CHANGES revision 280304
155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 5280304Sjkim Changes between 1.0.1l and 1.0.1m [19 Mar 2015] 6280304Sjkim 7280304Sjkim *) Segmentation fault in ASN1_TYPE_cmp fix 8280304Sjkim 9280304Sjkim The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is 10280304Sjkim made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check 11280304Sjkim certificate signature algorithm consistency this can be used to crash any 12280304Sjkim certificate verification operation and exploited in a DoS attack. Any 13280304Sjkim application which performs certificate verification is vulnerable including 14280304Sjkim OpenSSL clients and servers which enable client authentication. 15280304Sjkim (CVE-2015-0286) 16280304Sjkim [Stephen Henson] 17280304Sjkim 18280304Sjkim *) ASN.1 structure reuse memory corruption fix 19280304Sjkim 20280304Sjkim Reusing a structure in ASN.1 parsing may allow an attacker to cause 21280304Sjkim memory corruption via an invalid write. Such reuse is and has been 22280304Sjkim strongly discouraged and is believed to be rare. 23280304Sjkim 24280304Sjkim Applications that parse structures containing CHOICE or ANY DEFINED BY 25280304Sjkim components may be affected. Certificate parsing (d2i_X509 and related 26280304Sjkim functions) are however not affected. OpenSSL clients and servers are 27280304Sjkim not affected. 28280304Sjkim (CVE-2015-0287) 29280304Sjkim [Stephen Henson] 30280304Sjkim 31280304Sjkim *) PKCS7 NULL pointer dereferences fix 32280304Sjkim 33280304Sjkim The PKCS#7 parsing code does not handle missing outer ContentInfo 34280304Sjkim correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with 35280304Sjkim missing content and trigger a NULL pointer dereference on parsing. 36280304Sjkim 37280304Sjkim Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or 38280304Sjkim otherwise parse PKCS#7 structures from untrusted sources are 39280304Sjkim affected. OpenSSL clients and servers are not affected. 40280304Sjkim 41280304Sjkim This issue was reported to OpenSSL by Michal Zalewski (Google). 42280304Sjkim (CVE-2015-0289) 43280304Sjkim [Emilia K�sper] 44280304Sjkim 45280304Sjkim *) DoS via reachable assert in SSLv2 servers fix 46280304Sjkim 47280304Sjkim A malicious client can trigger an OPENSSL_assert (i.e., an abort) in 48280304Sjkim servers that both support SSLv2 and enable export cipher suites by sending 49280304Sjkim a specially crafted SSLv2 CLIENT-MASTER-KEY message. 50280304Sjkim 51280304Sjkim This issue was discovered by Sean Burford (Google) and Emilia K�sper 52280304Sjkim (OpenSSL development team). 53280304Sjkim (CVE-2015-0293) 54280304Sjkim [Emilia K�sper] 55280304Sjkim 56280304Sjkim *) Use After Free following d2i_ECPrivatekey error fix 57280304Sjkim 58280304Sjkim A malformed EC private key file consumed via the d2i_ECPrivateKey function 59280304Sjkim could cause a use after free condition. This, in turn, could cause a double 60280304Sjkim free in several private key parsing functions (such as d2i_PrivateKey 61280304Sjkim or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption 62280304Sjkim for applications that receive EC private keys from untrusted 63280304Sjkim sources. This scenario is considered rare. 64280304Sjkim 65280304Sjkim This issue was discovered by the BoringSSL project and fixed in their 66280304Sjkim commit 517073cd4b. 67280304Sjkim (CVE-2015-0209) 68280304Sjkim [Matt Caswell] 69280304Sjkim 70280304Sjkim *) X509_to_X509_REQ NULL pointer deref fix 71280304Sjkim 72280304Sjkim The function X509_to_X509_REQ will crash with a NULL pointer dereference if 73280304Sjkim the certificate key is invalid. This function is rarely used in practice. 74280304Sjkim 75280304Sjkim This issue was discovered by Brian Carpenter. 76280304Sjkim (CVE-2015-0288) 77280304Sjkim [Stephen Henson] 78280304Sjkim 79280304Sjkim *) Removed the export ciphers from the DEFAULT ciphers 80280304Sjkim [Kurt Roeckx] 81280304Sjkim 82277597Sjkim Changes between 1.0.1k and 1.0.1l [15 Jan 2015] 83277597Sjkim 84277597Sjkim *) Build fixes for the Windows and OpenVMS platforms 85277597Sjkim [Matt Caswell and Richard Levitte] 86277597Sjkim 87276864Sjkim Changes between 1.0.1j and 1.0.1k [8 Jan 2015] 88276864Sjkim 89276864Sjkim *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS 90276864Sjkim message can cause a segmentation fault in OpenSSL due to a NULL pointer 91276864Sjkim dereference. This could lead to a Denial Of Service attack. Thanks to 92276864Sjkim Markus Stenberg of Cisco Systems, Inc. for reporting this issue. 93276864Sjkim (CVE-2014-3571) 94276864Sjkim [Steve Henson] 95276864Sjkim 96276864Sjkim *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the 97276864Sjkim dtls1_buffer_record function under certain conditions. In particular this 98276864Sjkim could occur if an attacker sent repeated DTLS records with the same 99276864Sjkim sequence number but for the next epoch. The memory leak could be exploited 100276864Sjkim by an attacker in a Denial of Service attack through memory exhaustion. 101276864Sjkim Thanks to Chris Mueller for reporting this issue. 102276864Sjkim (CVE-2015-0206) 103276864Sjkim [Matt Caswell] 104276864Sjkim 105276864Sjkim *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is 106276864Sjkim built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl 107276864Sjkim method would be set to NULL which could later result in a NULL pointer 108276864Sjkim dereference. Thanks to Frank Schmirler for reporting this issue. 109276864Sjkim (CVE-2014-3569) 110276864Sjkim [Kurt Roeckx] 111276864Sjkim 112276864Sjkim *) Abort handshake if server key exchange message is omitted for ephemeral 113276864Sjkim ECDH ciphersuites. 114276864Sjkim 115276864Sjkim Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for 116276864Sjkim reporting this issue. 117276864Sjkim (CVE-2014-3572) 118276864Sjkim [Steve Henson] 119276864Sjkim 120276864Sjkim *) Remove non-export ephemeral RSA code on client and server. This code 121276864Sjkim violated the TLS standard by allowing the use of temporary RSA keys in 122276864Sjkim non-export ciphersuites and could be used by a server to effectively 123276864Sjkim downgrade the RSA key length used to a value smaller than the server 124276864Sjkim certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at 125276864Sjkim INRIA or reporting this issue. 126276864Sjkim (CVE-2015-0204) 127276864Sjkim [Steve Henson] 128276864Sjkim 129276864Sjkim *) Fixed issue where DH client certificates are accepted without verification. 130276864Sjkim An OpenSSL server will accept a DH certificate for client authentication 131276864Sjkim without the certificate verify message. This effectively allows a client to 132276864Sjkim authenticate without the use of a private key. This only affects servers 133276864Sjkim which trust a client certificate authority which issues certificates 134276864Sjkim containing DH keys: these are extremely rare and hardly ever encountered. 135276864Sjkim Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting 136276864Sjkim this issue. 137276864Sjkim (CVE-2015-0205) 138276864Sjkim [Steve Henson] 139276864Sjkim 140276864Sjkim *) Ensure that the session ID context of an SSL is updated when its 141276864Sjkim SSL_CTX is updated via SSL_set_SSL_CTX. 142276864Sjkim 143276864Sjkim The session ID context is typically set from the parent SSL_CTX, 144276864Sjkim and can vary with the CTX. 145276864Sjkim [Adam Langley] 146276864Sjkim 147276864Sjkim *) Fix various certificate fingerprint issues. 148276864Sjkim 149276864Sjkim By using non-DER or invalid encodings outside the signed portion of a 150276864Sjkim certificate the fingerprint can be changed without breaking the signature. 151276864Sjkim Although no details of the signed portion of the certificate can be changed 152276864Sjkim this can cause problems with some applications: e.g. those using the 153276864Sjkim certificate fingerprint for blacklists. 154276864Sjkim 155276864Sjkim 1. Reject signatures with non zero unused bits. 156276864Sjkim 157276864Sjkim If the BIT STRING containing the signature has non zero unused bits reject 158276864Sjkim the signature. All current signature algorithms require zero unused bits. 159276864Sjkim 160276864Sjkim 2. Check certificate algorithm consistency. 161276864Sjkim 162276864Sjkim Check the AlgorithmIdentifier inside TBS matches the one in the 163276864Sjkim certificate signature. NB: this will result in signature failure 164276864Sjkim errors for some broken certificates. 165276864Sjkim 166276864Sjkim Thanks to Konrad Kraszewski from Google for reporting this issue. 167276864Sjkim 168276864Sjkim 3. Check DSA/ECDSA signatures use DER. 169276864Sjkim 170276864Sjkim Reencode DSA/ECDSA signatures and compare with the original received 171276864Sjkim signature. Return an error if there is a mismatch. 172276864Sjkim 173276864Sjkim This will reject various cases including garbage after signature 174276864Sjkim (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS 175276864Sjkim program for discovering this case) and use of BER or invalid ASN.1 INTEGERs 176276864Sjkim (negative or with leading zeroes). 177276864Sjkim 178276864Sjkim Further analysis was conducted and fixes were developed by Stephen Henson 179276864Sjkim of the OpenSSL core team. 180276864Sjkim 181276864Sjkim (CVE-2014-8275) 182276864Sjkim [Steve Henson] 183276864Sjkim 184276864Sjkim *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect 185276864Sjkim results on some platforms, including x86_64. This bug occurs at random 186276864Sjkim with a very low probability, and is not known to be exploitable in any 187276864Sjkim way, though its exact impact is difficult to determine. Thanks to Pieter 188276864Sjkim Wuille (Blockstream) who reported this issue and also suggested an initial 189276864Sjkim fix. Further analysis was conducted by the OpenSSL development team and 190276864Sjkim Adam Langley of Google. The final fix was developed by Andy Polyakov of 191276864Sjkim the OpenSSL core team. 192276864Sjkim (CVE-2014-3570) 193276864Sjkim [Andy Polyakov] 194276864Sjkim 195276864Sjkim *) Do not resume sessions on the server if the negotiated protocol 196276864Sjkim version does not match the session's version. Resuming with a different 197276864Sjkim version, while not strictly forbidden by the RFC, is of questionable 198276864Sjkim sanity and breaks all known clients. 199276864Sjkim [David Benjamin, Emilia K�sper] 200276864Sjkim 201276864Sjkim *) Tighten handling of the ChangeCipherSpec (CCS) message: reject 202276864Sjkim early CCS messages during renegotiation. (Note that because 203276864Sjkim renegotiation is encrypted, this early CCS was not exploitable.) 204276864Sjkim [Emilia K�sper] 205276864Sjkim 206276864Sjkim *) Tighten client-side session ticket handling during renegotiation: 207276864Sjkim ensure that the client only accepts a session ticket if the server sends 208276864Sjkim the extension anew in the ServerHello. Previously, a TLS client would 209276864Sjkim reuse the old extension state and thus accept a session ticket if one was 210276864Sjkim announced in the initial ServerHello. 211276864Sjkim 212276864Sjkim Similarly, ensure that the client requires a session ticket if one 213276864Sjkim was advertised in the ServerHello. Previously, a TLS client would 214276864Sjkim ignore a missing NewSessionTicket message. 215276864Sjkim [Emilia K�sper] 216276864Sjkim 217273149Sjkim Changes between 1.0.1i and 1.0.1j [15 Oct 2014] 218273149Sjkim 219273149Sjkim *) SRTP Memory Leak. 220273149Sjkim 221273149Sjkim A flaw in the DTLS SRTP extension parsing code allows an attacker, who 222273149Sjkim sends a carefully crafted handshake message, to cause OpenSSL to fail 223273149Sjkim to free up to 64k of memory causing a memory leak. This could be 224273149Sjkim exploited in a Denial Of Service attack. This issue affects OpenSSL 225273149Sjkim 1.0.1 server implementations for both SSL/TLS and DTLS regardless of 226273149Sjkim whether SRTP is used or configured. Implementations of OpenSSL that 227273149Sjkim have been compiled with OPENSSL_NO_SRTP defined are not affected. 228273149Sjkim 229273149Sjkim The fix was developed by the OpenSSL team. 230273149Sjkim (CVE-2014-3513) 231273149Sjkim [OpenSSL team] 232273149Sjkim 233273149Sjkim *) Session Ticket Memory Leak. 234273149Sjkim 235273149Sjkim When an OpenSSL SSL/TLS/DTLS server receives a session ticket the 236273149Sjkim integrity of that ticket is first verified. In the event of a session 237273149Sjkim ticket integrity check failing, OpenSSL will fail to free memory 238273149Sjkim causing a memory leak. By sending a large number of invalid session 239273149Sjkim tickets an attacker could exploit this issue in a Denial Of Service 240273149Sjkim attack. 241273149Sjkim (CVE-2014-3567) 242273149Sjkim [Steve Henson] 243273149Sjkim 244273149Sjkim *) Build option no-ssl3 is incomplete. 245273149Sjkim 246273149Sjkim When OpenSSL is configured with "no-ssl3" as a build option, servers 247273149Sjkim could accept and complete a SSL 3.0 handshake, and clients could be 248273149Sjkim configured to send them. 249273149Sjkim (CVE-2014-3568) 250273149Sjkim [Akamai and the OpenSSL team] 251273149Sjkim 252273149Sjkim *) Add support for TLS_FALLBACK_SCSV. 253273149Sjkim Client applications doing fallback retries should call 254273149Sjkim SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV). 255273149Sjkim (CVE-2014-3566) 256273149Sjkim [Adam Langley, Bodo Moeller] 257273149Sjkim 258273149Sjkim *) Add additional DigestInfo checks. 259273149Sjkim 260273149Sjkim Reencode DigestInto in DER and check against the original when 261273149Sjkim verifying RSA signature: this will reject any improperly encoded 262273149Sjkim DigestInfo structures. 263273149Sjkim 264273149Sjkim Note: this is a precautionary measure and no attacks are currently known. 265273149Sjkim 266273149Sjkim [Steve Henson] 267273149Sjkim 268269686Sjkim Changes between 1.0.1h and 1.0.1i [6 Aug 2014] 269269686Sjkim 270269686Sjkim *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the 271269686Sjkim SRP code can be overrun an internal buffer. Add sanity check that 272269686Sjkim g, A, B < N to SRP code. 273269686Sjkim 274269686Sjkim Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC 275269686Sjkim Group for discovering this issue. 276269686Sjkim (CVE-2014-3512) 277269686Sjkim [Steve Henson] 278269686Sjkim 279269686Sjkim *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate 280269686Sjkim TLS 1.0 instead of higher protocol versions when the ClientHello message 281269686Sjkim is badly fragmented. This allows a man-in-the-middle attacker to force a 282269686Sjkim downgrade to TLS 1.0 even if both the server and the client support a 283269686Sjkim higher protocol version, by modifying the client's TLS records. 284269686Sjkim 285269686Sjkim Thanks to David Benjamin and Adam Langley (Google) for discovering and 286269686Sjkim researching this issue. 287269686Sjkim (CVE-2014-3511) 288269686Sjkim [David Benjamin] 289269686Sjkim 290269686Sjkim *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject 291269686Sjkim to a denial of service attack. A malicious server can crash the client 292269686Sjkim with a null pointer dereference (read) by specifying an anonymous (EC)DH 293269686Sjkim ciphersuite and sending carefully crafted handshake messages. 294269686Sjkim 295269686Sjkim Thanks to Felix Gr�bert (Google) for discovering and researching this 296269686Sjkim issue. 297269686Sjkim (CVE-2014-3510) 298269686Sjkim [Emilia K�sper] 299269686Sjkim 300269686Sjkim *) By sending carefully crafted DTLS packets an attacker could cause openssl 301269686Sjkim to leak memory. This can be exploited through a Denial of Service attack. 302269686Sjkim Thanks to Adam Langley for discovering and researching this issue. 303269686Sjkim (CVE-2014-3507) 304269686Sjkim [Adam Langley] 305269686Sjkim 306269686Sjkim *) An attacker can force openssl to consume large amounts of memory whilst 307269686Sjkim processing DTLS handshake messages. This can be exploited through a 308269686Sjkim Denial of Service attack. 309269686Sjkim Thanks to Adam Langley for discovering and researching this issue. 310269686Sjkim (CVE-2014-3506) 311269686Sjkim [Adam Langley] 312269686Sjkim 313269686Sjkim *) An attacker can force an error condition which causes openssl to crash 314269686Sjkim whilst processing DTLS packets due to memory being freed twice. This 315269686Sjkim can be exploited through a Denial of Service attack. 316269686Sjkim Thanks to Adam Langley and Wan-Teh Chang for discovering and researching 317269686Sjkim this issue. 318269686Sjkim (CVE-2014-3505) 319269686Sjkim [Adam Langley] 320269686Sjkim 321269686Sjkim *) If a multithreaded client connects to a malicious server using a resumed 322269686Sjkim session and the server sends an ec point format extension it could write 323269686Sjkim up to 255 bytes to freed memory. 324269686Sjkim 325269686Sjkim Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this 326269686Sjkim issue. 327269686Sjkim (CVE-2014-3509) 328269686Sjkim [Gabor Tyukasz] 329269686Sjkim 330269686Sjkim *) A malicious server can crash an OpenSSL client with a null pointer 331269686Sjkim dereference (read) by specifying an SRP ciphersuite even though it was not 332269686Sjkim properly negotiated with the client. This can be exploited through a 333269686Sjkim Denial of Service attack. 334269686Sjkim 335269686Sjkim Thanks to Joonas Kuorilehto and Riku Hietam�ki (Codenomicon) for 336269686Sjkim discovering and researching this issue. 337269686Sjkim (CVE-2014-5139) 338269686Sjkim [Steve Henson] 339269686Sjkim 340269686Sjkim *) A flaw in OBJ_obj2txt may cause pretty printing functions such as 341269686Sjkim X509_name_oneline, X509_name_print_ex et al. to leak some information 342269686Sjkim from the stack. Applications may be affected if they echo pretty printing 343269686Sjkim output to the attacker. 344269686Sjkim 345269686Sjkim Thanks to Ivan Fratric (Google) for discovering this issue. 346269686Sjkim (CVE-2014-3508) 347269686Sjkim [Emilia K�sper, and Steve Henson] 348269686Sjkim 349269686Sjkim *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) 350269686Sjkim for corner cases. (Certain input points at infinity could lead to 351269686Sjkim bogus results, with non-infinity inputs mapped to infinity too.) 352269686Sjkim [Bodo Moeller] 353269686Sjkim 354267258Sjkim Changes between 1.0.1g and 1.0.1h [5 Jun 2014] 355267258Sjkim 356267258Sjkim *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted 357267258Sjkim handshake can force the use of weak keying material in OpenSSL 358267258Sjkim SSL/TLS clients and servers. 359267258Sjkim 360267258Sjkim Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and 361267258Sjkim researching this issue. (CVE-2014-0224) 362267258Sjkim [KIKUCHI Masashi, Steve Henson] 363267258Sjkim 364267258Sjkim *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an 365267258Sjkim OpenSSL DTLS client the code can be made to recurse eventually crashing 366267258Sjkim in a DoS attack. 367267258Sjkim 368267258Sjkim Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue. 369267258Sjkim (CVE-2014-0221) 370267258Sjkim [Imre Rad, Steve Henson] 371267258Sjkim 372267258Sjkim *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can 373267258Sjkim be triggered by sending invalid DTLS fragments to an OpenSSL DTLS 374267258Sjkim client or server. This is potentially exploitable to run arbitrary 375267258Sjkim code on a vulnerable client or server. 376267258Sjkim 377267258Sjkim Thanks to J�ri Aedla for reporting this issue. (CVE-2014-0195) 378267258Sjkim [J�ri Aedla, Steve Henson] 379267258Sjkim 380267258Sjkim *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites 381267258Sjkim are subject to a denial of service attack. 382267258Sjkim 383267258Sjkim Thanks to Felix Gr�bert and Ivan Fratric at Google for discovering 384267258Sjkim this issue. (CVE-2014-3470) 385267258Sjkim [Felix Gr�bert, Ivan Fratric, Steve Henson] 386267258Sjkim 387267258Sjkim *) Harmonize version and its documentation. -f flag is used to display 388267258Sjkim compilation flags. 389267258Sjkim [mancha <mancha1@zoho.com>] 390267258Sjkim 391267258Sjkim *) Fix eckey_priv_encode so it immediately returns an error upon a failure 392267258Sjkim in i2d_ECPrivateKey. 393267258Sjkim [mancha <mancha1@zoho.com>] 394267258Sjkim 395267258Sjkim *) Fix some double frees. These are not thought to be exploitable. 396267258Sjkim [mancha <mancha1@zoho.com>] 397267258Sjkim 398264331Sjkim Changes between 1.0.1f and 1.0.1g [7 Apr 2014] 399264331Sjkim 400264331Sjkim *) A missing bounds check in the handling of the TLS heartbeat extension 401264331Sjkim can be used to reveal up to 64k of memory to a connected client or 402264331Sjkim server. 403264331Sjkim 404264331Sjkim Thanks for Neel Mehta of Google Security for discovering this bug and to 405264331Sjkim Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for 406264331Sjkim preparing the fix (CVE-2014-0160) 407264331Sjkim [Adam Langley, Bodo Moeller] 408264331Sjkim 409264331Sjkim *) Fix for the attack described in the paper "Recovering OpenSSL 410264331Sjkim ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack" 411264331Sjkim by Yuval Yarom and Naomi Benger. Details can be obtained from: 412264331Sjkim http://eprint.iacr.org/2014/140 413264331Sjkim 414264331Sjkim Thanks to Yuval Yarom and Naomi Benger for discovering this 415264331Sjkim flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076) 416264331Sjkim [Yuval Yarom and Naomi Benger] 417264331Sjkim 418264331Sjkim *) TLS pad extension: draft-agl-tls-padding-03 419264331Sjkim 420264331Sjkim Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the 421264331Sjkim TLS client Hello record length value would otherwise be > 255 and 422264331Sjkim less that 512 pad with a dummy extension containing zeroes so it 423264331Sjkim is at least 512 bytes long. 424264331Sjkim 425264331Sjkim [Adam Langley, Steve Henson] 426264331Sjkim 427264331Sjkim Changes between 1.0.1e and 1.0.1f [6 Jan 2014] 428264331Sjkim 429264331Sjkim *) Fix for TLS record tampering bug. A carefully crafted invalid 430264331Sjkim handshake could crash OpenSSL with a NULL pointer exception. 431264331Sjkim Thanks to Anton Johansson for reporting this issues. 432264331Sjkim (CVE-2013-4353) 433264331Sjkim 434264331Sjkim *) Keep original DTLS digest and encryption contexts in retransmission 435264331Sjkim structures so we can use the previous session parameters if they need 436264331Sjkim to be resent. (CVE-2013-6450) 437264331Sjkim [Steve Henson] 438264331Sjkim 439264331Sjkim *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which 440264331Sjkim avoids preferring ECDHE-ECDSA ciphers when the client appears to be 441264331Sjkim Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for 442264331Sjkim several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug 443264331Sjkim is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing 444264331Sjkim 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer. 445264331Sjkim [Rob Stradling, Adam Langley] 446264331Sjkim 447246772Sjkim Changes between 1.0.1d and 1.0.1e [11 Feb 2013] 448246772Sjkim 449264331Sjkim *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI 450264331Sjkim supporting platforms or when small records were transferred. 451264331Sjkim [Andy Polyakov, Steve Henson] 452246772Sjkim 453246772Sjkim Changes between 1.0.1c and 1.0.1d [5 Feb 2013] 454246772Sjkim 455246772Sjkim *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 456246772Sjkim 457246772Sjkim This addresses the flaw in CBC record processing discovered by 458246772Sjkim Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 459246772Sjkim at: http://www.isg.rhul.ac.uk/tls/ 460246772Sjkim 461246772Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 462246772Sjkim Security Group at Royal Holloway, University of London 463246772Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 464246772Sjkim Emilia K�sper for the initial patch. 465246772Sjkim (CVE-2013-0169) 466246772Sjkim [Emilia K�sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 467246772Sjkim 468246772Sjkim *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode 469246772Sjkim ciphersuites which can be exploited in a denial of service attack. 470246772Sjkim Thanks go to and to Adam Langley <agl@chromium.org> for discovering 471246772Sjkim and detecting this bug and to Wolfgang Ettlinger 472246772Sjkim <wolfgang.ettlinger@gmail.com> for independently discovering this issue. 473246772Sjkim (CVE-2012-2686) 474246772Sjkim [Adam Langley] 475246772Sjkim 476246772Sjkim *) Return an error when checking OCSP signatures when key is NULL. 477246772Sjkim This fixes a DoS attack. (CVE-2013-0166) 478246772Sjkim [Steve Henson] 479246772Sjkim 480246772Sjkim *) Make openssl verify return errors. 481246772Sjkim [Chris Palmer <palmer@google.com> and Ben Laurie] 482246772Sjkim 483246772Sjkim *) Call OCSP Stapling callback after ciphersuite has been chosen, so 484246772Sjkim the right response is stapled. Also change SSL_get_certificate() 485246772Sjkim so it returns the certificate actually sent. 486246772Sjkim See http://rt.openssl.org/Ticket/Display.html?id=2836. 