Updated the README file. 1040755714Skris [Ralf S. Engelschall] 1040855714Skris 1040955714Skris *) Added various .cvsignore files in the CVS repository subdirs 1041055714Skris to make a "cvs update" really silent. 1041155714Skris [Ralf S. Engelschall] 1041255714Skris 1041355714Skris *) Recompiled the error-definition header files and added 1041455714Skris missing symbols to the Win32 linker tables. 1041555714Skris [Ralf S. Engelschall] 1041655714Skris 1041755714Skris *) Cleaned up the top-level documents; 1041855714Skris o new files: CHANGES and LICENSE 1041955714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 1042055714Skris o merged COPYRIGHT into LICENSE 1042155714Skris o removed obsolete TODO file 1042255714Skris o renamed MICROSOFT to INSTALL.W32 1042355714Skris [Ralf S. Engelschall] 1042455714Skris 1042555714Skris *) Removed dummy files from the 0.9.1b source tree: 1042655714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 1042755714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 1042855714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 1042955714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 1043055714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 1043155714Skris [Ralf S. Engelschall] 1043255714Skris 1043355714Skris *) Added various platform portability fixes. 1043455714Skris [Mark J. Cox] 1043555714Skris 1043655714Skris *) The Genesis of the OpenSSL rpject: 1043755714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 1043855714Skris Young and Tim J. Hudson created while they were working for C2Net until 1043955714Skris summer 1998. 1044055714Skris [The OpenSSL Project] 1044155714Skris 1044255714Skris 1044355714Skris Changes between 0.9.0b and 0.9.1b [not released] 1044455714Skris 1044555714Skris *) Updated a few CA certificates under certs/ 1044655714Skris [Eric A. Young] 1044755714Skris 1044855714Skris *) Changed some BIGNUM api stuff. 1044955714Skris [Eric A. Young] 1045055714Skris 1045155714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 1045255714Skris DGUX x86, Linux Alpha, etc. 1045355714Skris [Eric A. Young] 1045455714Skris 1045555714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 1045655714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 1045755714Skris available). 1045855714Skris [Eric A. Young] 1045955714Skris 1046055714Skris *) Add -strparse option to asn1pars program which parses nested 1046155714Skris binary structures 1046255714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 1046355714Skris 1046455714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 1046555714Skris [Eric A. Young] 1046655714Skris 1046755714Skris *) DSA fix for "ca" program. 1046855714Skris [Eric A. Young] 1046955714Skris 1047055714Skris *) Added "-genkey" option to "dsaparam" program. 1047155714Skris [Eric A. Young] 1047255714Skris 1047355714Skris *) Added RIPE MD160 (rmd160) message digest. 1047455714Skris [Eric A. Young] 1047555714Skris 1047655714Skris *) Added -a (all) option to "ssleay version" command. 1047755714Skris [Eric A. Young] 1047855714Skris 1047955714Skris *) Added PLATFORM define which is the id given to Configure. 1048055714Skris [Eric A. Young] 1048155714Skris 1048255714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 1048355714Skris [Eric A. Young] 1048455714Skris 1048555714Skris *) Extended the ASN.1 parser routines. 1048655714Skris [Eric A. Young] 1048755714Skris 1048855714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 1048955714Skris [Eric A. Young] 1049055714Skris 1049155714Skris *) Added a BN_CTX to the BN library. 1049255714Skris [Eric A. Young] 1049355714Skris 1049455714Skris *) Fixed the weak key values in DES library 1049555714Skris [Eric A. Young] 1049655714Skris 1049755714Skris *) Changed API in EVP library for cipher aliases. 1049855714Skris [Eric A. Young] 1049955714Skris 1050055714Skris *) Added support for RC2/64bit cipher. 1050155714Skris [Eric A. Young] 1050255714Skris 1050355714Skris *) Converted the lhash library to the crypto/mem.c functions. 1050455714Skris [Eric A. Young] 1050555714Skris 1050655714Skris *) Added more recognized ASN.1 object ids. 1050755714Skris [Eric A. Young] 1050855714Skris 1050955714Skris *) Added more RSA padding checks for SSL/TLS. 1051055714Skris [Eric A. Young] 1051155714Skris 1051255714Skris *) Added BIO proxy/filter functionality. 1051355714Skris [Eric A. Young] 1051455714Skris 1051555714Skris *) Added extra_certs to SSL_CTX which can be used 1051655714Skris send extra CA certificates to the client in the CA cert chain sending 1051755714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 1051855714Skris [Eric A. Young] 1051955714Skris 1052055714Skris *) Now Fortezza is denied in the authentication phase because 1052155714Skris this is key exchange mechanism is not supported by SSLeay at all. 1052255714Skris [Eric A. Young] 1052355714Skris 1052455714Skris *) Additional PKCS1 checks. 1052555714Skris [Eric A. Young] 1052655714Skris 1052755714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 1052855714Skris [Eric A. Young] 1052955714Skris 1053055714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 1053155714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 1053255714Skris [Eric A. Young] 1053355714Skris 1053455714Skris *) Fixed a few memory leaks. 1053555714Skris [Eric A. Young] 1053655714Skris 1053755714Skris *) Fixed various code and comment typos. 1053855714Skris [Eric A. Young] 1053955714Skris 1054055714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 1054155714Skris bytes sent in the client random. 1054255714Skris [Edward Bishop <ebishop@spyglass.com>] 1054355714Skris 10544