487246772Sjkim [Rob Stradling <rob.stradling@comodo.com>] 488246772Sjkim 489246772Sjkim *) Fix possible deadlock when decoding public keys. 490246772Sjkim [Steve Henson] 491246772Sjkim 492246772Sjkim *) Don't use TLS 1.0 record version number in initial client hello 493246772Sjkim if renegotiating. 494246772Sjkim [Steve Henson] 495246772Sjkim 496238405Sjkim Changes between 1.0.1b and 1.0.1c [10 May 2012] 497237657Sjkim 498238405Sjkim *) Sanity check record length before skipping explicit IV in TLS 499238405Sjkim 1.2, 1.1 and DTLS to fix DoS attack. 500237657Sjkim 501237657Sjkim Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 502237657Sjkim fuzzing as a service testing platform. 503237657Sjkim (CVE-2012-2333) 504237657Sjkim [Steve Henson] 505237657Sjkim 506237657Sjkim *) Initialise tkeylen properly when encrypting CMS messages. 507237657Sjkim Thanks to Solar Designer of Openwall for reporting this issue. 508237657Sjkim [Steve Henson] 509237657Sjkim 510238405Sjkim *) In FIPS mode don't try to use composite ciphers as they are not 511238405Sjkim approved. 512238405Sjkim [Steve Henson] 513237657Sjkim 514238405Sjkim Changes between 1.0.1a and 1.0.1b [26 Apr 2012] 515237657Sjkim 516238405Sjkim *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and 517238405Sjkim 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately 518238405Sjkim mean any application compiled against OpenSSL 1.0.0 headers setting 519238405Sjkim SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng 520238405Sjkim TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to 521238405Sjkim 0x10000000L Any application which was previously compiled against 522238405Sjkim OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1 523238405Sjkim will need to be recompiled as a result. Letting be results in 524238405Sjkim inability to disable specifically TLS 1.1 and in client context, 525238405Sjkim in unlike event, limit maximum offered version to TLS 1.0 [see below]. 526238405Sjkim [Steve Henson] 527237657Sjkim 528238405Sjkim *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not 529238405Sjkim disable just protocol X, but all protocols above X *if* there are 530238405Sjkim protocols *below* X still enabled. In more practical terms it means 531238405Sjkim that if application wants to disable TLS1.0 in favor of TLS1.1 and 532238405Sjkim above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass 533238405Sjkim SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to 534238405Sjkim client side. 535238405Sjkim [Andy Polyakov] 536238405Sjkim 537238405Sjkim Changes between 1.0.1 and 1.0.1a [19 Apr 2012] 538238405Sjkim 539237657Sjkim *) Check for potentially exploitable overflows in asn1_d2i_read_bio 540237657Sjkim BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 541237657Sjkim in CRYPTO_realloc_clean. 542237657Sjkim 543237657Sjkim Thanks to Tavis Ormandy, Google Security Team, for discovering this 544237657Sjkim issue and to Adam Langley <agl@chromium.org> for fixing it. 545237657Sjkim (CVE-2012-2110) 546237657Sjkim [Adam Langley (Google), Tavis Ormandy, Google Security Team] 547237657Sjkim 548238405Sjkim *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections. 549238405Sjkim [Adam Langley] 550237657Sjkim 551238405Sjkim *) Workarounds for some broken servers that "hang" if a client hello 552238405Sjkim record length exceeds 255 bytes. 553238405Sjkim 554238405Sjkim 1. Do not use record version number > TLS 1.0 in initial client 555238405Sjkim hello: some (but not all) hanging servers will now work. 556238405Sjkim 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate 557238405Sjkim the number of ciphers sent in the client hello. This should be 558238405Sjkim set to an even number, such as 50, for example by passing: 559238405Sjkim -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure. 560238405Sjkim Most broken servers should now work. 561238405Sjkim 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable 562238405Sjkim TLS 1.2 client support entirely. 563238405Sjkim [Steve Henson] 564238405Sjkim 565238405Sjkim *) Fix SEGV in Vector Permutation AES module observed in OpenSSH. 566238405Sjkim [Andy Polyakov] 567238405Sjkim 568238405Sjkim Changes between 1.0.0h and 1.0.1 [14 Mar 2012] 569238405Sjkim 570238405Sjkim *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET 571238405Sjkim STRING form instead of a DigestInfo. 572238405Sjkim [Steve Henson] 573238405Sjkim 574238405Sjkim *) The format used for MDC2 RSA signatures is inconsistent between EVP 575238405Sjkim and the RSA_sign/RSA_verify functions. This was made more apparent when 576238405Sjkim OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular 577238405Sjkim those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect 578238405Sjkim the correct format in RSA_verify so both forms transparently work. 579238405Sjkim [Steve Henson] 580238405Sjkim 581238405Sjkim *) Some servers which support TLS 1.0 can choke if we initially indicate 582238405Sjkim support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA 583238405Sjkim encrypted premaster secret. As a workaround use the maximum pemitted 584238405Sjkim client version in client hello, this should keep such servers happy 585238405Sjkim and still work with previous versions of OpenSSL. 586238405Sjkim [Steve Henson] 587238405Sjkim 588238405Sjkim *) Add support for TLS/DTLS heartbeats. 589238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 590238405Sjkim 591238405Sjkim *) Add support for SCTP. 592238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 593238405Sjkim 594238405Sjkim *) Improved PRNG seeding for VOS. 595238405Sjkim [Paul Green <Paul.Green@stratus.com>] 596238405Sjkim 597238405Sjkim *) Extensive assembler packs updates, most notably: 598238405Sjkim 599238405Sjkim - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support; 600238405Sjkim - x86[_64]: SSSE3 support (SHA1, vector-permutation AES); 601238405Sjkim - x86_64: bit-sliced AES implementation; 602238405Sjkim - ARM: NEON support, contemporary platforms optimizations; 603238405Sjkim - s390x: z196 support; 604238405Sjkim - *: GHASH and GF(2^m) multiplication implementations; 605238405Sjkim 606238405Sjkim [Andy Polyakov] 607238405Sjkim 608238405Sjkim *) Make TLS-SRP code conformant with RFC 5054 API cleanup 609238405Sjkim (removal of unnecessary code) 610238405Sjkim [Peter Sylvester <peter.sylvester@edelweb.fr>] 611238405Sjkim 612238405Sjkim *) Add TLS key material exporter from RFC 5705. 613238405Sjkim [Eric Rescorla] 614238405Sjkim 615238405Sjkim *) Add DTLS-SRTP negotiation from RFC 5764. 616238405Sjkim [Eric Rescorla] 617238405Sjkim 618238405Sjkim *) Add Next Protocol Negotiation, 619238405Sjkim http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be 620238405Sjkim disabled with a no-npn flag to config or Configure. Code donated 621238405Sjkim by Google. 622238405Sjkim [Adam Langley <agl@google.com> and Ben Laurie] 623238405Sjkim 624238405Sjkim *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224, 625238405Sjkim NIST-P256, NIST-P521, with constant-time single point multiplication on 626238405Sjkim typical inputs. Compiler support for the nonstandard type __uint128_t is 627238405Sjkim required to use this (present in gcc 4.4 and later, for 64-bit builds). 628238405Sjkim Code made available under Apache License version 2.0. 629238405Sjkim 630238405Sjkim Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command 631238405Sjkim line to include this in your build of OpenSSL, and run "make depend" (or 632238405Sjkim "make update"). This enables the following EC_METHODs: 633238405Sjkim 634238405Sjkim EC_GFp_nistp224_method() 635238405Sjkim EC_GFp_nistp256_method() 636238405Sjkim EC_GFp_nistp521_method() 637238405Sjkim 638238405Sjkim EC_GROUP_new_by_curve_name() will automatically use these (while 639238405Sjkim EC_GROUP_new_curve_GFp() currently prefers the more flexible 640238405Sjkim implementations). 641238405Sjkim [Emilia K�sper, Adam Langley, Bodo Moeller (Google)] 642238405Sjkim 643238405Sjkim *) Use type ossl_ssize_t instad of ssize_t which isn't available on 644238405Sjkim all platforms. Move ssize_t definition from e_os.h to the public 645238405Sjkim header file e_os2.h as it now appears in public header file cms.h 646238405Sjkim [Steve Henson] 647238405Sjkim 648238405Sjkim *) New -sigopt option to the ca, req and x509 utilities. Additional 649238405Sjkim signature parameters can be passed using this option and in 650238405Sjkim particular PSS. 651238405Sjkim [Steve Henson] 652238405Sjkim 653238405Sjkim *) Add RSA PSS signing function. This will generate and set the 654238405Sjkim appropriate AlgorithmIdentifiers for PSS based on those in the 655238405Sjkim corresponding EVP_MD_CTX structure. No application support yet. 656238405Sjkim [Steve Henson] 657238405Sjkim 658238405Sjkim *) Support for companion algorithm specific ASN1 signing routines. 659238405Sjkim New function ASN1_item_sign_ctx() signs a pre-initialised 660238405Sjkim EVP_MD_CTX structure and sets AlgorithmIdentifiers based on 661238405Sjkim the appropriate parameters. 662238405Sjkim [Steve Henson] 663238405Sjkim 664238405Sjkim *) Add new algorithm specific ASN1 verification initialisation function 665238405Sjkim to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1 666238405Sjkim handling will be the same no matter what EVP_PKEY_METHOD is used. 667238405Sjkim Add a PSS handler to support verification of PSS signatures: checked 668238405Sjkim against a number of sample certificates. 669238405Sjkim [Steve Henson] 670238405Sjkim 671238405Sjkim *) Add signature printing for PSS. Add PSS OIDs. 672238405Sjkim [Steve Henson, Martin Kaiser <lists@kaiser.cx>] 673238405Sjkim 674238405Sjkim *) Add algorithm specific signature printing. An individual ASN1 method 675238405Sjkim can now print out signatures instead of the standard hex dump. 676238405Sjkim 677238405Sjkim More complex signatures (e.g. PSS) can print out more meaningful 678238405Sjkim information. Include DSA version that prints out the signature 679238405Sjkim parameters r, s. 680238405Sjkim [Steve Henson] 681238405Sjkim 682238405Sjkim *) Password based recipient info support for CMS library: implementing 683238405Sjkim RFC3211. 684238405Sjkim [Steve Henson] 685238405Sjkim 686238405Sjkim *) Split password based encryption into PBES2 and PBKDF2 functions. This 687238405Sjkim neatly separates the code into cipher and PBE sections and is required 688238405Sjkim for some algorithms that split PBES2 into separate pieces (such as 689238405Sjkim password based CMS). 690238405Sjkim [Steve Henson] 691238405Sjkim 692238405Sjkim *) Session-handling fixes: 693238405Sjkim - Fix handling of connections that are resuming with a session ID, 694238405Sjkim but also support Session Tickets. 695238405Sjkim - Fix a bug that suppressed issuing of a new ticket if the client 696238405Sjkim presented a ticket with an expired session. 697238405Sjkim - Try to set the ticket lifetime hint to something reasonable. 698238405Sjkim - Make tickets shorter by excluding irrelevant information. 699238405Sjkim - On the client side, don't ignore renewed tickets. 700238405Sjkim [Adam Langley, Bodo Moeller (Google)] 701238405Sjkim 702238405Sjkim *) Fix PSK session representation. 703238405Sjkim [Bodo Moeller] 704238405Sjkim 705238405Sjkim *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations. 706238405Sjkim 707238405Sjkim This work was sponsored by Intel. 708238405Sjkim [Andy Polyakov] 709238405Sjkim 710238405Sjkim *) Add GCM support to TLS library. Some custom code is needed to split 711238405Sjkim the IV between the fixed (from PRF) and explicit (from TLS record) 712238405Sjkim portions. This adds all GCM ciphersuites supported by RFC5288 and 713238405Sjkim RFC5289. Generalise some AES* cipherstrings to inlclude GCM and 714238405Sjkim add a special AESGCM string for GCM only. 715238405Sjkim [Steve Henson] 716238405Sjkim 717238405Sjkim *) Expand range of ctrls for AES GCM. Permit setting invocation 718238405Sjkim field on decrypt and retrieval of invocation field only on encrypt. 719238405Sjkim [Steve Henson] 720238405Sjkim 721238405Sjkim *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support. 722238405Sjkim As required by RFC5289 these ciphersuites cannot be used if for 723238405Sjkim versions of TLS earlier than 1.2. 724238405Sjkim [Steve Henson] 725238405Sjkim 726238405Sjkim *) For FIPS capable OpenSSL interpret a NULL default public key method 727238405Sjkim as unset and return the appopriate default but do *not* set the default. 728238405Sjkim This means we can return the appopriate method in applications that 729238405Sjkim swicth between FIPS and non-FIPS modes. 730238405Sjkim [Steve Henson] 731238405Sjkim 732238405Sjkim *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an 733238405Sjkim ENGINE is used then we cannot handle that in the FIPS module so we 734238405Sjkim keep original code iff non-FIPS operations are allowed. 735238405Sjkim [Steve Henson] 736238405Sjkim 737238405Sjkim *) Add -attime option to openssl utilities. 738238405Sjkim [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson] 739238405Sjkim 740238405Sjkim *) Redirect DSA and DH operations to FIPS module in FIPS mode. 741238405Sjkim [Steve Henson] 742238405Sjkim 743238405Sjkim *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use 744238405Sjkim FIPS EC methods unconditionally for now. 745238405Sjkim [Steve Henson] 746238405Sjkim 747238405Sjkim *) New build option no-ec2m to disable characteristic 2 code. 748238405Sjkim [Steve Henson] 749238405Sjkim 750238405Sjkim *) Backport libcrypto audit of return value checking from 1.1.0-dev; not 751238405Sjkim all cases can be covered as some introduce binary incompatibilities. 752238405Sjkim [Steve Henson] 753238405Sjkim 754238405Sjkim *) Redirect RSA operations to FIPS module including keygen, 755238405Sjkim encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods. 756238405Sjkim [Steve Henson] 757238405Sjkim 758238405Sjkim *) Add similar low level API blocking to ciphers. 759238405Sjkim [Steve Henson] 760238405Sjkim 761238405Sjkim *) Low level digest APIs are not approved in FIPS mode: any attempt 762238405Sjkim to use these will cause a fatal error. Applications that *really* want 763238405Sjkim to use them can use the private_* version instead. 764238405Sjkim [Steve Henson] 765238405Sjkim 766238405Sjkim *) Redirect cipher operations to FIPS module for FIPS builds. 767238405Sjkim [Steve Henson] 768238405Sjkim 769238405Sjkim *) Redirect digest operations to FIPS module for FIPS builds. 770238405Sjkim [Steve Henson] 771238405Sjkim 772238405Sjkim *) Update build system to add "fips" flag which will link in fipscanister.o 773238405Sjkim for static and shared library builds embedding a signature if needed. 774238405Sjkim [Steve Henson] 775238405Sjkim 776238405Sjkim *) Output TLS supported curves in preference order instead of numerical 777238405Sjkim order. This is currently hardcoded for the highest order curves first. 778238405Sjkim This should be configurable so applications can judge speed vs strength. 779238405Sjkim [Steve Henson] 780238405Sjkim 781238405Sjkim *) Add TLS v1.2 server support for client authentication. 782238405Sjkim [Steve Henson] 783238405Sjkim 784238405Sjkim *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers 785238405Sjkim and enable MD5. 786238405Sjkim [Steve Henson] 787238405Sjkim 788238405Sjkim *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying 789238405Sjkim FIPS modules versions. 790238405Sjkim [Steve Henson] 791238405Sjkim 792238405Sjkim *) Add TLS v1.2 client side support for client authentication. Keep cache 793238405Sjkim of handshake records longer as we don't know the hash algorithm to use 794238405Sjkim until after the certificate request message is received. 795238405Sjkim [Steve Henson] 796238405Sjkim 797238405Sjkim *) Initial TLS v1.2 client support. Add a default signature algorithms 798238405Sjkim extension including all the algorithms we support. Parse new signature 799238405Sjkim format in client key exchange. Relax some ECC signing restrictions for 800238405Sjkim TLS v1.2 as indicated in RFC5246. 801238405Sjkim [Steve Henson] 802238405Sjkim 803238405Sjkim *) Add server support for TLS v1.2 signature algorithms extension. Switch 804238405Sjkim to new signature format when needed using client digest preference. 805238405Sjkim All server ciphersuites should now work correctly in TLS v1.2. No client 806238405Sjkim support yet and no support for client certificates. 807238405Sjkim [Steve Henson] 808238405Sjkim 809238405Sjkim *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch 810238405Sjkim to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based 811238405Sjkim ciphersuites. At present only RSA key exchange ciphersuites work with 812238405Sjkim TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete 813238405Sjkim SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods 814238405Sjkim and version checking. 815238405Sjkim [Steve Henson] 816238405Sjkim 817238405Sjkim *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled 818238405Sjkim with this defined it will not be affected by any changes to ssl internal 819238405Sjkim structures. Add several utility functions to allow openssl application 820238405Sjkim to work with OPENSSL_NO_SSL_INTERN defined. 821238405Sjkim [Steve Henson] 822238405Sjkim 823238405Sjkim *) Add SRP support. 824238405Sjkim [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie] 825238405Sjkim 826238405Sjkim *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id. 827238405Sjkim [Steve Henson] 828238405Sjkim 829238405Sjkim *) Permit abbreviated handshakes when renegotiating using the function 830238405Sjkim SSL_renegotiate_abbreviated(). 831238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 832238405Sjkim 833238405Sjkim *) Add call to ENGINE_register_all_complete() to 834238405Sjkim ENGINE_load_builtin_engines(), so some implementations get used 835238405Sjkim automatically instead of needing explicit application support. 836238405Sjkim [Steve Henson] 837238405Sjkim 838238405Sjkim *) Add support for TLS key exporter as described in RFC5705. 839238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson] 840238405Sjkim 841238405Sjkim *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only 842238405Sjkim a few changes are required: 843238405Sjkim 844238405Sjkim Add SSL_OP_NO_TLSv1_1 flag. 845238405Sjkim Add TLSv1_1 methods. 846238405Sjkim Update version checking logic to handle version 1.1. 847238405Sjkim Add explicit IV handling (ported from DTLS code). 848238405Sjkim Add command line options to s_client/s_server. 849238405Sjkim [Steve Henson] 850238405Sjkim 851238405Sjkim Changes between 1.0.0g and 1.0.0h [12 Mar 2012] 852238405Sjkim 853237657Sjkim *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 854237657Sjkim in CMS and PKCS7 code. When RSA decryption fails use a random key for 855237657Sjkim content decryption and always return the same error. Note: this attack 856237657Sjkim needs on average 2^20 messages so it only affects automated senders. The 857237657Sjkim old behaviour can be reenabled in the CMS code by setting the 858237657Sjkim CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 859237657Sjkim an MMA defence is not necessary. 860237657Sjkim Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 861237657Sjkim this issue. (CVE-2012-0884) 862237657Sjkim [Steve Henson] 863237657Sjkim 864237657Sjkim *) Fix CVE-2011-4619: make sure we really are receiving a 865237657Sjkim client hello before rejecting multiple SGC restarts. Thanks to 866237657Sjkim Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 867237657Sjkim [Steve Henson] 868237657Sjkim 869238405Sjkim Changes between 1.0.0f and 1.0.0g [18 Jan 2012] 870238405Sjkim 871238405Sjkim *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 872238405Sjkim Thanks to Antonio Martin, Enterprise Secure Access Research and 873238405Sjkim Development, Cisco Systems, Inc. for discovering this bug and 874238405Sjkim preparing a fix. (CVE-2012-0050) 875238405Sjkim [Antonio Martin] 876238405Sjkim 877238405Sjkim Changes between 1.0.0e and 1.0.0f [4 Jan 2012] 878238405Sjkim 879238405Sjkim *) Nadhem Alfardan and Kenny Paterson have discovered an extension 880238405Sjkim of the Vaudenay padding oracle attack on CBC mode encryption 881238405Sjkim which enables an efficient plaintext recovery attack against 882238405Sjkim the OpenSSL implementation of DTLS. Their attack exploits timing 883238405Sjkim differences arising during decryption processing. A research 884238405Sjkim paper describing this attack can be found at: 885238405Sjkim http://www.isg.rhul.ac.uk/~kp/dtls.pdf 886238405Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 887238405Sjkim Security Group at Royal Holloway, University of London 888238405Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 889238405Sjkim <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 890238405Sjkim for preparing the fix. (CVE-2011-4108) 891238405Sjkim [Robin Seggelmann, Michael Tuexen] 892238405Sjkim 893238405Sjkim *) Clear bytes used for block padding of SSL 3.0 records. 894238405Sjkim (CVE-2011-4576) 895238405Sjkim [Adam Langley (Google)] 896238405Sjkim 897238405Sjkim *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 898238405Sjkim Kadianakis <desnacked@gmail.com> for discovering this issue and 899238405Sjkim Adam Langley for preparing the fix. (CVE-2011-4619) 900238405Sjkim [Adam Langley (Google)] 901238405Sjkim 902238405Sjkim *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027) 903238405Sjkim [Andrey Kulikov <amdeich@gmail.com>] 904238405Sjkim 905238405Sjkim *) Prevent malformed RFC3779 data triggering an assertion failure. 906238405Sjkim Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 907238405Sjkim and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 908238405Sjkim [Rob Austein <sra@hactrn.net>] 909238405Sjkim 910238405Sjkim *) Improved PRNG seeding for VOS. 911238405Sjkim [Paul Green <Paul.Green@stratus.com>] 912238405Sjkim 913238405Sjkim *) Fix ssl_ciph.c set-up race. 914238405Sjkim [Adam Langley (Google)] 915238405Sjkim 916238405Sjkim *) Fix spurious failures in ecdsatest.c. 917238405Sjkim [Emilia K�sper (Google)] 918238405Sjkim 919238405Sjkim *) Fix the BIO_f_buffer() implementation (which was mixing different 920238405Sjkim interpretations of the '..._len' fields). 921238405Sjkim [Adam Langley (Google)] 922238405Sjkim 923238405Sjkim *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 924238405Sjkim BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 925238405Sjkim threads won't reuse the same blinding coefficients. 926238405Sjkim 927238405Sjkim This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 928238405Sjkim lock to call BN_BLINDING_invert_ex, and avoids one use of 929238405Sjkim BN_BLINDING_update for each BN_BLINDING structure (previously, 930238405Sjkim the last update always remained unused). 931238405Sjkim [Emilia K�sper (Google)] 932238405Sjkim 933238405Sjkim *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf. 934238405Sjkim [Bob Buckholz (Google)] 935238405Sjkim 936238405Sjkim Changes between 1.0.0d and 1.0.0e [6 Sep 2011] 937238405Sjkim 938238405Sjkim *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted 939238405Sjkim by initialising X509_STORE_CTX properly. (CVE-2011-3207) 940238405Sjkim [Kaspar Brand <ossl@velox.ch>] 941238405Sjkim 942238405Sjkim *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 943238405Sjkim for multi-threaded use of ECDH. (CVE-2011-3210) 944238405Sjkim [Adam Langley (Google)] 945238405Sjkim 946238405Sjkim *) Fix x509_name_ex_d2i memory leak on bad inputs. 947238405Sjkim [Bodo Moeller] 948238405Sjkim 949238405Sjkim *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check 950238405Sjkim signature public key algorithm by using OID xref utilities instead. 951238405Sjkim Before this you could only use some ECC ciphersuites with SHA1 only. 952238405Sjkim [Steve Henson] 953238405Sjkim 954238405Sjkim *) Add protection against ECDSA timing attacks as mentioned in the paper 955238405Sjkim by Billy Bob Brumley and Nicola Tuveri, see: 956238405Sjkim 957238405Sjkim http://eprint.iacr.org/2011/232.pdf 958238405Sjkim 959238405Sjkim [Billy Bob Brumley and Nicola Tuveri] 960238405Sjkim 961238405Sjkim Changes between 1.0.0c and 1.0.0d [8 Feb 2011] 962238405Sjkim 963238405Sjkim *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 964238405Sjkim [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 965238405Sjkim 966238405Sjkim *) Fix bug in string printing code: if *any* escaping is enabled we must 967238405Sjkim escape the escape character (backslash) or the resulting string is 968238405Sjkim ambiguous. 969238405Sjkim [Steve Henson] 970238405Sjkim 971238405Sjkim Changes between 1.0.0b and 1.0.0c [2 Dec 2010] 972238405Sjkim 973238405Sjkim *) Disable code workaround for ancient and obsolete Netscape browsers 974238405Sjkim and servers: an attacker can use it in a ciphersuite downgrade attack. 975238405Sjkim Thanks to Martin Rex for discovering this bug. CVE-2010-4180 976238405Sjkim [Steve Henson] 977238405Sjkim 978238405Sjkim *) Fixed J-PAKE implementation error, originally discovered by 979238405Sjkim Sebastien Martini, further info and confirmation from Stefan 980238405Sjkim Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 981238405Sjkim [Ben Laurie] 982238405Sjkim 983238405Sjkim Changes between 1.0.0a and 1.0.0b [16 Nov 2010] 984238405Sjkim 985238405Sjkim *) Fix extension code to avoid race conditions which can result in a buffer 986238405Sjkim overrun vulnerability: resumed sessions must not be modified as they can 987238405Sjkim be shared by multiple threads. CVE-2010-3864 988238405Sjkim [Steve Henson] 989238405Sjkim 990238405Sjkim *) Fix WIN32 build system to correctly link an ENGINE directory into 991238405Sjkim a DLL. 992238405Sjkim [Steve Henson] 993238405Sjkim 994238405Sjkim Changes between 1.0.0 and 1.0.0a [01 Jun 2010] 995238405Sjkim 996238405Sjkim *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover 997238405Sjkim (CVE-2010-1633) 998238405Sjkim [Steve Henson, Peter-Michael Hager <hager@dortmund.net>] 999238405Sjkim 1000238405Sjkim Changes between 0.9.8n and 1.0.0 [29 Mar 2010] 1001238405Sjkim 1002238405Sjkim *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher 1003238405Sjkim context. The operation can be customised via the ctrl mechanism in 1004238405Sjkim case ENGINEs want to include additional functionality. 1005238405Sjkim [Steve Henson] 1006238405Sjkim 1007238405Sjkim *) Tolerate yet another broken PKCS#8 key format: private key value negative. 1008238405Sjkim [Steve Henson] 1009238405Sjkim 1010238405Sjkim *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to 1011238405Sjkim output hashes compatible with older versions of OpenSSL. 1012238405Sjkim [Willy Weisz <weisz@vcpc.univie.ac.at>] 1013238405Sjkim 1014238405Sjkim *) Fix compression algorithm handling: if resuming a session use the 1015238405Sjkim compression algorithm of the resumed session instead of determining 1016238405Sjkim it from client hello again. Don't allow server to change algorithm. 1017238405Sjkim [Steve Henson] 1018238405Sjkim 1019238405Sjkim *) Add load_crls() function to apps tidying load_certs() too. Add option 1020238405Sjkim to verify utility to allow additional CRLs to be included. 1021238405Sjkim [Steve Henson] 1022238405Sjkim 1023238405Sjkim *) Update OCSP request code to permit adding custom headers to the request: 1024238405Sjkim some responders need this. 1025238405Sjkim [Steve Henson] 1026238405Sjkim 1027238405Sjkim *) The function EVP_PKEY_sign() returns <=0 on error: check return code 1028238405Sjkim correctly. 1029238405Sjkim [Julia Lawall <julia@diku.dk>] 1030238405Sjkim 1031238405Sjkim *) Update verify callback code in apps/s_cb.c and apps/verify.c, it 1032238405Sjkim needlessly dereferenced structures, used obsolete functions and 1033238405Sjkim didn't handle all updated verify codes correctly. 1034238405Sjkim [Steve Henson] 1035238405Sjkim 1036238405Sjkim *) Disable MD2 in the default configuration. 1037238405Sjkim [Steve Henson] 1038238405Sjkim 1039238405Sjkim *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to 1040238405Sjkim indicate the initial BIO being pushed or popped. This makes it possible 1041238405Sjkim to determine whether the BIO is the one explicitly called or as a result 1042238405Sjkim of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so 1043238405Sjkim it handles reference counts correctly and doesn't zero out the I/O bio 1044238405Sjkim when it is not being explicitly popped. WARNING: applications which 1045238405Sjkim included workarounds for the old buggy behaviour will need to be modified 1046238405Sjkim or they could free up already freed BIOs. 1047238405Sjkim [Steve Henson] 1048238405Sjkim 1049238405Sjkim *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni 1050238405Sjkim renaming to all platforms (within the 0.9.8 branch, this was 1051238405Sjkim done conditionally on Netware platforms to avoid a name clash). 1052238405Sjkim [Guenter <lists@gknw.net>] 1053238405Sjkim 1054238405Sjkim *) Add ECDHE and PSK support to DTLS. 1055238405Sjkim [Michael Tuexen <tuexen@fh-muenster.de>] 1056238405Sjkim 1057238405Sjkim *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't 1058238405Sjkim be used on C++. 1059238405Sjkim [Steve Henson] 1060238405Sjkim 1061238405Sjkim *) Add "missing" function EVP_MD_flags() (without this the only way to 1062238405Sjkim retrieve a digest flags is by accessing the structure directly. Update 1063238405Sjkim EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest 1064238405Sjkim or cipher is registered as in the "from" argument. Print out all 1065238405Sjkim registered digests in the dgst usage message instead of manually 1066238405Sjkim attempting to work them out. 1067238405Sjkim [Steve Henson] 1068238405Sjkim 1069238405Sjkim *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello: 1070238405Sjkim this allows the use of compression and extensions. Change default cipher 1071238405Sjkim string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2 1072238405Sjkim by default unless an application cipher string requests it. 1073238405Sjkim [Steve Henson] 1074238405Sjkim 1075238405Sjkim *) Alter match criteria in PKCS12_parse(). It used to try to use local 1076238405Sjkim key ids to find matching certificates and keys but some PKCS#12 files 1077238405Sjkim don't follow the (somewhat unwritten) rules and this strategy fails. 1078238405Sjkim Now just gather all certificates together and the first private key 1079238405Sjkim then look for the first certificate that matches the key. 1080238405Sjkim [Steve Henson] 1081238405Sjkim 1082238405Sjkim *) Support use of registered digest and cipher names for dgst and cipher 1083238405Sjkim commands instead of having to add each one as a special case. So now 1084238405Sjkim you can do: 1085238405Sjkim 1086238405Sjkim openssl sha256 foo 1087238405Sjkim 1088238405Sjkim as well as: 1089238405Sjkim 1090238405Sjkim openssl dgst -sha256 foo 1091238405Sjkim 1092238405Sjkim and this works for ENGINE based algorithms too. 1093238405Sjkim 1094238405Sjkim [Steve Henson] 1095238405Sjkim 1096238405Sjkim *) Update Gost ENGINE to support parameter files. 1097238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 1098238405Sjkim 1099238405Sjkim *) Support GeneralizedTime in ca utility. 1100238405Sjkim [Oliver Martin <oliver@volatilevoid.net>, Steve Henson] 1101238405Sjkim 1102238405Sjkim *) Enhance the hash format used for certificate directory links. The new 1103238405Sjkim form uses the canonical encoding (meaning equivalent names will work 1104238405Sjkim even if they aren't identical) and uses SHA1 instead of MD5. This form 1105238405Sjkim is incompatible with the older format and as a result c_rehash should 1106238405Sjkim be used to rebuild symbolic links. 1107238405Sjkim [Steve Henson] 1108238405Sjkim 1109238405Sjkim *) Make PKCS#8 the default write format for private keys, replacing the 1110238405Sjkim traditional format. This form is standardised, more secure and doesn't 1111238405Sjkim include an implicit MD5 dependency. 1112238405Sjkim [Steve Henson] 1113238405Sjkim 1114238405Sjkim *) Add a $gcc_devteam_warn option to Configure. The idea is that any code 1115238405Sjkim committed to OpenSSL should pass this lot as a minimum. 1116238405Sjkim [Steve Henson] 1117238405Sjkim 1118238405Sjkim *) Add session ticket override functionality for use by EAP-FAST. 1119238405Sjkim [Jouni Malinen <j@w1.fi>] 1120238405Sjkim 1121238405Sjkim *) Modify HMAC functions to return a value. Since these can be implemented 1122238405Sjkim in an ENGINE errors can occur. 1123238405Sjkim [Steve Henson] 1124238405Sjkim 1125238405Sjkim *) Type-checked OBJ_bsearch_ex. 1126238405Sjkim [Ben Laurie] 1127238405Sjkim 1128238405Sjkim *) Type-checked OBJ_bsearch. Also some constification necessitated 1129238405Sjkim by type-checking. Still to come: TXT_DB, bsearch(?), 1130238405Sjkim OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING, 1131238405Sjkim CONF_VALUE. 1132238405Sjkim [Ben Laurie] 1133238405Sjkim 1134238405Sjkim *) New function OPENSSL_gmtime_adj() to add a specific number of days and 1135238405Sjkim seconds to a tm structure directly, instead of going through OS 1136238405Sjkim specific date routines. This avoids any issues with OS routines such 1137238405Sjkim as the year 2038 bug. New *_adj() functions for ASN1 time structures 1138238405Sjkim and X509_time_adj_ex() to cover the extended range. The existing 1139238405Sjkim X509_time_adj() is still usable and will no longer have any date issues. 1140238405Sjkim [Steve Henson] 1141238405Sjkim 1142238405Sjkim *) Delta CRL support. New use deltas option which will attempt to locate 1143238405Sjkim and search any appropriate delta CRLs available. 1144238405Sjkim 1145238405Sjkim This work was sponsored by Google. 1146238405Sjkim [Steve Henson] 1147238405Sjkim 1148238405Sjkim *) Support for CRLs partitioned by reason code. Reorganise CRL processing 1149238405Sjkim code and add additional score elements. Validate alternate CRL paths 1150238405Sjkim as part of the CRL checking and indicate a new error "CRL path validation 1151238405Sjkim error" in this case. Applications wanting additional details can use 1152238405Sjkim the verify callback and check the new "parent" field. If this is not 1153238405Sjkim NULL CRL path validation is taking place. Existing applications wont 1154238405Sjkim see this because it requires extended CRL support which is off by 1155238405Sjkim default. 1156238405Sjkim 1157238405Sjkim This work was sponsored by Google. 1158238405Sjkim [Steve Henson] 1159238405Sjkim 1160238405Sjkim *) Support for freshest CRL extension. 1161238405Sjkim 1162238405Sjkim This work was sponsored by Google. 1163238405Sjkim [Steve Henson] 1164238405Sjkim 1165238405Sjkim *) Initial indirect CRL support. Currently only supported in the CRLs 1166238405Sjkim passed directly and not via lookup. Process certificate issuer 1167238405Sjkim CRL entry extension and lookup CRL entries by bother issuer name 1168238405Sjkim and serial number. Check and process CRL issuer entry in IDP extension. 1169238405Sjkim 1170238405Sjkim This work was sponsored by Google. 1171238405Sjkim [Steve Henson] 1172238405Sjkim 1173238405Sjkim *) Add support for distinct certificate and CRL paths. The CRL issuer 1174238405Sjkim certificate is validated separately in this case. Only enabled if 1175238405Sjkim an extended CRL support flag is set: this flag will enable additional 1176238405Sjkim CRL functionality in future. 1177238405Sjkim 1178238405Sjkim This work was sponsored by Google. 1179238405Sjkim [Steve Henson] 1180238405Sjkim 1181238405Sjkim *) Add support for policy mappings extension. 1182238405Sjkim 1183238405Sjkim This work was sponsored by Google. 1184238405Sjkim [Steve Henson] 1185238405Sjkim 1186238405Sjkim *) Fixes to pathlength constraint, self issued certificate handling, 1187238405Sjkim policy processing to align with RFC3280 and PKITS tests. 1188238405Sjkim 1189238405Sjkim This work was sponsored by Google. 1190238405Sjkim [Steve Henson] 1191238405Sjkim 1192238405Sjkim *) Support for name constraints certificate extension. DN, email, DNS 1193238405Sjkim and URI types are currently supported. 1194238405Sjkim 1195238405Sjkim This work was sponsored by Google. 1196238405Sjkim [Steve Henson] 1197238405Sjkim 1198238405Sjkim *) To cater for systems that provide a pointer-based thread ID rather 1199238405Sjkim than numeric, deprecate the current numeric thread ID mechanism and 1200238405Sjkim replace it with a structure and associated callback type. This 1201238405Sjkim mechanism allows a numeric "hash" to be extracted from a thread ID in 1202238405Sjkim either case, and on platforms where pointers are larger than 'long', 1203238405Sjkim mixing is done to help ensure the numeric 'hash' is usable even if it 1204238405Sjkim can't be guaranteed unique. The default mechanism is to use "&errno" 1205238405Sjkim as a pointer-based thread ID to distinguish between threads. 1206238405Sjkim 1207238405Sjkim Applications that want to provide their own thread IDs should now use 1208238405Sjkim CRYPTO_THREADID_set_callback() to register a callback that will call 1209238405Sjkim either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer(). 1210238405Sjkim 1211238405Sjkim Note that ERR_remove_state() is now deprecated, because it is tied 1212238405Sjkim to the assumption that thread IDs are numeric. ERR_remove_state(0) 1213238405Sjkim to free the current thread's error state should be replaced by 1214238405Sjkim ERR_remove_thread_state(NULL). 1215238405Sjkim 1216238405Sjkim (This new approach replaces the functions CRYPTO_set_idptr_callback(), 1217238405Sjkim CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in 1218238405Sjkim OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an 1219238405Sjkim application was previously providing a numeric thread callback that 1220238405Sjkim was inappropriate for distinguishing threads, then uniqueness might 1221238405Sjkim have been obtained with &errno that happened immediately in the 1222238405Sjkim intermediate development versions of OpenSSL; this is no longer the 1223238405Sjkim case, the numeric thread callback will now override the automatic use 1224238405Sjkim of &errno.) 1225238405Sjkim [Geoff Thorpe, with help from Bodo Moeller] 1226238405Sjkim 1227238405Sjkim *) Initial support for different CRL issuing certificates. This covers a 1228238405Sjkim simple case where the self issued certificates in the chain exist and 1229238405Sjkim the real CRL issuer is higher in the existing chain. 1230238405Sjkim 1231238405Sjkim This work was sponsored by Google. 1232238405Sjkim [Steve Henson] 1233238405Sjkim 1234238405Sjkim *) Removed effectively defunct crypto/store from the build. 1235238405Sjkim [Ben Laurie] 1236238405Sjkim 1237238405Sjkim *) Revamp of STACK to provide stronger type-checking. Still to come: 1238238405Sjkim TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE, 1239238405Sjkim ASN1_STRING, CONF_VALUE. 1240238405Sjkim [Ben Laurie] 1241238405Sjkim 1242238405Sjkim *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer 1243238405Sjkim RAM on SSL connections. This option can save about 34k per idle SSL. 1244238405Sjkim [Nick Mathewson] 1245238405Sjkim 1246238405Sjkim *) Revamp of LHASH to provide stronger type-checking. Still to come: 1247238405Sjkim STACK, TXT_DB, bsearch, qsort. 1248238405Sjkim [Ben Laurie] 1249238405Sjkim 1250238405Sjkim *) Initial support for Cryptographic Message Syntax (aka CMS) based 1251238405Sjkim on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility, 1252238405Sjkim support for data, signedData, compressedData, digestedData and 1253238405Sjkim encryptedData, envelopedData types included. Scripts to check against 1254238405Sjkim RFC4134 examples draft and interop and consistency checks of many 1255238405Sjkim content types and variants. 1256238405Sjkim [Steve Henson] 1257238405Sjkim 1258238405Sjkim *) Add options to enc utility to support use of zlib compression BIO. 1259238405Sjkim [Steve Henson] 1260238405Sjkim 1261238405Sjkim *) Extend mk1mf to support importing of options and assembly language 1262238405Sjkim files from Configure script, currently only included in VC-WIN32. 1263238405Sjkim The assembly language rules can now optionally generate the source 1264238405Sjkim files from the associated perl scripts. 1265238405Sjkim [Steve Henson] 1266238405Sjkim 1267238405Sjkim *) Implement remaining functionality needed to support GOST ciphersuites. 1268238405Sjkim Interop testing has been performed using CryptoPro implementations. 1269238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 1270238405Sjkim 1271238405Sjkim *) s390x assembler pack. 1272238405Sjkim [Andy Polyakov] 1273238405Sjkim 1274238405Sjkim *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU 1275238405Sjkim "family." 1276238405Sjkim [Andy Polyakov] 1277238405Sjkim 1278238405Sjkim *) Implement Opaque PRF Input TLS extension as specified in 1279238405Sjkim draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an 1280238405Sjkim official specification yet and no extension type assignment by 1281238405Sjkim IANA exists, this extension (for now) will have to be explicitly 1282238405Sjkim enabled when building OpenSSL by providing the extension number 1283238405Sjkim to use. For example, specify an option 1284238405Sjkim 1285238405Sjkim -DTLSEXT_TYPE_opaque_prf_input=0x9527 1286238405Sjkim 1287238405Sjkim to the "config" or "Configure" script to enable the extension, 1288238405Sjkim assuming extension number 0x9527 (which is a completely arbitrary 1289238405Sjkim and unofficial assignment based on the MD5 hash of the Internet 1290238405Sjkim Draft). Note that by doing so, you potentially lose 1291238405Sjkim interoperability with other TLS implementations since these might 1292238405Sjkim be using the same extension number for other purposes. 1293238405Sjkim 1294238405Sjkim SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the 1295238405Sjkim opaque PRF input value to use in the handshake. This will create 1296238405Sjkim an interal copy of the length-'len' string at 'src', and will 1297238405Sjkim return non-zero for success. 1298238405Sjkim 1299238405Sjkim To get more control and flexibility, provide a callback function 1300238405Sjkim by using 1301238405Sjkim 1302238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb) 1303238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg) 1304238405Sjkim 1305238405Sjkim where 1306238405Sjkim 1307238405Sjkim int (*cb)(SSL *, void *peerinput, size_t len, void *arg); 1308238405Sjkim void *arg; 1309238405Sjkim 1310238405Sjkim Callback function 'cb' will be called in handshakes, and is 1311238405Sjkim expected to use SSL_set_tlsext_opaque_prf_input() as appropriate. 1312238405Sjkim Argument 'arg' is for application purposes (the value as given to 1313238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly 1314238405Sjkim be provided to the callback function). The callback function 1315238405Sjkim has to return non-zero to report success: usually 1 to use opaque 1316238405Sjkim PRF input just if possible, or 2 to enforce use of the opaque PRF 1317238405Sjkim input. In the latter case, the library will abort the handshake 1318238405Sjkim if opaque PRF input is not successfully negotiated. 1319238405Sjkim 1320238405Sjkim Arguments 'peerinput' and 'len' given to the callback function 1321238405Sjkim will always be NULL and 0 in the case of a client. A server will 1322238405Sjkim see the client's opaque PRF input through these variables if 1323238405Sjkim available (NULL and 0 otherwise). Note that if the server 1324238405Sjkim provides an opaque PRF input, the length must be the same as the 1325238405Sjkim length of the client's opaque PRF input. 1326238405Sjkim 1327238405Sjkim Note that the callback function will only be called when creating 1328238405Sjkim a new session (session resumption can resume whatever was 1329238405Sjkim previously negotiated), and will not be called in SSL 2.0 1330238405Sjkim handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or 1331238405Sjkim SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended 1332238405Sjkim for applications that need to enforce opaque PRF input. 1333238405Sjkim 1334238405Sjkim [Bodo Moeller] 1335238405Sjkim 1336238405Sjkim *) Update ssl code to support digests other than SHA1+MD5 for handshake 1337238405Sjkim MAC. 1338238405Sjkim 1339238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 1340238405Sjkim 1341238405Sjkim *) Add RFC4507 support to OpenSSL. This includes the corrections in 1342238405Sjkim RFC4507bis. The encrypted ticket format is an encrypted encoded 1343238405Sjkim SSL_SESSION structure, that way new session features are automatically 1344238405Sjkim supported. 1345238405Sjkim 1346238405Sjkim If a client application caches session in an SSL_SESSION structure 1347238405Sjkim support is transparent because tickets are now stored in the encoded 1348238405Sjkim SSL_SESSION. 1349238405Sjkim 1350238405Sjkim The SSL_CTX structure automatically generates keys for ticket 1351238405Sjkim protection in servers so again support should be possible 1352238405Sjkim with no application modification. 1353238405Sjkim 1354238405Sjkim If a client or server wishes to disable RFC4507 support then the option 1355238405Sjkim SSL_OP_NO_TICKET can be set. 1356238405Sjkim 1357238405Sjkim Add a TLS extension debugging callback to allow the contents of any client 1358238405Sjkim or server extensions to be examined. 1359238405Sjkim 1360238405Sjkim This work was sponsored by Google. 1361238405Sjkim [Steve Henson] 1362238405Sjkim 1363238405Sjkim *) Final changes to avoid use of pointer pointer casts in OpenSSL. 1364238405Sjkim OpenSSL should now compile cleanly on gcc 4.2 1365238405Sjkim [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson] 1366238405Sjkim 1367238405Sjkim *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC 1368238405Sjkim support including streaming MAC support: this is required for GOST 1369238405Sjkim ciphersuite support. 1370238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson] 1371238405Sjkim 1372238405Sjkim *) Add option -stream to use PKCS#7 streaming in smime utility. New 1373238405Sjkim function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream() 1374238405Sjkim to output in BER and PEM format. 1375238405Sjkim [Steve Henson] 1376238405Sjkim 1377238405Sjkim *) Experimental support for use of HMAC via EVP_PKEY interface. This 1378238405Sjkim allows HMAC to be handled via the EVP_DigestSign*() interface. The 1379238405Sjkim EVP_PKEY "key" in this case is the HMAC key, potentially allowing 1380238405Sjkim ENGINE support for HMAC keys which are unextractable. New -mac and 1381238405Sjkim -macopt options to dgst utility. 1382238405Sjkim [Steve Henson] 1383238405Sjkim 1384238405Sjkim *) New option -sigopt to dgst utility. Update dgst to use 1385238405Sjkim EVP_Digest{Sign,Verify}*. These two changes make it possible to use 1386238405Sjkim alternative signing paramaters such as X9.31 or PSS in the dgst 1387238405Sjkim utility. 1388238405Sjkim [Steve Henson] 1389238405Sjkim 1390238405Sjkim *) Change ssl_cipher_apply_rule(), the internal function that does 1391238405Sjkim the work each time a ciphersuite string requests enabling 1392238405Sjkim ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or 1393238405Sjkim removing ("!foo+bar") a class of ciphersuites: Now it maintains 1394238405Sjkim the order of disabled ciphersuites such that those ciphersuites 1395238405Sjkim that most recently went from enabled to disabled not only stay 1396238405Sjkim in order with respect to each other, but also have higher priority 1397238405Sjkim than other disabled ciphersuites the next time ciphersuites are 1398238405Sjkim enabled again. 1399238405Sjkim 1400238405Sjkim This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable 1401238405Sjkim the same ciphersuites as with "HIGH" alone, but in a specific 1402238405Sjkim order where the PSK ciphersuites come first (since they are the 1403238405Sjkim most recently disabled ciphersuites when "HIGH" is parsed). 1404238405Sjkim 1405238405Sjkim Also, change ssl_create_cipher_list() (using this new 1406238405Sjkim funcionality) such that between otherwise identical 1407238405Sjkim cihpersuites, ephemeral ECDH is preferred over ephemeral DH in 1408238405Sjkim the default order. 1409238405Sjkim [Bodo Moeller] 1410238405Sjkim 1411238405Sjkim *) Change ssl_create_cipher_list() so that it automatically 1412238405Sjkim arranges the ciphersuites in reasonable order before starting 1413238405Sjkim to process the rule string. Thus, the definition for "DEFAULT" 1414238405Sjkim (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but 1415238405Sjkim remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH". 1416238405Sjkim This makes it much easier to arrive at a reasonable default order 1417238405Sjkim in applications for which anonymous ciphers are OK (meaning 1418238405Sjkim that you can't actually use DEFAULT). 1419238405Sjkim [Bodo Moeller; suggested by Victor Duchovni] 1420238405Sjkim 1421238405Sjkim *) Split the SSL/TLS algorithm mask (as used for ciphersuite string 1422238405Sjkim processing) into multiple integers instead of setting 1423238405Sjkim "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK", 1424238405Sjkim "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer. 1425238405Sjkim (These masks as well as the individual bit definitions are hidden 1426238405Sjkim away into the non-exported interface ssl/ssl_locl.h, so this 1427238405Sjkim change to the definition of the SSL_CIPHER structure shouldn't 1428238405Sjkim affect applications.) This give us more bits for each of these 1429238405Sjkim categories, so there is no longer a need to coagulate AES128 and 1430238405Sjkim AES256 into a single algorithm bit, and to coagulate Camellia128 1431238405Sjkim and Camellia256 into a single algorithm bit, which has led to all 1432238405Sjkim kinds of kludges. 1433238405Sjkim 1434238405Sjkim Thus, among other things, the kludge introduced in 0.9.7m and 1435238405Sjkim 0.9.8e for masking out AES256 independently of AES128 or masking 1436238405Sjkim out Camellia256 independently of AES256 is not needed here in 0.9.9. 1437238405Sjkim 1438238405Sjkim With the change, we also introduce new ciphersuite aliases that 1439238405Sjkim so far were missing: "AES128", "AES256", "CAMELLIA128", and 1440238405Sjkim "CAMELLIA256". 1441238405Sjkim [Bodo Moeller] 1442238405Sjkim 1443238405Sjkim *) Add support for dsa-with-SHA224 and dsa-with-SHA256. 1444238405Sjkim Use the leftmost N bytes of the signature input if the input is 1445238405Sjkim larger than the prime q (with N being the size in bytes of q). 1446238405Sjkim [Nils Larsch] 1447238405Sjkim 1448238405Sjkim *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses 1449238405Sjkim it yet and it is largely untested. 1450238405Sjkim [Steve Henson] 1451238405Sjkim 1452238405Sjkim *) Add support for the ecdsa-with-SHA224/256/384/512 signature types. 1453238405Sjkim [Nils Larsch] 1454238405Sjkim 1455238405Sjkim *) Initial incomplete changes to avoid need for function casts in OpenSSL 1456238405Sjkim some compilers (gcc 4.2 and later) reject their use. Safestack is 1457238405Sjkim reimplemented. Update ASN1 to avoid use of legacy functions. 1458238405Sjkim [Steve Henson] 1459238405Sjkim 1460238405Sjkim *) Win32/64 targets are linked with Winsock2. 1461238405Sjkim [Andy Polyakov] 1462238405Sjkim 1463238405Sjkim *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected 1464238405Sjkim to external functions. This can be used to increase CRL handling 1465238405Sjkim efficiency especially when CRLs are very large by (for example) storing 1466238405Sjkim the CRL revoked certificates in a database. 1467238405Sjkim [Steve Henson] 1468238405Sjkim 1469238405Sjkim *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so 1470238405Sjkim new CRLs added to a directory can be used. New command line option 1471238405Sjkim -verify_return_error to s_client and s_server. This causes real errors 1472238405Sjkim to be returned by the verify callback instead of carrying on no matter 1473238405Sjkim what. This reflects the way a "real world" verify callback would behave. 1474238405Sjkim [Steve Henson] 1475238405Sjkim 1476238405Sjkim *) GOST engine, supporting several GOST algorithms and public key formats. 1477238405Sjkim Kindly donated by Cryptocom. 1478238405Sjkim [Cryptocom] 1479238405Sjkim 1480238405Sjkim *) Partial support for Issuing Distribution Point CRL extension. CRLs 1481238405Sjkim partitioned by DP are handled but no indirect CRL or reason partitioning 1482238405Sjkim (yet). Complete overhaul of CRL handling: now the most suitable CRL is 1483238405Sjkim selected via a scoring technique which handles IDP and AKID in CRLs. 1484238405Sjkim [Steve Henson] 1485238405Sjkim 1486238405Sjkim *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which 1487238405Sjkim will ultimately be used for all verify operations: this will remove the 1488238405Sjkim X509_STORE dependency on certificate verification and allow alternative 1489238405Sjkim lookup methods. X509_STORE based implementations of these two callbacks. 1490238405Sjkim [Steve Henson] 1491238405Sjkim 1492238405Sjkim *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names. 1493238405Sjkim Modify get_crl() to find a valid (unexpired) CRL if possible. 1494238405Sjkim [Steve Henson] 1495238405Sjkim 1496238405Sjkim *) New function X509_CRL_match() to check if two CRLs are identical. Normally 1497238405Sjkim this would be called X509_CRL_cmp() but that name is already used by 1498238405Sjkim a function that just compares CRL issuer names. Cache several CRL 1499238405Sjkim extensions in X509_CRL structure and cache CRLDP in X509. 1500238405Sjkim [Steve Henson] 1501238405Sjkim 1502238405Sjkim *) Store a "canonical" representation of X509_NAME structure (ASN1 Name) 1503238405Sjkim this maps equivalent X509_NAME structures into a consistent structure. 1504238405Sjkim Name comparison can then be performed rapidly using memcmp(). 1505238405Sjkim [Steve Henson] 1506238405Sjkim 1507238405Sjkim *) Non-blocking OCSP request processing. Add -timeout option to ocsp 1508238405Sjkim utility. 1509238405Sjkim [Steve Henson] 1510238405Sjkim 1511238405Sjkim *) Allow digests to supply their own micalg string for S/MIME type using 1512238405Sjkim the ctrl EVP_MD_CTRL_MICALG. 1513238405Sjkim [Steve Henson] 1514238405Sjkim 1515238405Sjkim *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the 1516238405Sjkim EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN 1517238405Sjkim ctrl. It can then customise the structure before and/or after signing 1518238405Sjkim if necessary. 1519238405Sjkim [Steve Henson] 1520238405Sjkim 1521238405Sjkim *) New function OBJ_add_sigid() to allow application defined signature OIDs 1522238405Sjkim to be added to OpenSSLs internal tables. New function OBJ_sigid_free() 1523238405Sjkim to free up any added signature OIDs. 1524238405Sjkim [Steve Henson] 1525238405Sjkim 1526238405Sjkim *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(), 1527238405Sjkim EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal 1528238405Sjkim digest and cipher tables. New options added to openssl utility: 1529238405Sjkim list-message-digest-algorithms and list-cipher-algorithms. 1530238405Sjkim [Steve Henson] 1531238405Sjkim 1532238405Sjkim *) Change the array representation of binary polynomials: the list 1533238405Sjkim of degrees of non-zero coefficients is now terminated with -1. 1534238405Sjkim Previously it was terminated with 0, which was also part of the 1535238405Sjkim value; thus, the array representation was not applicable to 1536238405Sjkim polynomials where t^0 has coefficient zero. This change makes 1537238405Sjkim the array representation useful in a more general context. 1538238405Sjkim [Douglas Stebila] 1539238405Sjkim 1540238405Sjkim *) Various modifications and fixes to SSL/TLS cipher string 1541238405Sjkim handling. For ECC, the code now distinguishes between fixed ECDH 1542238405Sjkim with RSA certificates on the one hand and with ECDSA certificates 1543238405Sjkim on the other hand, since these are separate ciphersuites. The 1544238405Sjkim unused code for Fortezza ciphersuites has been removed. 1545238405Sjkim 1546238405Sjkim For consistency with EDH, ephemeral ECDH is now called "EECDH" 1547238405Sjkim (not "ECDHE"). For consistency with the code for DH 1548238405Sjkim certificates, use of ECDH certificates is now considered ECDH 1549238405Sjkim authentication, not RSA or ECDSA authentication (the latter is 1550238405Sjkim merely the CA's signing algorithm and not actively used in the 1551238405Sjkim protocol). 1552238405Sjkim 1553238405Sjkim The temporary ciphersuite alias "ECCdraft" is no longer 1554238405Sjkim available, and ECC ciphersuites are no longer excluded from "ALL" 1555238405Sjkim and "DEFAULT". The following aliases now exist for RFC 4492 1556238405Sjkim ciphersuites, most of these by analogy with the DH case: 1557238405Sjkim 1558238405Sjkim kECDHr - ECDH cert, signed with RSA 1559238405Sjkim kECDHe - ECDH cert, signed with ECDSA 1560238405Sjkim kECDH - ECDH cert (signed with either RSA or ECDSA) 1561238405Sjkim kEECDH - ephemeral ECDH 1562238405Sjkim ECDH - ECDH cert or ephemeral ECDH 1563238405Sjkim 1564238405Sjkim aECDH - ECDH cert 1565238405Sjkim aECDSA - ECDSA cert 1566238405Sjkim ECDSA - ECDSA cert 1567238405Sjkim 1568238405Sjkim AECDH - anonymous ECDH 1569238405Sjkim EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH") 1570238405Sjkim 1571238405Sjkim [Bodo Moeller] 1572238405Sjkim 1573238405Sjkim *) Add additional S/MIME capabilities for AES and GOST ciphers if supported. 1574238405Sjkim Use correct micalg parameters depending on digest(s) in signed message. 1575238405Sjkim [Steve Henson] 1576238405Sjkim 1577238405Sjkim *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process 1578238405Sjkim an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code. 1579238405Sjkim [Steve Henson] 1580238405Sjkim 1581238405Sjkim *) Initial engine support for EVP_PKEY_METHOD. New functions to permit 1582238405Sjkim an engine to register a method. Add ENGINE lookups for methods and 1583238405Sjkim functional reference processing. 1584238405Sjkim [Steve Henson] 1585238405Sjkim 1586238405Sjkim *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of 1587238405Sjkim EVP_{Sign,Verify}* which allow an application to customise the signature 1588238405Sjkim process. 1589238405Sjkim [Steve Henson] 1590238405Sjkim 1591238405Sjkim *) New -resign option to smime utility. This adds one or more signers 1592238405Sjkim to an existing PKCS#7 signedData structure. Also -md option to use an 1593238405Sjkim alternative message digest algorithm for signing. 1594238405Sjkim [Steve Henson] 1595238405Sjkim 1596238405Sjkim *) Tidy up PKCS#7 routines and add new functions to make it easier to 1597238405Sjkim create PKCS7 structures containing multiple signers. Update smime 1598238405Sjkim application to support multiple signers. 1599238405Sjkim [Steve Henson] 1600238405Sjkim 1601238405Sjkim *) New -macalg option to pkcs12 utility to allow setting of an alternative 1602238405Sjkim digest MAC. 1603238405Sjkim [Steve Henson] 1604238405Sjkim 1605238405Sjkim *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC. 1606238405Sjkim Reorganize PBE internals to lookup from a static table using NIDs, 1607238405Sjkim add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl: 1608238405Sjkim EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative 1609238405Sjkim PRF which will be automatically used with PBES2. 1610238405Sjkim [Steve Henson] 1611238405Sjkim 1612238405Sjkim *) Replace the algorithm specific calls to generate keys in "req" with the 1613238405Sjkim new API. 1614238405Sjkim [Steve Henson] 1615238405Sjkim 1616238405Sjkim *) Update PKCS#7 enveloped data routines to use new API. This is now 1617238405Sjkim supported by any public key method supporting the encrypt operation. A 1618238405Sjkim ctrl is added to allow the public key algorithm to examine or modify 1619238405Sjkim the PKCS#7 RecipientInfo structure if it needs to: for RSA this is 1620238405Sjkim a no op. 1621238405Sjkim [Steve Henson] 1622238405Sjkim 1623238405Sjkim *) Add a ctrl to asn1 method to allow a public key algorithm to express 1624238405Sjkim a default digest type to use. In most cases this will be SHA1 but some 1625238405Sjkim algorithms (such as GOST) need to specify an alternative digest. The 1626238405Sjkim return value indicates how strong the prefernce is 1 means optional and 1627238405Sjkim 2 is mandatory (that is it is the only supported type). Modify 1628238405Sjkim ASN1_item_sign() to accept a NULL digest argument to indicate it should 1629238405Sjkim use the default md. Update openssl utilities to use the default digest 1630238405Sjkim type for signing if it is not explicitly indicated. 1631238405Sjkim [Steve Henson] 1632238405Sjkim 1633238405Sjkim *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New 1634238405Sjkim EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant 1635238405Sjkim signing method from the key type. This effectively removes the link 1636238405Sjkim between digests and public key types. 1637238405Sjkim [Steve Henson] 1638238405Sjkim 1639238405Sjkim *) Add an OID cross reference table and utility functions. Its purpose is to 1640238405Sjkim translate between signature OIDs such as SHA1WithrsaEncryption and SHA1, 1641238405Sjkim rsaEncryption. This will allow some of the algorithm specific hackery 1642238405Sjkim needed to use the correct OID to be removed. 1643238405Sjkim [Steve Henson] 1644238405Sjkim 1645238405Sjkim *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO 1646238405Sjkim structures for PKCS7_sign(). They are now set up by the relevant public 1647238405Sjkim key ASN1 method. 1648238405Sjkim [Steve Henson] 1649238405Sjkim 1650238405Sjkim *) Add provisional EC pkey method with support for ECDSA and ECDH. 1651238405Sjkim [Steve Henson] 1652238405Sjkim 1653238405Sjkim *) Add support for key derivation (agreement) in the API, DH method and 1654238405Sjkim pkeyutl. 1655238405Sjkim [Steve Henson] 1656238405Sjkim 1657238405Sjkim *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support 1658238405Sjkim public and private key formats. As a side effect these add additional 1659238405Sjkim command line functionality not previously available: DSA signatures can be 1660238405Sjkim generated and verified using pkeyutl and DH key support and generation in 1661238405Sjkim pkey, genpkey. 1662238405Sjkim [Steve Henson] 1663238405Sjkim 1664238405Sjkim *) BeOS support. 1665238405Sjkim [Oliver Tappe <zooey@hirschkaefer.de>] 1666238405Sjkim 1667238405Sjkim *) New make target "install_html_docs" installs HTML renditions of the 1668238405Sjkim manual pages. 1669238405Sjkim [Oliver Tappe <zooey@hirschkaefer.de>] 1670238405Sjkim 1671238405Sjkim *) New utility "genpkey" this is analagous to "genrsa" etc except it can 1672238405Sjkim generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to 1673238405Sjkim support key and parameter generation and add initial key generation 1674238405Sjkim functionality for RSA. 1675238405Sjkim [Steve Henson] 1676238405Sjkim 1677238405Sjkim *) Add functions for main EVP_PKEY_method operations. The undocumented 1678238405Sjkim functions EVP_PKEY_{encrypt,decrypt} have been renamed to 1679238405Sjkim EVP_PKEY_{encrypt,decrypt}_old. 1680238405Sjkim [Steve Henson] 1681238405Sjkim 1682238405Sjkim *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public 1683238405Sjkim key API, doesn't do much yet. 1684238405Sjkim [Steve Henson] 1685238405Sjkim 1686238405Sjkim *) New function EVP_PKEY_asn1_get0_info() to retrieve information about 1687238405Sjkim public key algorithms. New option to openssl utility: 1688238405Sjkim "list-public-key-algorithms" to print out info. 1689238405Sjkim [Steve Henson] 1690238405Sjkim 1691238405Sjkim *) Implement the Supported Elliptic Curves Extension for 1692238405Sjkim ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1693238405Sjkim [Douglas Stebila] 1694238405Sjkim 1695238405Sjkim *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or 1696238405Sjkim EVP_CIPHER structures to avoid later problems in EVP_cleanup(). 1697238405Sjkim [Steve Henson] 1698238405Sjkim 1699238405Sjkim *) New utilities pkey and pkeyparam. These are similar to algorithm specific 1700238405Sjkim utilities such as rsa, dsa, dsaparam etc except they process any key 1701238405Sjkim type. 1702238405Sjkim [Steve Henson] 1703238405Sjkim 1704238405Sjkim *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New 1705238405Sjkim functions EVP_PKEY_print_public(), EVP_PKEY_print_private(), 1706238405Sjkim EVP_PKEY_print_param() to print public key data from an EVP_PKEY 1707238405Sjkim structure. 1708238405Sjkim [Steve Henson] 1709238405Sjkim 1710238405Sjkim *) Initial support for pluggable public key ASN1. 1711238405Sjkim De-spaghettify the public key ASN1 handling. Move public and private 1712238405Sjkim key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate 1713238405Sjkim algorithm specific handling to a single module within the relevant 1714238405Sjkim algorithm directory. Add functions to allow (near) opaque processing 1715238405Sjkim of public and private key structures. 1716238405Sjkim [Steve Henson] 1717238405Sjkim 1718238405Sjkim *) Implement the Supported Point Formats Extension for 1719238405Sjkim ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 1720238405Sjkim [Douglas Stebila] 1721238405Sjkim 1722238405Sjkim *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members 1723238405Sjkim for the psk identity [hint] and the psk callback functions to the 1724238405Sjkim SSL_SESSION, SSL and SSL_CTX structure. 1725238405Sjkim 1726238405Sjkim New ciphersuites: 1727238405Sjkim PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA, 1728238405Sjkim PSK-AES256-CBC-SHA 1729238405Sjkim 1730238405Sjkim New functions: 1731238405Sjkim SSL_CTX_use_psk_identity_hint 1732238405Sjkim SSL_get_psk_identity_hint 1733238405Sjkim SSL_get_psk_identity 1734238405Sjkim SSL_use_psk_identity_hint 1735238405Sjkim 1736238405Sjkim [Mika Kousa and Pasi Eronen of Nokia Corporation] 1737238405Sjkim 1738238405Sjkim *) Add RFC 3161 compliant time stamp request creation, response generation 1739238405Sjkim and response verification functionality. 1740238405Sjkim [Zolt�n Gl�zik <zglozik@opentsa.org>, The OpenTSA Project] 1741238405Sjkim 1742238405Sjkim *) Add initial support for TLS extensions, specifically for the server_name 1743238405Sjkim extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 1744238405Sjkim have new members for a host name. The SSL data structure has an 1745238405Sjkim additional member SSL_CTX *initial_ctx so that new sessions can be 1746238405Sjkim stored in that context to allow for session resumption, even after the 1747238405Sjkim SSL has been switched to a new SSL_CTX in reaction to a client's 1748238405Sjkim server_name extension. 1749238405Sjkim 1750238405Sjkim New functions (subject to change): 1751238405Sjkim 1752238405Sjkim SSL_get_servername() 1753238405Sjkim SSL_get_servername_type() 1754238405Sjkim SSL_set_SSL_CTX() 1755238405Sjkim 1756238405Sjkim New CTRL codes and macros (subject to change): 1757238405Sjkim 1758238405Sjkim SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 1759238405Sjkim - SSL_CTX_set_tlsext_servername_callback() 1760238405Sjkim SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 1761238405Sjkim - SSL_CTX_set_tlsext_servername_arg() 1762238405Sjkim SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 1763238405Sjkim 1764238405Sjkim openssl s_client has a new '-servername ...' option. 1765238405Sjkim 1766238405Sjkim openssl s_server has new options '-servername_host ...', '-cert2 ...', 1767238405Sjkim '-key2 ...', '-servername_fatal' (subject to change). This allows 1768238405Sjkim testing the HostName extension for a specific single host name ('-cert' 1769238405Sjkim and '-key' remain fallbacks for handshakes without HostName 1770238405Sjkim negotiation). If the unrecogninzed_name alert has to be sent, this by 1771238405Sjkim default is a warning; it becomes fatal with the '-servername_fatal' 1772238405Sjkim option. 1773238405Sjkim 1774238405Sjkim [Peter Sylvester, Remy Allais, Christophe Renou] 1775238405Sjkim 1776238405Sjkim *) Whirlpool hash implementation is added. 1777238405Sjkim [Andy Polyakov] 1778238405Sjkim 1779238405Sjkim *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to 1780238405Sjkim bn(64,32). Because of instruction set limitations it doesn't have 1781238405Sjkim any negative impact on performance. This was done mostly in order 1782238405Sjkim to make it possible to share assembler modules, such as bn_mul_mont 1783238405Sjkim implementations, between 32- and 64-bit builds without hassle. 1784238405Sjkim [Andy Polyakov] 1785238405Sjkim 1786238405Sjkim *) Move code previously exiled into file crypto/ec/ec2_smpt.c 1787238405Sjkim to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP 1788238405Sjkim macro. 1789238405Sjkim [Bodo Moeller] 1790238405Sjkim 1791238405Sjkim *) New candidate for BIGNUM assembler implementation, bn_mul_mont, 1792238405Sjkim dedicated Montgomery multiplication procedure, is introduced. 1793238405Sjkim BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher 1794238405Sjkim "64-bit" performance on certain 32-bit targets. 1795238405Sjkim [Andy Polyakov] 1796238405Sjkim 1797238405Sjkim *) New option SSL_OP_NO_COMP to disable use of compression selectively 1798238405Sjkim in SSL structures. New SSL ctrl to set maximum send fragment size. 1799238405Sjkim Save memory by seeting the I/O buffer sizes dynamically instead of 1800238405Sjkim using the maximum available value. 1801238405Sjkim [Steve Henson] 1802238405Sjkim 1803238405Sjkim *) New option -V for 'openssl ciphers'. This prints the ciphersuite code 1804238405Sjkim in addition to the text details. 1805238405Sjkim [Bodo Moeller] 1806238405Sjkim 1807238405Sjkim *) Very, very preliminary EXPERIMENTAL support for printing of general 1808238405Sjkim ASN1 structures. This currently produces rather ugly output and doesn't 1809238405Sjkim handle several customised structures at all. 1810238405Sjkim [Steve Henson] 1811238405Sjkim 1812238405Sjkim *) Integrated support for PVK file format and some related formats such 1813238405Sjkim as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support 1814238405Sjkim these in the 'rsa' and 'dsa' utilities. 1815238405Sjkim [Steve Henson] 1816238405Sjkim 1817238405Sjkim *) Support for PKCS#1 RSAPublicKey format on rsa utility command line. 1818238405Sjkim [Steve Henson] 1819238405Sjkim 1820238405Sjkim *) Remove the ancient ASN1_METHOD code. This was only ever used in one 1821238405Sjkim place for the (very old) "NETSCAPE" format certificates which are now 1822238405Sjkim handled using new ASN1 code equivalents. 1823238405Sjkim [Steve Henson] 1824238405Sjkim 1825238405Sjkim *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD 1826238405Sjkim pointer and make the SSL_METHOD parameter in SSL_CTX_new, 1827238405Sjkim SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'. 1828238405Sjkim [Nils Larsch] 1829238405Sjkim 1830238405Sjkim *) Modify CRL distribution points extension code to print out previously 1831238405Sjkim unsupported fields. Enhance extension setting code to allow setting of 1832238405Sjkim all fields. 1833238405Sjkim [Steve Henson] 1834238405Sjkim 1835238405Sjkim *) Add print and set support for Issuing Distribution Point CRL extension. 1836238405Sjkim [Steve Henson] 1837238405Sjkim 1838238405Sjkim *) Change 'Configure' script to enable Camellia by default. 1839238405Sjkim [NTT] 1840238405Sjkim 1841206046Ssimon Changes between 0.9.8m and 0.9.8n [24 Mar 2010] 1842206046Ssimon 1843206046Ssimon *) When rejecting SSL/TLS records due to an incorrect version number, never 1844206046Ssimon update s->server with a new major version number. As of 1845206046Ssimon - OpenSSL 0.9.8m if 'short' is a 16-bit type, 1846206046Ssimon - OpenSSL 0.9.8f if 'short' is longer than 16 bits, 1847206046Ssimon the previous behavior could result in a read attempt at NULL when 1848206046Ssimon receiving specific incorrect SSL/TLS records once record payload 1849206046Ssimon protection is active. (CVE-2010-0740) 1850206046Ssimon [Bodo Moeller, Adam Langley <agl@chromium.org>] 1851206046Ssimon 1852206046Ssimon *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL 1853206046Ssimon could be crashed if the relevant tables were not present (e.g. chrooted). 1854206046Ssimon [Tomas Hoger <thoger@redhat.com>] 1855206046Ssimon 1856205128Ssimon Changes between 0.9.8l and 0.9.8m [25 Feb 2010] 1857205128Ssimon 1858205128Ssimon *) Always check bn_wexpend() return values for failure. (CVE-2009-3245) 1859205128Ssimon [Martin Olsson, Neel Mehta] 1860205128Ssimon 1861205128Ssimon *) Fix X509_STORE locking: Every 'objs' access requires a lock (to 1862205128Ssimon accommodate for stack sorting, always a write lock!). 1863205128Ssimon [Bodo Moeller] 1864205128Ssimon 1865205128Ssimon *) On some versions of WIN32 Heap32Next is very slow. This can cause 1866205128Ssimon excessive delays in the RAND_poll(): over a minute. As a workaround 1867205128Ssimon include a time check in the inner Heap32Next loop too. 1868205128Ssimon [Steve Henson] 1869205128Ssimon 1870205128Ssimon *) The code that handled flushing of data in SSL/TLS originally used the 1871205128Ssimon BIO_CTRL_INFO ctrl to see if any data was pending first. This caused 1872205128Ssimon the problem outlined in PR#1949. The fix suggested there however can 1873205128Ssimon trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions 1874205128Ssimon of Apache). So instead simplify the code to flush unconditionally. 1875205128Ssimon This should be fine since flushing with no data to flush is a no op. 1876205128Ssimon [Steve Henson] 1877205128Ssimon 1878205128Ssimon *) Handle TLS versions 2.0 and later properly and correctly use the 1879205128Ssimon highest version of TLS/SSL supported. Although TLS >= 2.0 is some way 1880205128Ssimon off ancient servers have a habit of sticking around for a while... 1881205128Ssimon [Steve Henson] 1882205128Ssimon 1883205128Ssimon *) Modify compression code so it frees up structures without using the 1884205128Ssimon ex_data callbacks. This works around a problem where some applications 1885205128Ssimon call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when 1886205128Ssimon restarting) then use compression (e.g. SSL with compression) later. 1887205128Ssimon This results in significant per-connection memory leaks and 1888205128Ssimon has caused some security issues including CVE-2008-1678 and 1889205128Ssimon CVE-2009-4355. 1890205128Ssimon [Steve Henson] 1891205128Ssimon 1892205128Ssimon *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't 1893205128Ssimon change when encrypting or decrypting. 1894205128Ssimon [Bodo Moeller] 1895205128Ssimon 1896205128Ssimon *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to 1897205128Ssimon connect and renegotiate with servers which do not support RI. 1898205128Ssimon Until RI is more widely deployed this option is enabled by default. 1899205128Ssimon [Steve Henson] 1900205128Ssimon 1901205128Ssimon *) Add "missing" ssl ctrls to clear options and mode. 1902205128Ssimon [Steve Henson] 1903205128Ssimon 1904205128Ssimon *) If client attempts to renegotiate and doesn't support RI respond with 1905205128Ssimon a no_renegotiation alert as required by RFC5746. Some renegotiating 1906205128Ssimon TLS clients will continue a connection gracefully when they receive 1907205128Ssimon the alert. Unfortunately OpenSSL mishandled this alert and would hang 1908205128Ssimon waiting for a server hello which it will never receive. Now we treat a 1909205128Ssimon received no_renegotiation alert as a fatal error. This is because 1910205128Ssimon applications requesting a renegotiation might well expect it to succeed 1911205128Ssimon and would have no code in place to handle the server denying it so the 1912205128Ssimon only safe thing to do is to terminate the connection. 1913205128Ssimon [Steve Henson] 1914205128Ssimon 1915205128Ssimon *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if 1916205128Ssimon peer supports secure renegotiation and 0 otherwise. Print out peer 1917205128Ssimon renegotiation support in s_client/s_server. 1918205128Ssimon [Steve Henson] 1919205128Ssimon 1920205128Ssimon *) Replace the highly broken and deprecated SPKAC certification method with 1921205128Ssimon the updated NID creation version. This should correctly handle UTF8. 1922205128Ssimon [Steve Henson] 1923205128Ssimon 1924205128Ssimon *) Implement RFC5746. Re-enable renegotiation but require the extension 1925205128Ssimon as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION 1926205128Ssimon turns out to be a bad idea. It has been replaced by 1927205128Ssimon SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with 1928205128Ssimon SSL_CTX_set_options(). This is really not recommended unless you 1929205128Ssimon know what you are doing. 1930205128Ssimon [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson] 1931205128Ssimon 1932205128Ssimon *) Fixes to stateless session resumption handling. Use initial_ctx when 1933205128Ssimon issuing and attempting to decrypt tickets in case it has changed during 1934205128Ssimon servername handling. Use a non-zero length session ID when attempting 1935205128Ssimon stateless session resumption: this makes it possible to determine if 1936205128Ssimon a resumption has occurred immediately after receiving server hello 1937205128Ssimon (several places in OpenSSL subtly assume this) instead of later in 1938205128Ssimon the handshake. 1939205128Ssimon [Steve Henson] 1940205128Ssimon 1941205128Ssimon *) The functions ENGINE_ctrl(), OPENSSL_isservice(), 1942205128Ssimon CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error 1943205128Ssimon fixes for a few places where the return code is not checked 1944205128Ssimon correctly. 1945205128Ssimon [Julia Lawall <julia@diku.dk>] 1946205128Ssimon 1947205128Ssimon *) Add --strict-warnings option to Configure script to include devteam 1948205128Ssimon warnings in other configurations. 1949205128Ssimon [Steve Henson] 1950205128Ssimon 1951205128Ssimon *) Add support for --libdir option and LIBDIR variable in makefiles. This 1952205128Ssimon makes it possible to install openssl libraries in locations which 1953205128Ssimon have names other than "lib", for example "/usr/lib64" which some 1954205128Ssimon systems need. 1955205128Ssimon [Steve Henson, based on patch from Jeremy Utley] 1956205128Ssimon 1957205128Ssimon *) Don't allow the use of leading 0x80 in OIDs. This is a violation of 1958205128Ssimon X690 8.9.12 and can produce some misleading textual output of OIDs. 1959205128Ssimon [Steve Henson, reported by Dan Kaminsky] 1960205128Ssimon 1961205128Ssimon *) Delete MD2 from algorithm tables. This follows the recommendation in 1962205128Ssimon several standards that it is not used in new applications due to 1963205128Ssimon several cryptographic weaknesses. For binary compatibility reasons 1964205128Ssimon the MD2 API is still compiled in by default. 1965205128Ssimon [Steve Henson] 1966205128Ssimon 1967205128Ssimon *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved 1968205128Ssimon and restored. 1969205128Ssimon [Steve Henson] 1970205128Ssimon 1971205128Ssimon *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and 1972205128Ssimon OPENSSL_asc2uni conditionally on Netware platforms to avoid a name 1973205128Ssimon clash. 1974205128Ssimon [Guenter <lists@gknw.net>] 1975205128Ssimon 1976205128Ssimon *) Fix the server certificate chain building code to use X509_verify_cert(), 1977205128Ssimon it used to have an ad-hoc builder which was unable to cope with anything 1978205128Ssimon other than a simple chain. 1979205128Ssimon [David Woodhouse <dwmw2@infradead.org>, Steve Henson] 1980205128Ssimon 1981205128Ssimon *) Don't check self signed certificate signatures in X509_verify_cert() 1982205128Ssimon by default (a flag can override this): it just wastes time without 1983205128Ssimon adding any security. As a useful side effect self signed root CAs 1984205128Ssimon with non-FIPS digests are now usable in FIPS mode. 1985205128Ssimon [Steve Henson] 1986205128Ssimon 1987205128Ssimon *) In dtls1_process_out_of_seq_message() the check if the current message 1988205128Ssimon is already buffered was missing. For every new message was memory 1989205128Ssimon allocated, allowing an attacker to perform an denial of service attack 1990205128Ssimon with sending out of seq handshake messages until there is no memory 1991205128Ssimon left. Additionally every future messege was buffered, even if the 1992205128Ssimon sequence number made no sense and would be part of another handshake. 1993205128Ssimon So only messages with sequence numbers less than 10 in advance will be 1994205128Ssimon buffered. (CVE-2009-1378) 1995205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 1996205128Ssimon 1997205128Ssimon *) Records are buffered if they arrive with a future epoch to be 1998205128Ssimon processed after finishing the corresponding handshake. There is 1999205128Ssimon currently no limitation to this buffer allowing an attacker to perform 2000205128Ssimon a DOS attack with sending records with future epochs until there is no 2001205128Ssimon memory left. This patch adds the pqueue_size() function to detemine 2002205128Ssimon the size of a buffer and limits the record buffer to 100 entries. 2003205128Ssimon (CVE-2009-1377) 2004205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 2005205128Ssimon 2006205128Ssimon *) Keep a copy of frag->msg_header.frag_len so it can be used after the 2007205128Ssimon parent structure is freed. (CVE-2009-1379) 2008205128Ssimon [Daniel Mentz] 2009205128Ssimon 2010205128Ssimon *) Handle non-blocking I/O properly in SSL_shutdown() call. 2011205128Ssimon [Darryl Miles <darryl-mailinglists@netbauds.net>] 2012205128Ssimon 2013205128Ssimon *) Add 2.5.4.* OIDs 2014205128Ssimon [Ilya O. <vrghost@gmail.com>] 2015205128Ssimon 2016205128Ssimon Changes between 0.9.8k and 0.9.8l [5 Nov 2009] 2017205128Ssimon 2018205128Ssimon *) Disable renegotiation completely - this fixes a severe security 2019205128Ssimon problem (CVE-2009-3555) at the cost of breaking all 2020205128Ssimon renegotiation. Renegotiation can be re-enabled by setting 2021205128Ssimon SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at 2022205128Ssimon run-time. This is really not recommended unless you know what 2023205128Ssimon you're doing. 2024205128Ssimon [Ben Laurie] 2025205128Ssimon 2026194206Ssimon Changes between 0.9.8j and 0.9.8k [25 Mar 2009] 2027167612Ssimon 2028194206Ssimon *) Don't set val to NULL when freeing up structures, it is freed up by 2029194206Ssimon underlying code. If sizeof(void *) > sizeof(long) this can result in 2030194206Ssimon zeroing past the valid field. (CVE-2009-0789) 2031194206Ssimon [Paolo Ganci <Paolo.Ganci@AdNovum.CH>] 2032194206Ssimon 2033194206Ssimon *) Fix bug where return value of CMS_SignerInfo_verify_content() was not 2034194206Ssimon checked correctly. This would allow some invalid signed attributes to 2035194206Ssimon appear to verify correctly. (CVE-2009-0591) 2036194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 2037194206Ssimon 2038194206Ssimon *) Reject UniversalString and BMPString types with invalid lengths. This 2039194206Ssimon prevents a crash in ASN1_STRING_print_ex() which assumes the strings have 2040194206Ssimon a legal length. (CVE-2009-0590) 2041194206Ssimon [Steve Henson] 2042194206Ssimon 2043194206Ssimon *) Set S/MIME signing as the default purpose rather than setting it 2044194206Ssimon unconditionally. This allows applications to override it at the store 2045194206Ssimon level. 2046194206Ssimon [Steve Henson] 2047194206Ssimon 2048194206Ssimon *) Permit restricted recursion of ASN1 strings. This is needed in practice 2049194206Ssimon to handle some structures. 2050194206Ssimon [Steve Henson] 2051194206Ssimon 2052194206Ssimon *) Improve efficiency of mem_gets: don't search whole buffer each time 2053194206Ssimon for a '\n' 2054194206Ssimon [Jeremy Shapiro <jnshapir@us.ibm.com>] 2055194206Ssimon 2056194206Ssimon *) New -hex option for openssl rand. 2057194206Ssimon [Matthieu Herrb] 2058194206Ssimon 2059194206Ssimon *) Print out UTF8String and NumericString when parsing ASN1. 2060194206Ssimon [Steve Henson] 2061194206Ssimon 2062194206Ssimon *) Support NumericString type for name components. 2063194206Ssimon [Steve Henson] 2064194206Ssimon 2065194206Ssimon *) Allow CC in the environment to override the automatically chosen 2066194206Ssimon compiler. Note that nothing is done to ensure flags work with the 2067194206Ssimon chosen compiler. 2068194206Ssimon [Ben Laurie] 2069194206Ssimon 2070194206Ssimon Changes between 0.9.8i and 0.9.8j [07 Jan 2009] 2071194206Ssimon 2072194206Ssimon *) Properly check EVP_VerifyFinal() and similar return values 2073194206Ssimon (CVE-2008-5077). 2074194206Ssimon [Ben Laurie, Bodo Moeller, Google Security Team] 2075194206Ssimon 2076194206Ssimon *) Enable TLS extensions by default. 2077194206Ssimon [Ben Laurie] 2078194206Ssimon 2079194206Ssimon *) Allow the CHIL engine to be loaded, whether the application is 2080194206Ssimon multithreaded or not. (This does not release the developer from the 2081194206Ssimon obligation to set up the dynamic locking callbacks.) 2082194206Ssimon [Sander Temme <sander@temme.net>] 2083194206Ssimon 2084194206Ssimon *) Use correct exit code if there is an error in dgst command. 2085194206Ssimon [Steve Henson; problem pointed out by Roland Dirlewanger] 2086194206Ssimon 2087194206Ssimon *) Tweak Configure so that you need to say "experimental-jpake" to enable 2088194206Ssimon JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. 2089194206Ssimon [Bodo Moeller] 2090194206Ssimon 2091194206Ssimon *) Add experimental JPAKE support, including demo authentication in 2092194206Ssimon s_client and s_server. 2093194206Ssimon [Ben Laurie] 2094194206Ssimon 2095194206Ssimon *) Set the comparison function in v3_addr_canonize(). 2096194206Ssimon [Rob Austein <sra@hactrn.net>] 2097194206Ssimon 2098194206Ssimon *) Add support for XMPP STARTTLS in s_client. 2099194206Ssimon [Philip Paeps <philip@freebsd.org>] 2100194206Ssimon 2101194206Ssimon *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior 2102194206Ssimon to ensure that even with this option, only ciphersuites in the 2103194206Ssimon server's preference list will be accepted. (Note that the option 2104194206Ssimon applies only when resuming a session, so the earlier behavior was 2105194206Ssimon just about the algorithm choice for symmetric cryptography.) 2106194206Ssimon [Bodo Moeller] 2107194206Ssimon 2108194206Ssimon Changes between 0.9.8h and 0.9.8i [15 Sep 2008] 2109194206Ssimon 2110205128Ssimon *) Fix NULL pointer dereference if a DTLS server received 2111205128Ssimon ChangeCipherSpec as first record (CVE-2009-1386). 2112205128Ssimon [PR #1679] 2113205128Ssimon 2114194206Ssimon *) Fix a state transitition in s3_srvr.c and d1_srvr.c 2115194206Ssimon (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...). 2116194206Ssimon [Nagendra Modadugu] 2117194206Ssimon 2118194206Ssimon *) The fix in 0.9.8c that supposedly got rid of unsafe 2119194206Ssimon double-checked locking was incomplete for RSA blinding, 2120194206Ssimon addressing just one layer of what turns out to have been 2121194206Ssimon doubly unsafe triple-checked locking. 2122194206Ssimon 2123194206Ssimon So now fix this for real by retiring the MONT_HELPER macro 2124194206Ssimon in crypto/rsa/rsa_eay.c. 2125194206Ssimon 2126194206Ssimon [Bodo Moeller; problem pointed out by Marius Schilder] 2127194206Ssimon 2128194206Ssimon *) Various precautionary measures: 2129194206Ssimon 2130194206Ssimon - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). 2131194206Ssimon 2132194206Ssimon - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). 2133194206Ssimon (NB: This would require knowledge of the secret session ticket key 2134194206Ssimon to exploit, in which case you'd be SOL either way.) 2135194206Ssimon 2136194206Ssimon - Change bn_nist.c so that it will properly handle input BIGNUMs 2137194206Ssimon outside the expected range. 2138194206Ssimon 2139194206Ssimon - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG 2140194206Ssimon builds. 2141194206Ssimon 2142194206Ssimon [Neel Mehta, Bodo Moeller] 2143194206Ssimon 2144194206Ssimon *) Allow engines to be "soft loaded" - i.e. optionally don't die if 2145194206Ssimon the load fails. Useful for distros. 2146194206Ssimon [Ben Laurie and the FreeBSD team] 2147194206Ssimon 2148194206Ssimon *) Add support for Local Machine Keyset attribute in PKCS#12 files. 2149194206Ssimon [Steve Henson] 2150194206Ssimon 2151194206Ssimon *) Fix BN_GF2m_mod_arr() top-bit cleanup code. 2152194206Ssimon [Huang Ying] 2153194206Ssimon 2154194206Ssimon *) Expand ENGINE to support engine supplied SSL client certificate functions. 2155194206Ssimon 2156194206Ssimon This work was sponsored by Logica. 2157194206Ssimon [Steve Henson] 2158194206Ssimon 2159194206Ssimon *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows 2160194206Ssimon keystores. Support for SSL/TLS client authentication too. 2161194206Ssimon Not compiled unless enable-capieng specified to Configure. 2162194206Ssimon 2163194206Ssimon This work was sponsored by Logica. 2164194206Ssimon [Steve Henson] 2165194206Ssimon 2166194206Ssimon *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using 2167194206Ssimon ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain 2168194206Ssimon attribute creation routines such as certifcate requests and PKCS#12 2169194206Ssimon files. 2170194206Ssimon [Steve Henson] 2171194206Ssimon 2172194206Ssimon Changes between 0.9.8g and 0.9.8h [28 May 2008] 2173194206Ssimon 2174194206Ssimon *) Fix flaw if 'Server Key exchange message' is omitted from a TLS 2175194206Ssimon handshake which could lead to a cilent crash as found using the 2176194206Ssimon Codenomicon TLS test suite (CVE-2008-1672) 2177194206Ssimon [Steve Henson, Mark Cox] 2178194206Ssimon 2179194206Ssimon *) Fix double free in TLS server name extensions which could lead to 2180194206Ssimon a remote crash found by Codenomicon TLS test suite (CVE-2008-0891) 2181194206Ssimon [Joe Orton] 2182194206Ssimon 2183194206Ssimon *) Clear error queue in SSL_CTX_use_certificate_chain_file() 2184194206Ssimon 2185194206Ssimon Clear the error queue to ensure that error entries left from 2186194206Ssimon older function calls do not interfere with the correct operation. 2187194206Ssimon [Lutz Jaenicke, Erik de Castro Lopo] 2188194206Ssimon 2189194206Ssimon *) Remove root CA certificates of commercial CAs: 2190194206Ssimon 2191194206Ssimon The OpenSSL project does not recommend any specific CA and does not 2192194206Ssimon have any policy with respect to including or excluding any CA. 2193194206Ssimon Therefore it does not make any sense to ship an arbitrary selection 2194194206Ssimon of root CA certificates with the OpenSSL software. 2195194206Ssimon [Lutz Jaenicke] 2196194206Ssimon 2197194206Ssimon *) RSA OAEP patches to fix two separate invalid memory reads. 2198194206Ssimon The first one involves inputs when 'lzero' is greater than 2199194206Ssimon 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes 2200194206Ssimon before the beginning of from). The second one involves inputs where 2201194206Ssimon the 'db' section contains nothing but zeroes (there is a one-byte 2202194206Ssimon invalid read after the end of 'db'). 2203194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 2204194206Ssimon 2205194206Ssimon *) Partial backport from 0.9.9-dev: 2206194206Ssimon 2207194206Ssimon Introduce bn_mul_mont (dedicated Montgomery multiplication 2208194206Ssimon procedure) as a candidate for BIGNUM assembler implementation. 2209194206Ssimon While 0.9.9-dev uses assembler for various architectures, only 2210194206Ssimon x86_64 is available by default here in the 0.9.8 branch, and 2211194206Ssimon 32-bit x86 is available through a compile-time setting. 2212194206Ssimon 2213194206Ssimon To try the 32-bit x86 assembler implementation, use Configure 2214194206Ssimon option "enable-montasm" (which exists only for this backport). 2215194206Ssimon 2216194206Ssimon As "enable-montasm" for 32-bit x86 disclaims code stability 2217194206Ssimon anyway, in this constellation we activate additional code 2218194206Ssimon backported from 0.9.9-dev for further performance improvements, 2219194206Ssimon namely BN_from_montgomery_word. (To enable this otherwise, 2220194206Ssimon e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".) 2221194206Ssimon 2222194206Ssimon [Andy Polyakov (backport partially by Bodo Moeller)] 2223194206Ssimon 2224194206Ssimon *) Add TLS session ticket callback. This allows an application to set 2225194206Ssimon TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed 2226194206Ssimon values. This is useful for key rollover for example where several key 2227194206Ssimon sets may exist with different names. 2228194206Ssimon [Steve Henson] 2229194206Ssimon 2230194206Ssimon *) Reverse ENGINE-internal logic for caching default ENGINE handles. 2231194206Ssimon This was broken until now in 0.9.8 releases, such that the only way 2232194206Ssimon a registered ENGINE could be used (assuming it initialises 2233194206Ssimon successfully on the host) was to explicitly set it as the default 2234194206Ssimon for the relevant algorithms. This is in contradiction with 0.9.7 2235194206Ssimon behaviour and the documentation. With this fix, when an ENGINE is 2236194206Ssimon registered into a given algorithm's table of implementations, the 2237194206Ssimon 'uptodate' flag is reset so that auto-discovery will be used next 2238194206Ssimon time a new context for that algorithm attempts to select an 2239194206Ssimon implementation. 2240194206Ssimon [Ian Lister (tweaked by Geoff Thorpe)] 2241194206Ssimon 2242194206Ssimon *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 2243194206Ssimon implemention in the following ways: 2244194206Ssimon 2245194206Ssimon Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be 2246194206Ssimon hard coded. 2247194206Ssimon 2248194206Ssimon Lack of BER streaming support means one pass streaming processing is 2249194206Ssimon only supported if data is detached: setting the streaming flag is 2250194206Ssimon ignored for embedded content. 2251194206Ssimon 2252194206Ssimon CMS support is disabled by default and must be explicitly enabled 2253194206Ssimon with the enable-cms configuration option. 2254194206Ssimon [Steve Henson] 2255194206Ssimon 2256194206Ssimon *) Update the GMP engine glue to do direct copies between BIGNUM and 2257194206Ssimon mpz_t when openssl and GMP use the same limb size. Otherwise the 2258194206Ssimon existing "conversion via a text string export" trick is still used. 2259194206Ssimon [Paul Sheer <paulsheer@gmail.com>] 2260194206Ssimon 2261194206Ssimon *) Zlib compression BIO. This is a filter BIO which compressed and 2262194206Ssimon uncompresses any data passed through it. 2263194206Ssimon [Steve Henson] 2264194206Ssimon 2265194206Ssimon *) Add AES_wrap_key() and AES_unwrap_key() functions to implement 2266194206Ssimon RFC3394 compatible AES key wrapping. 2267194206Ssimon [Steve Henson] 2268194206Ssimon 2269194206Ssimon *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): 2270194206Ssimon sets string data without copying. X509_ALGOR_set0() and 2271194206Ssimon X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) 2272194206Ssimon data. Attribute function X509at_get0_data_by_OBJ(): retrieves data 2273194206Ssimon from an X509_ATTRIBUTE structure optionally checking it occurs only 2274194206Ssimon once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied 2275194206Ssimon data. 2276194206Ssimon [Steve Henson] 2277194206Ssimon 2278194206Ssimon *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() 2279194206Ssimon to get the expected BN_FLG_CONSTTIME behavior. 2280194206Ssimon [Bodo Moeller (Google)] 2281194206Ssimon 2282194206Ssimon *) Netware support: 2283194206Ssimon 2284194206Ssimon - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets 2285194206Ssimon - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) 2286194206Ssimon - added some more tests to do_tests.pl 2287194206Ssimon - fixed RunningProcess usage so that it works with newer LIBC NDKs too 2288194206Ssimon - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency 2289194206Ssimon - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, 2290194206Ssimon netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc 2291194206Ssimon - various changes to netware.pl to enable gcc-cross builds on Win32 2292194206Ssimon platform 2293194206Ssimon - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) 2294194206Ssimon - various changes to fix missing prototype warnings 2295194206Ssimon - fixed x86nasm.pl to create correct asm files for NASM COFF output 2296194206Ssimon - added AES, WHIRLPOOL and CPUID assembler code to build files 2297194206Ssimon - added missing AES assembler make rules to mk1mf.pl 2298194206Ssimon - fixed order of includes in apps/ocsp.c so that e_os.h settings apply 2299194206Ssimon [Guenter Knauf <eflash@gmx.net>] 2300194206Ssimon 2301194206Ssimon *) Implement certificate status request TLS extension defined in RFC3546. 2302194206Ssimon A client can set the appropriate parameters and receive the encoded 2303194206Ssimon OCSP response via a callback. A server can query the supplied parameters 2304194206Ssimon and set the encoded OCSP response in the callback. Add simplified examples 2305194206Ssimon to s_client and s_server. 2306194206Ssimon [Steve Henson] 2307194206Ssimon 2308194206Ssimon Changes between 0.9.8f and 0.9.8g [19 Oct 2007] 2309194206Ssimon 2310194206Ssimon *) Fix various bugs: 2311194206Ssimon + Binary incompatibility of ssl_ctx_st structure 2312194206Ssimon + DTLS interoperation with non-compliant servers 2313194206Ssimon + Don't call get_session_cb() without proposed session 2314194206Ssimon + Fix ia64 assembler code 2315194206Ssimon [Andy Polyakov, Steve Henson] 2316194206Ssimon 2317194206Ssimon Changes between 0.9.8e and 0.9.8f [11 Oct 2007] 2318194206Ssimon 2319194206Ssimon *) DTLS Handshake overhaul. There were longstanding issues with 2320194206Ssimon OpenSSL DTLS implementation, which were making it impossible for 2321194206Ssimon RFC 4347 compliant client to communicate with OpenSSL server. 2322194206Ssimon Unfortunately just fixing these incompatibilities would "cut off" 2323194206Ssimon pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e 2324194206Ssimon server keeps tolerating non RFC compliant syntax. The opposite is 2325194206Ssimon not true, 0.9.8f client can not communicate with earlier server. 2326194206Ssimon This update even addresses CVE-2007-4995. 2327194206Ssimon [Andy Polyakov] 2328194206Ssimon 2329194206Ssimon *) Changes to avoid need for function casts in OpenSSL: some compilers 2330194206Ssimon (gcc 4.2 and later) reject their use. 2331194206Ssimon [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, 2332194206Ssimon Steve Henson] 2333194206Ssimon 2334194206Ssimon *) Add RFC4507 support to OpenSSL. This includes the corrections in 2335194206Ssimon RFC4507bis. The encrypted ticket format is an encrypted encoded 2336194206Ssimon SSL_SESSION structure, that way new session features are automatically 2337194206Ssimon supported. 2338194206Ssimon 2339194206Ssimon If a client application caches session in an SSL_SESSION structure 2340194206Ssimon support is transparent because tickets are now stored in the encoded 2341194206Ssimon SSL_SESSION. 2342194206Ssimon 2343194206Ssimon The SSL_CTX structure automatically generates keys for ticket 2344194206Ssimon protection in servers so again support should be possible 2345194206Ssimon with no application modification. 2346194206Ssimon 2347194206Ssimon If a client or server wishes to disable RFC4507 support then the option 2348194206Ssimon SSL_OP_NO_TICKET can be set. 2349194206Ssimon 2350194206Ssimon Add a TLS extension debugging callback to allow the contents of any client 2351194206Ssimon or server extensions to be examined. 2352194206Ssimon 2353194206Ssimon This work was sponsored by Google. 2354194206Ssimon [Steve Henson] 2355194206Ssimon 2356194206Ssimon *) Add initial support for TLS extensions, specifically for the server_name 2357194206Ssimon extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 2358194206Ssimon have new members for a host name. The SSL data structure has an 2359194206Ssimon additional member SSL_CTX *initial_ctx so that new sessions can be 2360194206Ssimon stored in that context to allow for session resumption, even after the 2361194206Ssimon SSL has been switched to a new SSL_CTX in reaction to a client's 2362194206Ssimon server_name extension. 2363194206Ssimon 2364194206Ssimon New functions (subject to change): 2365194206Ssimon 2366194206Ssimon SSL_get_servername() 2367194206Ssimon SSL_get_servername_type() 2368194206Ssimon SSL_set_SSL_CTX() 2369194206Ssimon 2370194206Ssimon New CTRL codes and macros (subject to change): 2371194206Ssimon 2372194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 2373194206Ssimon - SSL_CTX_set_tlsext_servername_callback() 2374194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 2375194206Ssimon - SSL_CTX_set_tlsext_servername_arg() 2376194206Ssimon SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 2377194206Ssimon 2378194206Ssimon openssl s_client has a new '-servername ...' option. 2379194206Ssimon 2380194206Ssimon openssl s_server has new options '-servername_host ...', '-cert2 ...', 2381194206Ssimon '-key2 ...', '-servername_fatal' (subject to change). This allows 2382194206Ssimon testing the HostName extension for a specific single host name ('-cert' 2383194206Ssimon and '-key' remain fallbacks for handshakes without HostName 2384194206Ssimon negotiation). If the unrecogninzed_name alert has to be sent, this by 2385194206Ssimon default is a warning; it becomes fatal with the '-servername_fatal' 2386194206Ssimon option. 2387194206Ssimon 2388194206Ssimon [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson] 2389194206Ssimon 2390194206Ssimon *) Add AES and SSE2 assembly language support to VC++ build. 2391194206Ssimon [Steve Henson] 2392194206Ssimon 2393194206Ssimon *) Mitigate attack on final subtraction in Montgomery reduction. 2394194206Ssimon [Andy Polyakov] 2395194206Ssimon 2396194206Ssimon *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 2397194206Ssimon (which previously caused an internal error). 2398194206Ssimon [Bodo Moeller] 2399194206Ssimon 2400194206Ssimon *) Squeeze another 10% out of IGE mode when in != out. 2401194206Ssimon [Ben Laurie] 2402194206Ssimon 2403194206Ssimon *) AES IGE mode speedup. 2404194206Ssimon [Dean Gaudet (Google)] 2405194206Ssimon 2406194206Ssimon *) Add the Korean symmetric 128-bit cipher SEED (see 2407194206Ssimon http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and 2408194206Ssimon add SEED ciphersuites from RFC 4162: 2409194206Ssimon 2410194206Ssimon TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" 2411194206Ssimon TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" 2412194206Ssimon TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" 2413194206Ssimon TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" 2414194206Ssimon 2415194206Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 2416194206Ssimon series, SEED remains excluded from compilation unless OpenSSL 2417194206Ssimon is configured with 'enable-seed'. 2418194206Ssimon [KISA, Bodo Moeller] 2419194206Ssimon 2420194206Ssimon *) Mitigate branch prediction attacks, which can be practical if a 2421194206Ssimon single processor is shared, allowing a spy process to extract 2422194206Ssimon information. For detailed background information, see 2423194206Ssimon http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron, 2424194206Ssimon J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL 2425194206Ssimon and Necessary Software Countermeasures"). The core of the change 2426194206Ssimon are new versions BN_div_no_branch() and 2427194206Ssimon BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), 2428194206Ssimon respectively, which are slower, but avoid the security-relevant 2429194206Ssimon conditional branches. These are automatically called by BN_div() 2430194206Ssimon and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one 2431194206Ssimon of the input BIGNUMs. Also, BN_is_bit_set() has been changed to 2432194206Ssimon remove a conditional branch. 2433194206Ssimon 2434194206Ssimon BN_FLG_CONSTTIME is the new name for the previous 2435194206Ssimon BN_FLG_EXP_CONSTTIME flag, since it now affects more than just 2436194206Ssimon modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag 2437194206Ssimon in the exponent causes BN_mod_exp_mont() to use the alternative 2438194206Ssimon implementation in BN_mod_exp_mont_consttime().) The old name 2439194206Ssimon remains as a deprecated alias. 2440194206Ssimon 2441194206Ssimon Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general 2442194206Ssimon RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses 2443194206Ssimon constant-time implementations for more than just exponentiation. 2444194206Ssimon Here too the old name is kept as a deprecated alias. 2445194206Ssimon 2446194206Ssimon BN_BLINDING_new() will now use BN_dup() for the modulus so that 2447194206Ssimon the BN_BLINDING structure gets an independent copy of the 2448194206Ssimon modulus. This means that the previous "BIGNUM *m" argument to 2449194206Ssimon BN_BLINDING_new() and to BN_BLINDING_create_param() now 2450194206Ssimon essentially becomes "const BIGNUM *m", although we can't actually 2451194206Ssimon change this in the header file before 0.9.9. It allows 2452194206Ssimon RSA_setup_blinding() to use BN_with_flags() on the modulus to 2453194206Ssimon enable BN_FLG_CONSTTIME. 2454194206Ssimon 2455194206Ssimon [Matthew D Wood (Intel Corp)] 2456194206Ssimon 2457194206Ssimon *) In the SSL/TLS server implementation, be strict about session ID 2458194206Ssimon context matching (which matters if an application uses a single 2459194206Ssimon external cache for different purposes). Previously, 2460194206Ssimon out-of-context reuse was forbidden only if SSL_VERIFY_PEER was 2461194206Ssimon set. This did ensure strict client verification, but meant that, 2462194206Ssimon with applications using a single external cache for quite 2463194206Ssimon different requirements, clients could circumvent ciphersuite 2464194206Ssimon restrictions for a given session ID context by starting a session 2465194206Ssimon in a different context. 2466194206Ssimon [Bodo Moeller] 2467194206Ssimon 2468167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 2469167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 2470167612Ssimon authentication-only ciphersuites. 2471167612Ssimon [Bodo Moeller] 2472167612Ssimon 2473194206Ssimon *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was 2474194206Ssimon not complete and could lead to a possible single byte overflow 2475194206Ssimon (CVE-2007-5135) [Ben Laurie] 2476194206Ssimon 2477194206Ssimon Changes between 0.9.8d and 0.9.8e [23 Feb 2007] 2478194206Ssimon 2479167612Ssimon *) Since AES128 and AES256 (and similarly Camellia128 and 2480167612Ssimon Camellia256) share a single mask bit in the logic of 2481167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 2482167612Ssimon kludge to work properly if AES128 is available and AES256 isn't 2483167612Ssimon (or if Camellia128 is available and Camellia256 isn't). 2484167612Ssimon [Victor Duchovni] 2485167612Ssimon 2486167612Ssimon *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c 2487167612Ssimon (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): 2488167612Ssimon When a point or a seed is encoded in a BIT STRING, we need to 2489167612Ssimon prevent the removal of trailing zero bits to get the proper DER 2490167612Ssimon encoding. (By default, crypto/asn1/a_bitstr.c assumes the case 2491167612Ssimon of a NamedBitList, for which trailing 0 bits need to be removed.) 2492167612Ssimon [Bodo Moeller] 2493167612Ssimon 2494167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 2495167612Ssimon protocol version while receiving ClientHello even if the 2496167612Ssimon ClientHello is fragmented. (The server can't insist on the 2497167612Ssimon particular protocol version it has chosen before the ServerHello 2498167612Ssimon message has informed the client about his choice.) 2499167612Ssimon [Bodo Moeller] 2500167612Ssimon 2501167612Ssimon *) Add RFC 3779 support. 2502167612Ssimon [Rob Austein for ARIN, Ben Laurie] 2503167612Ssimon 2504167612Ssimon *) Load error codes if they are not already present instead of using a 2505167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 2506167612Ssimon Improve header file function name parsing. 2507167612Ssimon [Steve Henson] 2508167612Ssimon 2509167612Ssimon *) extend SMTP and IMAP protocol emulation in s_client to use EHLO 2510167612Ssimon or CAPABILITY handshake as required by RFCs. 2511167612Ssimon [Goetz Babin-Ebell] 2512167612Ssimon 2513162911Ssimon Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 2514162911Ssimon 2515162911Ssimon *) Introduce limits to prevent malicious keys being able to 2516162911Ssimon cause a denial of service. (CVE-2006-2940) 2517162911Ssimon [Steve Henson, Bodo Moeller] 2518162911Ssimon 2519162911Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 2520162911Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 2521162911Ssimon 2522162911Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 2523162911Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 2524162911Ssimon 2525162911Ssimon *) Fix SSL client code which could crash if connecting to a 2526162911Ssimon malicious SSLv2 server. (CVE-2006-4343) 2527162911Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 2528162911Ssimon 2529162911Ssimon *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 2530162911Ssimon match only those. Before that, "AES256-SHA" would be interpreted 2531162911Ssimon as a pattern and match "AES128-SHA" too (since AES128-SHA got 2532162911Ssimon the same strength classification in 0.9.7h) as we currently only 2533162911Ssimon have a single AES bit in the ciphersuite description bitmap. 2534162911Ssimon That change, however, also applied to ciphersuite strings such as 2535162911Ssimon "RC4-MD5" that intentionally matched multiple ciphersuites -- 2536162911Ssimon namely, SSL 2.0 ciphersuites in addition to the more common ones 2537162911Ssimon from SSL 3.0/TLS 1.0. 2538162911Ssimon 2539162911Ssimon So we change the selection algorithm again: Naming an explicit 2540162911Ssimon ciphersuite selects this one ciphersuite, and any other similar 2541162911Ssimon ciphersuite (same bitmap) from *other* protocol versions. 2542162911Ssimon Thus, "RC4-MD5" again will properly select both the SSL 2.0 2543162911Ssimon ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 2544162911Ssimon 2545162911Ssimon Since SSL 2.0 does not have any ciphersuites for which the 2546162911Ssimon 128/256 bit distinction would be relevant, this works for now. 2547162911Ssimon The proper fix will be to use different bits for AES128 and 2548162911Ssimon AES256, which would have avoided the problems from the beginning; 2549162911Ssimon however, bits are scarce, so we can only do this in a new release 2550162911Ssimon (not just a patchlevel) when we can change the SSL_CIPHER 2551162911Ssimon definition to split the single 'unsigned long mask' bitmap into 2552162911Ssimon multiple values to extend the available space. 2553162911Ssimon 2554162911Ssimon [Bodo Moeller] 2555162911Ssimon 2556162911Ssimon Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 2557162911Ssimon 2558162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 2559162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 2560162911Ssimon 2561162911Ssimon *) Add AES IGE and biIGE modes. 2562162911Ssimon [Ben Laurie] 2563162911Ssimon 2564162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 2565162911Ssimon possible instead of select(), since the latter has some 2566162911Ssimon undesirable limitations. 2567162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 2568162911Ssimon 2569162911Ssimon *) Disable "ECCdraft" ciphersuites more thoroughly. Now special 2570162911Ssimon treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 2571162911Ssimon cannot be implicitly activated as part of, e.g., the "AES" alias. 2572162911Ssimon However, please upgrade to OpenSSL 0.9.9[-dev] for 2573162911Ssimon non-experimental use of the ECC ciphersuites to get TLS extension 2574162911Ssimon support, which is required for curve and point format negotiation 2575162911Ssimon to avoid potential handshake problems. 2576162911Ssimon [Bodo Moeller] 2577162911Ssimon 2578162911Ssimon *) Disable rogue ciphersuites: 2579162911Ssimon 2580162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 2581162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 2582162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 2583162911Ssimon 2584162911Ssimon The latter two were purportedly from 2585162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 2586162911Ssimon appear there. 2587162911Ssimon 2588167612Ssimon Also deactivate the remaining ciphersuites from 2589162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 2590162911Ssimon unofficial, and the ID has long expired. 2591162911Ssimon [Bodo Moeller] 2592162911Ssimon 2593162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 2594162911Ssimon dual-core machines) and other potential thread-safety issues. 2595162911Ssimon [Bodo Moeller] 2596162911Ssimon 2597162911Ssimon *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 2598162911Ssimon versions), which is now available for royalty-free use 2599162911Ssimon (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). 2600162911Ssimon Also, add Camellia TLS ciphersuites from RFC 4132. 2601162911Ssimon 2602162911Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 2603162911Ssimon series, Camellia remains excluded from compilation unless OpenSSL 2604162911Ssimon is configured with 'enable-camellia'. 2605162911Ssimon [NTT] 2606162911Ssimon 2607162911Ssimon *) Disable the padding bug check when compression is in use. The padding 2608162911Ssimon bug check assumes the first packet is of even length, this is not 2609162911Ssimon necessarily true if compresssion is enabled and can result in false 2610162911Ssimon positives causing handshake failure. The actual bug test is ancient 2611162911Ssimon code so it is hoped that implementations will either have fixed it by 2612162911Ssimon now or any which still have the bug do not support compression. 2613162911Ssimon [Steve Henson] 2614162911Ssimon 2615160814Ssimon Changes between 0.9.8a and 0.9.8b [04 May 2006] 2616160814Ssimon 2617160814Ssimon *) When applying a cipher rule check to see if string match is an explicit 2618160814Ssimon cipher suite and only match that one cipher suite if it is. 2619160814Ssimon [Steve Henson] 2620160814Ssimon 2621160814Ssimon *) Link in manifests for VC++ if needed. 2622160814Ssimon [Austin Ziegler <halostatue@gmail.com>] 2623160814Ssimon 2624160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 2625160814Ssimon draft-ietf-tls-ecc-12.txt with proposed changes (but without 2626160814Ssimon TLS extensions, which are supported starting with the 0.9.9 2627160814Ssimon branch, not in the OpenSSL 0.9.8 branch). 2628160814Ssimon [Douglas Stebila] 2629160814Ssimon 2630160814Ssimon *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 2631160814Ssimon opaque EVP_CIPHER_CTX handling. 2632160814Ssimon [Steve Henson] 2633160814Ssimon 2634160814Ssimon *) Fixes and enhancements to zlib compression code. We now only use 2635160814Ssimon "zlib1.dll" and use the default __cdecl calling convention on Win32 2636160814Ssimon to conform with the standards mentioned here: 2637160814Ssimon http://www.zlib.net/DLL_FAQ.txt 2638160814Ssimon Static zlib linking now works on Windows and the new --with-zlib-include 2639160814Ssimon --with-zlib-lib options to Configure can be used to supply the location 2640160814Ssimon of the headers and library. Gracefully handle case where zlib library 2641160814Ssimon can't be loaded. 2642160814Ssimon [Steve Henson] 2643160814Ssimon 2644160814Ssimon *) Several fixes and enhancements to the OID generation code. The old code 2645160814Ssimon sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 2646160814Ssimon handle numbers larger than ULONG_MAX, truncated printing and had a 2647160814Ssimon non standard OBJ_obj2txt() behaviour. 2648160814Ssimon [Steve Henson] 2649160814Ssimon 2650160814Ssimon *) Add support for building of engines under engine/ as shared libraries 2651160814Ssimon under VC++ build system. 2652160814Ssimon [Steve Henson] 2653160814Ssimon 2654160814Ssimon *) Corrected the numerous bugs in the Win32 path splitter in DSO. 2655160814Ssimon Hopefully, we will not see any false combination of paths any more. 2656160814Ssimon [Richard Levitte] 2657160814Ssimon 2658160814Ssimon Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 2659160814Ssimon 2660160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 2661160814Ssimon (part of SSL_OP_ALL). This option used to disable the 2662160814Ssimon countermeasure against man-in-the-middle protocol-version 2663160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 2664160814Ssimon idea. (CVE-2005-2969) 2665160814Ssimon 2666160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 2667160814Ssimon for Information Security, National Institute of Advanced Industrial 2668160814Ssimon Science and Technology [AIST], Japan)] 2669160814Ssimon 2670160814Ssimon *) Add two function to clear and return the verify parameter flags. 2671160814Ssimon [Steve Henson] 2672160814Ssimon 2673160814Ssimon *) Keep cipherlists sorted in the source instead of sorting them at 2674160814Ssimon runtime, thus removing the need for a lock. 2675160814Ssimon [Nils Larsch] 2676160814Ssimon 2677160814Ssimon *) Avoid some small subgroup attacks in Diffie-Hellman. 2678160814Ssimon [Nick Mathewson and Ben Laurie] 2679160814Ssimon 2680160814Ssimon *) Add functions for well-known primes. 2681160814Ssimon [Nick Mathewson] 2682160814Ssimon 2683160814Ssimon *) Extended Windows CE support. 2684160814Ssimon [Satoshi Nakamura and Andy Polyakov] 2685160814Ssimon 2686160814Ssimon *) Initialize SSL_METHOD structures at compile time instead of during 2687160814Ssimon runtime, thus removing the need for a lock. 2688160814Ssimon [Steve Henson] 2689160814Ssimon 2690160814Ssimon *) Make PKCS7_decrypt() work even if no certificate is supplied by 2691160814Ssimon attempting to decrypt each encrypted key in turn. Add support to 2692160814Ssimon smime utility. 2693160814Ssimon [Steve Henson] 2694160814Ssimon 2695160814Ssimon Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 2696160814Ssimon 2697162911Ssimon [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 2698162911Ssimon OpenSSL 0.9.8.] 2699162911Ssimon 2700160814Ssimon *) Add libcrypto.pc and libssl.pc for those who feel they need them. 2701160814Ssimon [Richard Levitte] 2702160814Ssimon 2703160814Ssimon *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 2704160814Ssimon key into the same file any more. 2705160814Ssimon [Richard Levitte] 2706160814Ssimon 2707160814Ssimon *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 2708160814Ssimon [Andy Polyakov] 2709160814Ssimon 2710160814Ssimon *) Add -utf8 command line and config file option to 'ca'. 2711160814Ssimon [Stefan <stf@udoma.org] 2712160814Ssimon 2713160814Ssimon *) Removed the macro des_crypt(), as it seems to conflict with some 2714160814Ssimon libraries. Use DES_crypt(). 2715160814Ssimon [Richard Levitte] 2716160814Ssimon 2717160814Ssimon *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 2718160814Ssimon involves renaming the source and generated shared-libs for 2719160814Ssimon both. The engines will accept the corrected or legacy ids 2720160814Ssimon ('ncipher' and '4758_cca' respectively) when binding. NB, 2721160814Ssimon this only applies when building 'shared'. 2722160814Ssimon [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 2723160814Ssimon 2724160814Ssimon *) Add attribute functions to EVP_PKEY structure. Modify 2725160814Ssimon PKCS12_create() to recognize a CSP name attribute and 2726160814Ssimon use it. Make -CSP option work again in pkcs12 utility. 2727160814Ssimon [Steve Henson] 2728160814Ssimon 2729160814Ssimon *) Add new functionality to the bn blinding code: 2730160814Ssimon - automatic re-creation of the BN_BLINDING parameters after 2731160814Ssimon a fixed number of uses (currently 32) 2732160814Ssimon - add new function for parameter creation 2733160814Ssimon - introduce flags to control the update behaviour of the 2734160814Ssimon BN_BLINDING parameters 2735160814Ssimon - hide BN_BLINDING structure 2736160814Ssimon Add a second BN_BLINDING slot to the RSA structure to improve 2737160814Ssimon performance when a single RSA object is shared among several 2738160814Ssimon threads. 2739160814Ssimon [Nils Larsch] 2740160814Ssimon 2741160814Ssimon *) Add support for DTLS. 2742160814Ssimon [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 2743160814Ssimon 2744160814Ssimon *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 2745160814Ssimon to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 2746160814Ssimon [Walter Goulet] 2747160814Ssimon 2748160814Ssimon *) Remove buggy and incompletet DH cert support from 2749160814Ssimon ssl/ssl_rsa.c and ssl/s3_both.c 2750160814Ssimon [Nils Larsch] 2751160814Ssimon 2752160814Ssimon *) Use SHA-1 instead of MD5 as the default digest algorithm for 2753160814Ssimon the apps/openssl applications. 2754160814Ssimon [Nils Larsch] 2755160814Ssimon 2756160814Ssimon *) Compile clean with "-Wall -Wmissing-prototypes 2757160814Ssimon -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 2758160814Ssimon DEBUG_SAFESTACK must also be set. 2759160814Ssimon [Ben Laurie] 2760160814Ssimon 2761160814Ssimon *) Change ./Configure so that certain algorithms can be disabled by default. 2762160814Ssimon The new counterpiece to "no-xxx" is "enable-xxx". 2763160814Ssimon 2764160814Ssimon The patented RC5 and MDC2 algorithms will now be disabled unless 2765160814Ssimon "enable-rc5" and "enable-mdc2", respectively, are specified. 2766160814Ssimon 2767160814Ssimon (IDEA remains enabled despite being patented. This is because IDEA 2768160814Ssimon is frequently required for interoperability, and there is no license 2769160814Ssimon fee for non-commercial use. As before, "no-idea" can be used to 2770160814Ssimon avoid this algorithm.) 2771160814Ssimon 2772160814Ssimon [Bodo Moeller] 2773160814Ssimon 2774160814Ssimon *) Add processing of proxy certificates (see RFC 3820). This work was 2775160814Ssimon sponsored by KTH (The Royal Institute of Technology in Stockholm) and 2776160814Ssimon EGEE (Enabling Grids for E-science in Europe). 2777160814Ssimon [Richard Levitte] 2778160814Ssimon 2779160814Ssimon *) RC4 performance overhaul on modern architectures/implementations, such 2780160814Ssimon as Intel P4, IA-64 and AMD64. 2781160814Ssimon [Andy Polyakov] 2782160814Ssimon 2783160814Ssimon *) New utility extract-section.pl. This can be used specify an alternative 2784160814Ssimon section number in a pod file instead of having to treat each file as 2785160814Ssimon a separate case in Makefile. This can be done by adding two lines to the 2786160814Ssimon pod file: 2787160814Ssimon 2788160814Ssimon =for comment openssl_section:XXX 2789160814Ssimon 2790160814Ssimon The blank line is mandatory. 2791160814Ssimon 2792160814Ssimon [Steve Henson] 2793160814Ssimon 2794160814Ssimon *) New arguments -certform, -keyform and -pass for s_client and s_server 2795160814Ssimon to allow alternative format key and certificate files and passphrase 2796160814Ssimon sources. 2797160814Ssimon [Steve Henson] 2798160814Ssimon 2799160814Ssimon *) New structure X509_VERIFY_PARAM which combines current verify parameters, 2800160814Ssimon update associated structures and add various utility functions. 2801160814Ssimon 2802160814Ssimon Add new policy related verify parameters, include policy checking in 2803160814Ssimon standard verify code. Enhance 'smime' application with extra parameters 2804160814Ssimon to support policy checking and print out. 2805160814Ssimon [Steve Henson] 2806160814Ssimon 2807160814Ssimon *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 2808160814Ssimon Nehemiah processors. These extensions support AES encryption in hardware 2809160814Ssimon as well as RNG (though RNG support is currently disabled). 2810160814Ssimon [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 2811160814Ssimon 2812160814Ssimon *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 2813160814Ssimon [Geoff Thorpe] 2814160814Ssimon 2815160814Ssimon *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 2816160814Ssimon [Andy Polyakov and a number of other people] 2817160814Ssimon 2818160814Ssimon *) Improved PowerPC platform support. Most notably BIGNUM assembler 2819160814Ssimon implementation contributed by IBM. 2820160814Ssimon [Suresh Chari, Peter Waltenberg, Andy Polyakov] 2821160814Ssimon 2822160814Ssimon *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 2823160814Ssimon exponent rather than 'unsigned long'. There is a corresponding change to 2824160814Ssimon the new 'rsa_keygen' element of the RSA_METHOD structure. 2825160814Ssimon [Jelte Jansen, Geoff Thorpe] 2826160814Ssimon 2827160814Ssimon *) Functionality for creating the initial serial number file is now 2828160814Ssimon moved from CA.pl to the 'ca' utility with a new option -create_serial. 2829160814Ssimon 2830160814Ssimon (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 2831160814Ssimon number file to 1, which is bound to cause problems. To avoid 2832160814Ssimon the problems while respecting compatibility between different 0.9.7 2833160814Ssimon patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 2834160814Ssimon CA.pl for serial number initialization. With the new release 0.9.8, 2835160814Ssimon we can fix the problem directly in the 'ca' utility.) 2836160814Ssimon [Steve Henson] 2837160814Ssimon 2838160814Ssimon *) Reduced header interdepencies by declaring more opaque objects in 2839160814Ssimon ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 2840160814Ssimon give fewer recursive includes, which could break lazy source code - so 2841160814Ssimon this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 2842160814Ssimon developers should define this symbol when building and using openssl to 2843160814Ssimon ensure they track the recommended behaviour, interfaces, [etc], but 2844160814Ssimon backwards-compatible behaviour prevails when this isn't defined. 2845160814Ssimon [Geoff Thorpe] 2846160814Ssimon 2847160814Ssimon *) New function X509_POLICY_NODE_print() which prints out policy nodes. 2848160814Ssimon [Steve Henson] 2849160814Ssimon 2850160814Ssimon *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 2851160814Ssimon This will generate a random key of the appropriate length based on the 2852160814Ssimon cipher context. The EVP_CIPHER can provide its own random key generation 2853160814Ssimon routine to support keys of a specific form. This is used in the des and 2854160814Ssimon 3des routines to generate a key of the correct parity. Update S/MIME 2855160814Ssimon code to use new functions and hence generate correct parity DES keys. 2856160814Ssimon Add EVP_CHECK_DES_KEY #define to return an error if the key is not 2857160814Ssimon valid (weak or incorrect parity). 2858160814Ssimon [Steve Henson] 2859160814Ssimon 2860160814Ssimon *) Add a local set of CRLs that can be used by X509_verify_cert() as well 2861160814Ssimon as looking them up. This is useful when the verified structure may contain 2862160814Ssimon CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 2863160814Ssimon present unless the new PKCS7_NO_CRL flag is asserted. 2864160814Ssimon [Steve Henson] 2865160814Ssimon 2866160814Ssimon *) Extend ASN1 oid configuration module. It now additionally accepts the 2867160814Ssimon syntax: 2868160814Ssimon 2869160814Ssimon shortName = some long name, 1.2.3.4 2870160814Ssimon [Steve Henson] 2871160814Ssimon 2872160814Ssimon *) Reimplemented the BN_CTX implementation. There is now no more static 2873160814Ssimon limitation on the number of variables it can handle nor the depth of the 2874160814Ssimon "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 2875160814Ssimon information can now expand as required, and rather than having a single 2876160814Ssimon static array of bignums, BN_CTX now uses a linked-list of such arrays 2877160814Ssimon allowing it to expand on demand whilst maintaining the usefulness of 2878160814Ssimon BN_CTX's "bundling". 2879160814Ssimon [Geoff Thorpe] 2880160814Ssimon 2881160814Ssimon *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 2882160814Ssimon to allow all RSA operations to function using a single BN_CTX. 2883160814Ssimon [Geoff Thorpe] 2884160814Ssimon 2885160814Ssimon *) Preliminary support for certificate policy evaluation and checking. This 2886160814Ssimon is initially intended to pass the tests outlined in "Conformance Testing 2887160814Ssimon of Relying Party Client Certificate Path Processing Logic" v1.07. 2888160814Ssimon [Steve Henson] 2889160814Ssimon 2890160814Ssimon *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 2891160814Ssimon remained unused and not that useful. A variety of other little bignum 2892160814Ssimon tweaks and fixes have also been made continuing on from the audit (see 2893160814Ssimon below). 2894160814Ssimon [Geoff Thorpe] 2895160814Ssimon 2896160814Ssimon *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 2897160814Ssimon associated ASN1, EVP and SSL functions and old ASN1 macros. 2898160814Ssimon [Richard Levitte] 2899160814Ssimon 2900160814Ssimon *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 2901160814Ssimon and this should never fail. So the return value from the use of 2902160814Ssimon BN_set_word() (which can fail due to needless expansion) is now deprecated; 2903160814Ssimon if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 2904160814Ssimon [Geoff Thorpe] 2905160814Ssimon 2906160814Ssimon *) BN_CTX_get() should return zero-valued bignums, providing the same 2907160814Ssimon initialised value as BN_new(). 2908160814Ssimon [Geoff Thorpe, suggested by Ulf M�ller] 2909160814Ssimon 2910160814Ssimon *) Support for inhibitAnyPolicy certificate extension. 2911160814Ssimon [Steve Henson] 2912160814Ssimon 2913160814Ssimon *) An audit of the BIGNUM code is underway, for which debugging code is 2914160814Ssimon enabled when BN_DEBUG is defined. This makes stricter enforcements on what 2915160814Ssimon is considered valid when processing BIGNUMs, and causes execution to 2916160814Ssimon assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 2917160814Ssimon further steps are taken to deliberately pollute unused data in BIGNUM 2918160814Ssimon structures to try and expose faulty code further on. For now, openssl will 2919160814Ssimon (in its default mode of operation) continue to tolerate the inconsistent 2920160814Ssimon forms that it has tolerated in the past, but authors and packagers should 2921160814Ssimon consider trying openssl and their own applications when compiled with 2922160814Ssimon these debugging symbols defined. It will help highlight potential bugs in 2923160814Ssimon their own code, and will improve the test coverage for OpenSSL itself. At 2924160814Ssimon some point, these tighter rules will become openssl's default to improve 2925160814Ssimon maintainability, though the assert()s and other overheads will remain only 2926160814Ssimon in debugging configurations. See bn.h for more details. 2927160814Ssimon [Geoff Thorpe, Nils Larsch, Ulf M�ller] 2928160814Ssimon 2929160814Ssimon *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 2930160814Ssimon that can only be obtained through BN_CTX_new() (which implicitly 2931160814Ssimon initialises it). The presence of this function only made it possible 2932160814Ssimon to overwrite an existing structure (and cause memory leaks). 2933160814Ssimon [Geoff Thorpe] 2934160814Ssimon 2935160814Ssimon *) Because of the callback-based approach for implementing LHASH as a 2936160814Ssimon template type, lh_insert() adds opaque objects to hash-tables and 2937160814Ssimon lh_doall() or lh_doall_arg() are typically used with a destructor callback 2938160814Ssimon to clean up those corresponding objects before destroying the hash table 2939160814Ssimon (and losing the object pointers). So some over-zealous constifications in 2940160814Ssimon LHASH have been relaxed so that lh_insert() does not take (nor store) the 2941160814Ssimon objects as "const" and the lh_doall[_arg] callback wrappers are not 2942160814Ssimon prototyped to have "const" restrictions on the object pointers they are 2943160814Ssimon given (and so aren't required to cast them away any more). 2944160814Ssimon [Geoff Thorpe] 2945160814Ssimon 2946160814Ssimon *) The tmdiff.h API was so ugly and minimal that our own timing utility 2947160814Ssimon (speed) prefers to use its own implementation. The two implementations 2948160814Ssimon haven't been consolidated as yet (volunteers?) but the tmdiff API has had 2949160814Ssimon its object type properly exposed (MS_TM) instead of casting to/from "char 2950160814Ssimon *". This may still change yet if someone realises MS_TM and "ms_time_***" 2951160814Ssimon aren't necessarily the greatest nomenclatures - but this is what was used 2952160814Ssimon internally to the implementation so I've used that for now. 2953160814Ssimon [Geoff Thorpe] 2954160814Ssimon 2955160814Ssimon *) Ensure that deprecated functions do not get compiled when 2956160814Ssimon OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 2957160814Ssimon the self-tests were still using deprecated key-generation functions so 2958160814Ssimon these have been updated also. 2959160814Ssimon [Geoff Thorpe] 2960160814Ssimon 2961160814Ssimon *) Reorganise PKCS#7 code to separate the digest location functionality 2962160814Ssimon into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 2963160814Ssimon New function PKCS7_set_digest() to set the digest type for PKCS#7 2964160814Ssimon digestedData type. Add additional code to correctly generate the 2965160814Ssimon digestedData type and add support for this type in PKCS7 initialization 2966160814Ssimon functions. 2967160814Ssimon [Steve Henson] 2968160814Ssimon 2969160814Ssimon *) New function PKCS7_set0_type_other() this initializes a PKCS7 2970160814Ssimon structure of type "other". 2971160814Ssimon [Steve Henson] 2972160814Ssimon 2973160814Ssimon *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 2974160814Ssimon sure the loop does correctly stop and breaking ("division by zero") 2975160814Ssimon modulus operations are not performed. The (pre-generated) prime 2976160814Ssimon table crypto/bn/bn_prime.h was already correct, but it could not be 2977160814Ssimon re-generated on some platforms because of the "division by zero" 2978160814Ssimon situation in the script. 2979160814Ssimon [Ralf S. Engelschall] 2980160814Ssimon 2981160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 2982160814Ssimon draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 2983160814Ssimon SHA-1 now is only used for "small" curves (where the 2984160814Ssimon representation of a field element takes up to 24 bytes); for 2985160814Ssimon larger curves, the field element resulting from ECDH is directly 2986160814Ssimon used as premaster secret. 2987160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 2988160814Ssimon 2989160814Ssimon *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 2990160814Ssimon curve secp160r1 to the tests. 2991160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 2992160814Ssimon 2993160814Ssimon *) Add the possibility to load symbols globally with DSO. 2994160814Ssimon [G�tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 2995160814Ssimon 2996160814Ssimon *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 2997160814Ssimon control of the error stack. 2998160814Ssimon [Richard Levitte] 2999160814Ssimon 3000160814Ssimon *) Add support for STORE in ENGINE. 3001160814Ssimon [Richard Levitte] 3002160814Ssimon 3003160814Ssimon *) Add the STORE type. The intention is to provide a common interface 3004160814Ssimon to certificate and key stores, be they simple file-based stores, or 3005160814Ssimon HSM-type store, or LDAP stores, or... 3006160814Ssimon NOTE: The code is currently UNTESTED and isn't really used anywhere. 3007160814Ssimon [Richard Levitte] 3008160814Ssimon 3009160814Ssimon *) Add a generic structure called OPENSSL_ITEM. This can be used to 3010160814Ssimon pass a list of arguments to any function as well as provide a way 3011160814Ssimon for a function to pass data back to the caller. 3012160814Ssimon [Richard Levitte] 3013160814Ssimon 3014160814Ssimon *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 3015160814Ssimon works like BUF_strdup() but can be used to duplicate a portion of 3016160814Ssimon a string. The copy gets NUL-terminated. BUF_memdup() duplicates 3017160814Ssimon a memory area. 3018160814Ssimon [Richard Levitte] 3019160814Ssimon 3020160814Ssimon *) Add the function sk_find_ex() which works like sk_find(), but will 3021160814Ssimon return an index to an element even if an exact match couldn't be 3022160814Ssimon found. The index is guaranteed to point at the element where the 3023160814Ssimon searched-for key would be inserted to preserve sorting order. 3024160814Ssimon [Richard Levitte] 3025160814Ssimon 3026160814Ssimon *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 3027160814Ssimon takes an extra flags argument for optional functionality. Currently, 3028160814Ssimon the following flags are defined: 3029160814Ssimon 3030160814Ssimon OBJ_BSEARCH_VALUE_ON_NOMATCH 3031160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 3032160814Ssimon element where the comparing function returns a negative or zero 3033160814Ssimon number. 3034160814Ssimon 3035160814Ssimon OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 3036160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 3037160814Ssimon element where the comparing function returns zero. This is useful 3038160814Ssimon if there are more than one element where the comparing function 3039160814Ssimon returns zero. 3040160814Ssimon [Richard Levitte] 3041160814Ssimon 3042160814Ssimon *) Make it possible to create self-signed certificates with 'openssl ca' 3043160814Ssimon in such a way that the self-signed certificate becomes part of the 3044160814Ssimon CA database and uses the same mechanisms for serial number generation 3045160814Ssimon as all other certificate signing. The new flag '-selfsign' enables 3046160814Ssimon this functionality. Adapt CA.sh and CA.pl.in. 3047160814Ssimon [Richard Levitte] 3048160814Ssimon 3049160814Ssimon *) Add functionality to check the public key of a certificate request 3050160814Ssimon against a given private. This is useful to check that a certificate 3051160814Ssimon request can be signed by that key (self-signing). 3052160814Ssimon [Richard Levitte] 3053160814Ssimon 3054160814Ssimon *) Make it possible to have multiple active certificates with the same 3055160814Ssimon subject in the CA index file. This is done only if the keyword 3056160814Ssimon 'unique_subject' is set to 'no' in the main CA section (default 3057160814Ssimon if 'CA_default') of the configuration file. The value is saved 3058160814Ssimon with the database itself in a separate index attribute file, 3059160814Ssimon named like the index file with '.attr' appended to the name. 3060160814Ssimon [Richard Levitte] 3061160814Ssimon 3062160814Ssimon *) Generate muti valued AVAs using '+' notation in config files for 3063160814Ssimon req and dirName. 3064160814Ssimon [Steve Henson] 3065160814Ssimon 3066160814Ssimon *) Support for nameConstraints certificate extension. 3067160814Ssimon [Steve Henson] 3068160814Ssimon 3069160814Ssimon *) Support for policyConstraints certificate extension. 3070160814Ssimon [Steve Henson] 3071160814Ssimon 3072160814Ssimon *) Support for policyMappings certificate extension. 3073160814Ssimon [Steve Henson] 3074160814Ssimon 3075160814Ssimon *) Make sure the default DSA_METHOD implementation only uses its 3076160814Ssimon dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 3077160814Ssimon and change its own handlers to be NULL so as to remove unnecessary 3078160814Ssimon indirection. This lets alternative implementations fallback to the 3079160814Ssimon default implementation more easily. 3080160814Ssimon [Geoff Thorpe] 3081160814Ssimon 3082160814Ssimon *) Support for directoryName in GeneralName related extensions 3083160814Ssimon in config files. 3084160814Ssimon [Steve Henson] 3085160814Ssimon 3086160814Ssimon *) Make it possible to link applications using Makefile.shared. 3087160814Ssimon Make that possible even when linking against static libraries! 3088160814Ssimon [Richard Levitte] 3089160814Ssimon 3090160814Ssimon *) Support for single pass processing for S/MIME signing. This now 3091160814Ssimon means that S/MIME signing can be done from a pipe, in addition 3092160814Ssimon cleartext signing (multipart/signed type) is effectively streaming 3093160814Ssimon and the signed data does not need to be all held in memory. 3094160814Ssimon 3095160814Ssimon This is done with a new flag PKCS7_STREAM. When this flag is set 3096160814Ssimon PKCS7_sign() only initializes the PKCS7 structure and the actual signing 3097160814Ssimon is done after the data is output (and digests calculated) in 3098160814Ssimon SMIME_write_PKCS7(). 3099160814Ssimon [Steve Henson] 3100160814Ssimon 3101160814Ssimon *) Add full support for -rpath/-R, both in shared libraries and 3102160814Ssimon applications, at least on the platforms where it's known how 3103160814Ssimon to do it. 3104160814Ssimon [Richard Levitte] 3105160814Ssimon 3106160814Ssimon *) In crypto/ec/ec_mult.c, implement fast point multiplication with 3107160814Ssimon precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 3108160814Ssimon will now compute a table of multiples of the generator that 3109160814Ssimon makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 3110160814Ssimon faster (notably in the case of a single point multiplication, 3111160814Ssimon scalar * generator). 3112160814Ssimon [Nils Larsch, Bodo Moeller] 3113160814Ssimon 3114160814Ssimon *) IPv6 support for certificate extensions. The various extensions 3115160814Ssimon which use the IP:a.b.c.d can now take IPv6 addresses using the 3116160814Ssimon formats of RFC1884 2.2 . IPv6 addresses are now also displayed 3117160814Ssimon correctly. 3118160814Ssimon [Steve Henson] 3119160814Ssimon 3120160814Ssimon *) Added an ENGINE that implements RSA by performing private key 3121160814Ssimon exponentiations with the GMP library. The conversions to and from 3122160814Ssimon GMP's mpz_t format aren't optimised nor are any montgomery forms 3123160814Ssimon cached, and on x86 it appears OpenSSL's own performance has caught up. 3124160814Ssimon However there are likely to be other architectures where GMP could 3125160814Ssimon provide a boost. This ENGINE is not built in by default, but it can be 3126160814Ssimon specified at Configure time and should be accompanied by the necessary 3127160814Ssimon linker additions, eg; 3128160814Ssimon ./config -DOPENSSL_USE_GMP -lgmp 3129160814Ssimon [Geoff Thorpe] 3130160814Ssimon 3131160814Ssimon *) "openssl engine" will not display ENGINE/DSO load failure errors when 3132160814Ssimon testing availability of engines with "-t" - the old behaviour is 3133160814Ssimon produced by increasing the feature's verbosity with "-tt". 3134160814Ssimon [Geoff Thorpe] 3135160814Ssimon 3136160814Ssimon *) ECDSA routines: under certain error conditions uninitialized BN objects 3137160814Ssimon could be freed. Solution: make sure initialization is performed early 3138160814Ssimon enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 3139160814Ssimon via PR#459) 3140160814Ssimon [Lutz Jaenicke] 3141160814Ssimon 3142160814Ssimon *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 3143160814Ssimon and DH_METHOD (eg. by ENGINE implementations) to override the normal 3144160814Ssimon software implementations. For DSA and DH, parameter generation can 3145160814Ssimon also be overriden by providing the appropriate method callbacks. 3146160814Ssimon [Geoff Thorpe] 3147160814Ssimon 3148160814Ssimon *) Change the "progress" mechanism used in key-generation and 3149160814Ssimon primality testing to functions that take a new BN_GENCB pointer in 3150160814Ssimon place of callback/argument pairs. The new API functions have "_ex" 3151160814Ssimon postfixes and the older functions are reimplemented as wrappers for 3152160814Ssimon the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 3153160814Ssimon declarations of the old functions to help (graceful) attempts to 3154160814Ssimon migrate to the new functions. Also, the new key-generation API 3155160814Ssimon functions operate on a caller-supplied key-structure and return 3156160814Ssimon success/failure rather than returning a key or NULL - this is to 3157160814Ssimon help make "keygen" another member function of RSA_METHOD etc. 3158160814Ssimon 3159160814Ssimon Example for using the new callback interface: 3160160814Ssimon 3161160814Ssimon int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 3162160814Ssimon void *my_arg = ...; 3163160814Ssimon BN_GENCB my_cb; 3164160814Ssimon 3165160814Ssimon BN_GENCB_set(&my_cb, my_callback, my_arg); 3166160814Ssimon 3167160814Ssimon return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 3168160814Ssimon /* For the meaning of a, b in calls to my_callback(), see the 3169160814Ssimon * documentation of the function that calls the callback. 3170160814Ssimon * cb will point to my_cb; my_arg can be retrieved as cb->arg. 3171160814Ssimon * my_callback should return 1 if it wants BN_is_prime_ex() 3172160814Ssimon * to continue, or 0 to stop. 3173160814Ssimon */ 3174160814Ssimon 3175160814Ssimon [Geoff Thorpe] 3176160814Ssimon 3177160814Ssimon *) Change the ZLIB compression method to be stateful, and make it 3178160814Ssimon available to TLS with the number defined in 3179160814Ssimon draft-ietf-tls-compression-04.txt. 3180160814Ssimon [Richard Levitte] 3181160814Ssimon 3182160814Ssimon *) Add the ASN.1 structures and functions for CertificatePair, which 3183160814Ssimon is defined as follows (according to X.509_4thEditionDraftV6.pdf): 3184160814Ssimon 3185160814Ssimon CertificatePair ::= SEQUENCE { 3186160814Ssimon forward [0] Certificate OPTIONAL, 3187160814Ssimon reverse [1] Certificate OPTIONAL, 3188160814Ssimon -- at least one of the pair shall be present -- } 3189160814Ssimon 3190160814Ssimon Also implement the PEM functions to read and write certificate 3191160814Ssimon pairs, and defined the PEM tag as "CERTIFICATE PAIR". 3192160814Ssimon 3193160814Ssimon This needed to be defined, mostly for the sake of the LDAP 3194160814Ssimon attribute crossCertificatePair, but may prove useful elsewhere as 3195160814Ssimon well. 3196160814Ssimon [Richard Levitte] 3197160814Ssimon 3198160814Ssimon *) Make it possible to inhibit symlinking of shared libraries in 3199160814Ssimon Makefile.shared, for Cygwin's sake. 3200160814Ssimon [Richard Levitte] 3201160814Ssimon 3202160814Ssimon *) Extend the BIGNUM API by creating a function 3203160814Ssimon void BN_set_negative(BIGNUM *a, int neg); 3204160814Ssimon and a macro that behave like 3205160814Ssimon int BN_is_negative(const BIGNUM *a); 3206160814Ssimon 3207160814Ssimon to avoid the need to access 'a->neg' directly in applications. 3208160814Ssimon [Nils Larsch] 3209160814Ssimon 3210160814Ssimon *) Implement fast modular reduction for pseudo-Mersenne primes 3211160814Ssimon used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 3212160814Ssimon EC_GROUP_new_curve_GFp() will now automatically use this 3213160814Ssimon if applicable. 3214160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3215160814Ssimon 3216160814Ssimon *) Add new lock type (CRYPTO_LOCK_BN). 3217160814Ssimon [Bodo Moeller] 3218160814Ssimon 3219160814Ssimon *) Change the ENGINE framework to automatically load engines 3220160814Ssimon dynamically from specific directories unless they could be 3221160814Ssimon found to already be built in or loaded. Move all the 3222160814Ssimon current engines except for the cryptodev one to a new 3223160814Ssimon directory engines/. 3224160814Ssimon The engines in engines/ are built as shared libraries if 3225160814Ssimon the "shared" options was given to ./Configure or ./config. 3226160814Ssimon Otherwise, they are inserted in libcrypto.a. 3227160814Ssimon /usr/local/ssl/engines is the default directory for dynamic 3228160814Ssimon engines, but that can be overriden at configure time through 3229160814Ssimon the usual use of --prefix and/or --openssldir, and at run 3230160814Ssimon time with the environment variable OPENSSL_ENGINES. 3231160814Ssimon [Geoff Thorpe and Richard Levitte] 3232160814Ssimon 3233160814Ssimon *) Add Makefile.shared, a helper makefile to build shared 3234160814Ssimon libraries. Addapt Makefile.org. 3235160814Ssimon [Richard Levitte] 3236160814Ssimon 3237160814Ssimon *) Add version info to Win32 DLLs. 3238160814Ssimon [Peter 'Luna' Runestig" <peter@runestig.com>] 3239160814Ssimon 3240160814Ssimon *) Add new 'medium level' PKCS#12 API. Certificates and keys 3241160814Ssimon can be added using this API to created arbitrary PKCS#12 3242160814Ssimon files while avoiding the low level API. 3243160814Ssimon 3244160814Ssimon New options to PKCS12_create(), key or cert can be NULL and 3245160814Ssimon will then be omitted from the output file. The encryption 3246160814Ssimon algorithm NIDs can be set to -1 for no encryption, the mac 3247160814Ssimon iteration count can be set to 0 to omit the mac. 3248160814Ssimon 3249160814Ssimon Enhance pkcs12 utility by making the -nokeys and -nocerts 3250160814Ssimon options work when creating a PKCS#12 file. New option -nomac 3251160814Ssimon to omit the mac, NONE can be set for an encryption algorithm. 3252160814Ssimon New code is modified to use the enhanced PKCS12_create() 3253160814Ssimon instead of the low level API. 3254160814Ssimon [Steve Henson] 3255160814Ssimon 3256160814Ssimon *) Extend ASN1 encoder to support indefinite length constructed 3257160814Ssimon encoding. This can output sequences tags and octet strings in 3258160814Ssimon this form. Modify pk7_asn1.c to support indefinite length 3259160814Ssimon encoding. This is experimental and needs additional code to 3260160814Ssimon be useful, such as an ASN1 bio and some enhanced streaming 3261160814Ssimon PKCS#7 code. 3262160814Ssimon 3263160814Ssimon Extend template encode functionality so that tagging is passed 3264160814Ssimon down to the template encoder. 3265160814Ssimon [Steve Henson] 3266160814Ssimon 3267160814Ssimon *) Let 'openssl req' fail if an argument to '-newkey' is not 3268160814Ssimon recognized instead of using RSA as a default. 3269160814Ssimon [Bodo Moeller] 3270160814Ssimon 3271160814Ssimon *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 3272160814Ssimon As these are not official, they are not included in "ALL"; 3273160814Ssimon the "ECCdraft" ciphersuite group alias can be used to select them. 3274160814Ssimon [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 3275160814Ssimon 3276160814Ssimon *) Add ECDH engine support. 3277160814Ssimon [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 3278160814Ssimon 3279160814Ssimon *) Add ECDH in new directory crypto/ecdh/. 3280160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 3281160814Ssimon 3282160814Ssimon *) Let BN_rand_range() abort with an error after 100 iterations 3283160814Ssimon without success (which indicates a broken PRNG). 3284160814Ssimon [Bodo Moeller] 3285160814Ssimon 3286160814Ssimon *) Change BN_mod_sqrt() so that it verifies that the input value 3287160814Ssimon is really the square of the return value. (Previously, 3288160814Ssimon BN_mod_sqrt would show GIGO behaviour.) 3289160814Ssimon [Bodo Moeller] 3290160814Ssimon 3291160814Ssimon *) Add named elliptic curves over binary fields from X9.62, SECG, 3292160814Ssimon and WAP/WTLS; add OIDs that were still missing. 3293160814Ssimon 3294160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3295160814Ssimon (Sun Microsystems Laboratories)] 3296160814Ssimon 3297160814Ssimon *) Extend the EC library for elliptic curves over binary fields 3298160814Ssimon (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 3299160814Ssimon New EC_METHOD: 3300160814Ssimon 3301160814Ssimon EC_GF2m_simple_method 3302160814Ssimon 3303160814Ssimon New API functions: 3304160814Ssimon 3305160814Ssimon EC_GROUP_new_curve_GF2m 3306160814Ssimon EC_GROUP_set_curve_GF2m 3307160814Ssimon EC_GROUP_get_curve_GF2m 3308160814Ssimon EC_POINT_set_affine_coordinates_GF2m 3309160814Ssimon EC_POINT_get_affine_coordinates_GF2m 3310160814Ssimon EC_POINT_set_compressed_coordinates_GF2m 3311160814Ssimon 3312160814Ssimon Point compression for binary fields is disabled by default for 3313160814Ssimon patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 3314160814Ssimon enable it). 3315160814Ssimon 3316160814Ssimon As binary polynomials are represented as BIGNUMs, various members 3317160814Ssimon of the EC_GROUP and EC_POINT data structures can be shared 3318160814Ssimon between the implementations for prime fields and binary fields; 3319160814Ssimon the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 3320160814Ssimon are essentially identical to their ..._GFp counterparts. 3321160814Ssimon (For simplicity, the '..._GFp' prefix has been dropped from 3322160814Ssimon various internal method names.) 3323160814Ssimon 3324160814Ssimon An internal 'field_div' method (similar to 'field_mul' and 3325160814Ssimon 'field_sqr') has been added; this is used only for binary fields. 3326160814Ssimon 3327160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3328160814Ssimon (Sun Microsystems Laboratories)] 3329160814Ssimon 3330160814Ssimon *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 3331160814Ssimon through methods ('mul', 'precompute_mult'). 3332160814Ssimon 3333160814Ssimon The generic implementations (now internally called 'ec_wNAF_mul' 3334160814Ssimon and 'ec_wNAF_precomputed_mult') remain the default if these 3335160814Ssimon methods are undefined. 3336160814Ssimon 3337160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3338160814Ssimon (Sun Microsystems Laboratories)] 3339160814Ssimon 3340160814Ssimon *) New function EC_GROUP_get_degree, which is defined through 3341160814Ssimon EC_METHOD. For curves over prime fields, this returns the bit 3342160814Ssimon length of the modulus. 3343160814Ssimon 3344160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3345160814Ssimon (Sun Microsystems Laboratories)] 3346160814Ssimon 3347160814Ssimon *) New functions EC_GROUP_dup, EC_POINT_dup. 3348160814Ssimon (These simply call ..._new and ..._copy). 3349160814Ssimon 3350160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3351160814Ssimon (Sun Microsystems Laboratories)] 3352160814Ssimon 3353160814Ssimon *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 3354160814Ssimon Polynomials are represented as BIGNUMs (where the sign bit is not 3355160814Ssimon used) in the following functions [macros]: 3356160814Ssimon 3357160814Ssimon BN_GF2m_add 3358160814Ssimon BN_GF2m_sub [= BN_GF2m_add] 3359160814Ssimon BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 3360160814Ssimon BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 3361160814Ssimon BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 3362160814Ssimon BN_GF2m_mod_inv 3363160814Ssimon BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 3364160814Ssimon BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 3365160814Ssimon BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 3366160814Ssimon BN_GF2m_cmp [= BN_ucmp] 3367160814Ssimon 3368160814Ssimon (Note that only the 'mod' functions are actually for fields GF(2^m). 3369160814Ssimon BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 3370160814Ssimon 3371160814Ssimon For some functions, an the irreducible polynomial defining a 3372160814Ssimon field can be given as an 'unsigned int[]' with strictly 3373160814Ssimon decreasing elements giving the indices of those bits that are set; 3374160814Ssimon i.e., p[] represents the polynomial 3375160814Ssimon f(t) = t^p[0] + t^p[1] + ... + t^p[k] 3376160814Ssimon where 3377160814Ssimon p[0] > p[1] > ... > p[k] = 0. 3378160814Ssimon This applies to the following functions: 3379160814Ssimon 3380160814Ssimon BN_GF2m_mod_arr 3381160814Ssimon BN_GF2m_mod_mul_arr 3382160814Ssimon BN_GF2m_mod_sqr_arr 3383160814Ssimon BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 3384160814Ssimon BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 3385160814Ssimon BN_GF2m_mod_exp_arr 3386160814Ssimon BN_GF2m_mod_sqrt_arr 3387160814Ssimon BN_GF2m_mod_solve_quad_arr 3388160814Ssimon BN_GF2m_poly2arr 3389160814Ssimon BN_GF2m_arr2poly 3390160814Ssimon 3391160814Ssimon Conversion can be performed by the following functions: 3392160814Ssimon 3393160814Ssimon BN_GF2m_poly2arr 3394160814Ssimon BN_GF2m_arr2poly 3395160814Ssimon 3396160814Ssimon bntest.c has additional tests for binary polynomial arithmetic. 3397160814Ssimon 3398160814Ssimon Two implementations for BN_GF2m_mod_div() are available. 3399160814Ssimon The default algorithm simply uses BN_GF2m_mod_inv() and 3400160814Ssimon BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 3401160814Ssimon if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 3402160814Ssimon copyright notice in crypto/bn/bn_gf2m.c before enabling it). 3403160814Ssimon 3404160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3405160814Ssimon (Sun Microsystems Laboratories)] 3406160814Ssimon 3407160814Ssimon *) Add new error code 'ERR_R_DISABLED' that can be used when some 3408160814Ssimon functionality is disabled at compile-time. 3409160814Ssimon [Douglas Stebila <douglas.stebila@sun.com>] 3410160814Ssimon 3411160814Ssimon *) Change default behaviour of 'openssl asn1parse' so that more 3412160814Ssimon information is visible when viewing, e.g., a certificate: 3413160814Ssimon 3414160814Ssimon Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 3415160814Ssimon mode the content of non-printable OCTET STRINGs is output in a 3416160814Ssimon style similar to INTEGERs, but with '[HEX DUMP]' prepended to 3417160814Ssimon avoid the appearance of a printable string. 3418160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3419160814Ssimon 3420160814Ssimon *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 3421160814Ssimon functions 3422160814Ssimon EC_GROUP_set_asn1_flag() 3423160814Ssimon EC_GROUP_get_asn1_flag() 3424160814Ssimon EC_GROUP_set_point_conversion_form() 3425160814Ssimon EC_GROUP_get_point_conversion_form() 3426160814Ssimon These control ASN1 encoding details: 3427160814Ssimon - Curves (i.e., groups) are encoded explicitly unless asn1_flag 3428160814Ssimon has been set to OPENSSL_EC_NAMED_CURVE. 3429160814Ssimon - Points are encoded in uncompressed form by default; options for 3430160814Ssimon asn1_for are as for point2oct, namely 3431160814Ssimon POINT_CONVERSION_COMPRESSED 3432160814Ssimon POINT_CONVERSION_UNCOMPRESSED 3433160814Ssimon POINT_CONVERSION_HYBRID 3434160814Ssimon 3435160814Ssimon Also add 'seed' and 'seed_len' members to EC_GROUP with access 3436160814Ssimon functions 3437160814Ssimon EC_GROUP_set_seed() 3438160814Ssimon EC_GROUP_get0_seed() 3439160814Ssimon EC_GROUP_get_seed_len() 3440160814Ssimon This is used only for ASN1 purposes (so far). 3441160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3442160814Ssimon 3443160814Ssimon *) Add 'field_type' member to EC_METHOD, which holds the NID 3444160814Ssimon of the appropriate field type OID. The new function 3445160814Ssimon EC_METHOD_get_field_type() returns this value. 3446160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3447160814Ssimon 3448160814Ssimon *) Add functions 3449160814Ssimon EC_POINT_point2bn() 3450160814Ssimon EC_POINT_bn2point() 3451160814Ssimon EC_POINT_point2hex() 3452160814Ssimon EC_POINT_hex2point() 3453160814Ssimon providing useful interfaces to EC_POINT_point2oct() and 3454160814Ssimon EC_POINT_oct2point(). 3455160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3456160814Ssimon 3457160814Ssimon *) Change internals of the EC library so that the functions 3458160814Ssimon EC_GROUP_set_generator() 3459160814Ssimon EC_GROUP_get_generator() 3460160814Ssimon EC_GROUP_get_order() 3461160814Ssimon EC_GROUP_get_cofactor() 3462160814Ssimon are implemented directly in crypto/ec/ec_lib.c and not dispatched 3463160814Ssimon to methods, which would lead to unnecessary code duplication when 3464160814Ssimon adding different types of curves. 3465160814Ssimon [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 3466160814Ssimon 3467160814Ssimon *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 3468160814Ssimon arithmetic, and such that modified wNAFs are generated 3469160814Ssimon (which avoid length expansion in many cases). 3470160814Ssimon [Bodo Moeller] 3471160814Ssimon 3472160814Ssimon *) Add a function EC_GROUP_check_discriminant() (defined via 3473160814Ssimon EC_METHOD) that verifies that the curve discriminant is non-zero. 3474160814Ssimon 3475160814Ssimon Add a function EC_GROUP_check() that makes some sanity tests 3476160814Ssimon on a EC_GROUP, its generator and order. This includes 3477160814Ssimon EC_GROUP_check_discriminant(). 3478160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3479160814Ssimon 3480160814Ssimon *) Add ECDSA in new directory crypto/ecdsa/. 3481160814Ssimon 3482160814Ssimon Add applications 'openssl ecparam' and 'openssl ecdsa' 3483160814Ssimon (these are based on 'openssl dsaparam' and 'openssl dsa'). 3484160814Ssimon 3485160814Ssimon ECDSA support is also included in various other files across the 3486160814Ssimon library. Most notably, 3487160814Ssimon - 'openssl req' now has a '-newkey ecdsa:file' option; 3488160814Ssimon - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 3489160814Ssimon - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 3490160814Ssimon d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 3491160814Ssimon them suitable for ECDSA where domain parameters must be 3492160814Ssimon extracted before the specific public key; 3493160814Ssimon - ECDSA engine support has been added. 3494160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3495160814Ssimon 3496160814Ssimon *) Include some named elliptic curves, and add OIDs from X9.62, 3497160814Ssimon SECG, and WAP/WTLS. Each curve can be obtained from the new 3498160814Ssimon function 3499160814Ssimon EC_GROUP_new_by_curve_name(), 3500160814Ssimon and the list of available named curves can be obtained with 3501160814Ssimon EC_get_builtin_curves(). 3502160814Ssimon Also add a 'curve_name' member to EC_GROUP objects, which can be 3503160814Ssimon accessed via 3504160814Ssimon EC_GROUP_set_curve_name() 3505160814Ssimon EC_GROUP_get_curve_name() 3506160814Ssimon [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 3507160814Ssimon 3508160814Ssimon *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 3509160814Ssimon was actually never needed) and in BN_mul(). The removal in BN_mul() 3510160814Ssimon required a small change in bn_mul_part_recursive() and the addition 3511160814Ssimon of the functions bn_cmp_part_words(), bn_sub_part_words() and 3512160814Ssimon bn_add_part_words(), which do the same thing as bn_cmp_words(), 3513160814Ssimon bn_sub_words() and bn_add_words() except they take arrays with 3514160814Ssimon differing sizes. 3515160814Ssimon [Richard Levitte] 3516160814Ssimon 3517194206Ssimon Changes between 0.9.7l and 0.9.7m [23 Feb 2007] 3518194206Ssimon 3519167612Ssimon *) Cleanse PEM buffers before freeing them since they may contain 3520167612Ssimon sensitive data. 3521167612Ssimon [Benjamin Bennett <ben@psc.edu>] 3522167612Ssimon 3523167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 3524167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 3525167612Ssimon authentication-only ciphersuites. 3526167612Ssimon [Bodo Moeller] 3527167612Ssimon 3528167612Ssimon *) Since AES128 and AES256 share a single mask bit in the logic of 3529167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 3530167612Ssimon kludge to work properly if AES128 is available and AES256 isn't. 3531167612Ssimon [Victor Duchovni] 3532167612Ssimon 3533194206Ssimon *) Expand security boundary to match 1.1.1 module. 3534194206Ssimon [Steve Henson] 3535194206Ssimon 3536194206Ssimon *) Remove redundant features: hash file source, editing of test vectors 3537194206Ssimon modify fipsld to use external fips_premain.c signature. 3538194206Ssimon [Steve Henson] 3539194206Ssimon 3540194206Ssimon *) New perl script mkfipsscr.pl to create shell scripts or batch files to 3541194206Ssimon run algorithm test programs. 3542194206Ssimon [Steve Henson] 3543194206Ssimon 3544194206Ssimon *) Make algorithm test programs more tolerant of whitespace. 3545194206Ssimon [Steve Henson] 3546194206Ssimon 3547167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 3548167612Ssimon protocol version while receiving ClientHello even if the 3549167612Ssimon ClientHello is fragmented. (The server can't insist on the 3550167612Ssimon particular protocol version it has chosen before the ServerHello 3551167612Ssimon message has informed the client about his choice.) 3552167612Ssimon [Bodo Moeller] 3553167612Ssimon 3554167612Ssimon *) Load error codes if they are not already present instead of using a 3555167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 3556167612Ssimon [Steve Henson] 3557167612Ssimon 3558167612Ssimon Changes between 0.9.7k and 0.9.7l [28 Sep 2006] 3559167612Ssimon 3560167612Ssimon *) Introduce limits to prevent malicious keys being able to 3561167612Ssimon cause a denial of service. (CVE-2006-2940) 3562167612Ssimon [Steve Henson, Bodo Moeller] 3563167612Ssimon 3564167612Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 3565167612Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 3566167612Ssimon 3567167612Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 3568167612Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 3569167612Ssimon 3570167612Ssimon *) Fix SSL client code which could crash if connecting to a 3571167612Ssimon malicious SSLv2 server. (CVE-2006-4343) 3572167612Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 3573167612Ssimon 3574162911Ssimon *) Change ciphersuite string processing so that an explicit 3575162911Ssimon ciphersuite selects this one ciphersuite (so that "AES256-SHA" 3576162911Ssimon will no longer include "AES128-SHA"), and any other similar 3577162911Ssimon ciphersuite (same bitmap) from *other* protocol versions (so that 3578162911Ssimon "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 3579162911Ssimon SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 3580162911Ssimon changes from 0.9.8b and 0.9.8d. 3581162911Ssimon [Bodo Moeller] 3582162911Ssimon 3583162911Ssimon Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 3584162911Ssimon 3585162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 3586162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 3587162911Ssimon 3588162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 3589162911Ssimon possible instead of select(), since the latter has some 3590162911Ssimon undesirable limitations. 3591162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 3592162911Ssimon 3593162911Ssimon *) Disable rogue ciphersuites: 3594162911Ssimon 3595162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 3596162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 3597162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 3598162911Ssimon 3599162911Ssimon The latter two were purportedly from 3600162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 3601162911Ssimon appear there. 3602162911Ssimon 3603162911Ssimon Also deactive the remaining ciphersuites from 3604162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 3605162911Ssimon unofficial, and the ID has long expired. 3606162911Ssimon [Bodo Moeller] 3607162911Ssimon 3608162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 3609162911Ssimon dual-core machines) and other potential thread-safety issues. 3610162911Ssimon [Bodo Moeller] 3611162911Ssimon 3612162911Ssimon Changes between 0.9.7i and 0.9.7j [04 May 2006] 3613162911Ssimon 3614162911Ssimon *) Adapt fipsld and the build system to link against the validated FIPS 3615162911Ssimon module in FIPS mode. 3616162911Ssimon [Steve Henson] 3617162911Ssimon 3618162911Ssimon *) Fixes for VC++ 2005 build under Windows. 3619162911Ssimon [Steve Henson] 3620162911Ssimon 3621162911Ssimon *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 3622162911Ssimon from a Windows bash shell such as MSYS. It is autodetected from the 3623162911Ssimon "config" script when run from a VC++ environment. Modify standard VC++ 3624162911Ssimon build to use fipscanister.o from the GNU make build. 3625162911Ssimon [Steve Henson] 3626162911Ssimon 3627160814Ssimon Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 3628160814Ssimon 3629160814Ssimon *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 3630160814Ssimon The value now differs depending on if you build for FIPS or not. 3631160814Ssimon BEWARE! A program linked with a shared FIPSed libcrypto can't be 3632160814Ssimon safely run with a non-FIPSed libcrypto, as it may crash because of 3633160814Ssimon the difference induced by this change. 3634160814Ssimon [Andy Polyakov] 3635160814Ssimon 3636160814Ssimon Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 3637160814Ssimon 3638160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 3639160814Ssimon (part of SSL_OP_ALL). This option used to disable the 3640160814Ssimon countermeasure against man-in-the-middle protocol-version 3641160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 3642160814Ssimon idea. (CVE-2005-2969) 3643160814Ssimon 3644160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 3645160814Ssimon for Information Security, National Institute of Advanced Industrial 3646160814Ssimon Science and Technology [AIST], Japan)] 3647160814Ssimon 3648160814Ssimon *) Minimal support for X9.31 signatures and PSS padding modes. This is 3649160814Ssimon mainly for FIPS compliance and not fully integrated at this stage. 3650160814Ssimon [Steve Henson] 3651160814Ssimon 3652160814Ssimon *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 3653160814Ssimon the exponentiation using a fixed-length exponent. (Otherwise, 3654160814Ssimon the information leaked through timing could expose the secret key 3655160814Ssimon after many signatures; cf. Bleichenbacher's attack on DSA with 3656160814Ssimon biased k.) 3657160814Ssimon [Bodo Moeller] 3658160814Ssimon 3659160814Ssimon *) Make a new fixed-window mod_exp implementation the default for 3660160814Ssimon RSA, DSA, and DH private-key operations so that the sequence of 3661160814Ssimon squares and multiplies and the memory access pattern are 3662160814Ssimon independent of the particular secret key. This will mitigate 3663160814Ssimon cache-timing and potential related attacks. 3664160814Ssimon 3665160814Ssimon BN_mod_exp_mont_consttime() is the new exponentiation implementation, 3666160814Ssimon and this is automatically used by BN_mod_exp_mont() if the new flag 3667160814Ssimon BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 3668160814Ssimon will use this BN flag for private exponents unless the flag 3669160814Ssimon RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 3670160814Ssimon DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 3671160814Ssimon 3672160814Ssimon [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 3673160814Ssimon 3674160814Ssimon *) Change the client implementation for SSLv23_method() and 3675160814Ssimon SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 3676160814Ssimon Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 3677160814Ssimon (Previously, the SSL 2.0 backwards compatible Client Hello 3678160814Ssimon message format would be used even with SSL_OP_NO_SSLv2.) 3679160814Ssimon [Bodo Moeller] 3680160814Ssimon 3681160814Ssimon *) Add support for smime-type MIME parameter in S/MIME messages which some 3682160814Ssimon clients need. 3683160814Ssimon [Steve Henson] 3684160814Ssimon 3685160814Ssimon *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 3686160814Ssimon a threadsafe manner. Modify rsa code to use new function and add calls 3687160814Ssimon to dsa and dh code (which had race conditions before). 3688160814Ssimon [Steve Henson] 3689160814Ssimon 3690160814Ssimon *) Include the fixed error library code in the C error file definitions 3691160814Ssimon instead of fixing them up at runtime. This keeps the error code 3692160814Ssimon structures constant. 3693160814Ssimon [Steve Henson] 3694160814Ssimon 3695160814Ssimon Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 3696160814Ssimon 3697160814Ssimon [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 3698160814Ssimon OpenSSL 0.9.8.] 3699160814Ssimon 3700160814Ssimon *) Fixes for newer kerberos headers. NB: the casts are needed because 3701160814Ssimon the 'length' field is signed on one version and unsigned on another 3702160814Ssimon with no (?) obvious way to tell the difference, without these VC++ 3703160814Ssimon complains. Also the "definition" of FAR (blank) is no longer included 3704160814Ssimon nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 3705160814Ssimon some needed definitions. 3706160814Ssimon [Steve Henson] 3707160814Ssimon 3708160814Ssimon *) Undo Cygwin change. 3709160814Ssimon [Ulf M�ller] 3710160814Ssimon 3711160814Ssimon *) Added support for proxy certificates according to RFC 3820. 3712160814Ssimon Because they may be a security thread to unaware applications, 3713160814Ssimon they must be explicitely allowed in run-time. See 3714160814Ssimon docs/HOWTO/proxy_certificates.txt for further information. 3715160814Ssimon [Richard Levitte] 3716160814Ssimon 3717160814Ssimon Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 3718160814Ssimon 3719160814Ssimon *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 3720160814Ssimon server and client random values. Previously 3721160814Ssimon (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 3722160814Ssimon less random data when sizeof(time_t) > 4 (some 64 bit platforms). 3723160814Ssimon 3724160814Ssimon This change has negligible security impact because: 3725160814Ssimon 3726160814Ssimon 1. Server and client random values still have 24 bytes of pseudo random 3727160814Ssimon data. 3728160814Ssimon 3729160814Ssimon 2. Server and client random values are sent in the clear in the initial 3730160814Ssimon handshake. 3731160814Ssimon 3732160814Ssimon 3. The master secret is derived using the premaster secret (48 bytes in 3733160814Ssimon size for static RSA ciphersuites) as well as client server and random 3734160814Ssimon values. 3735160814Ssimon 3736160814Ssimon The OpenSSL team would like to thank the UK NISCC for bringing this issue 3737160814Ssimon to our attention. 3738160814Ssimon 3739160814Ssimon [Stephen Henson, reported by UK NISCC] 3740160814Ssimon 3741160814Ssimon *) Use Windows randomness collection on Cygwin. 3742160814Ssimon [Ulf M�ller] 3743160814Ssimon 3744160814Ssimon *) Fix hang in EGD/PRNGD query when communication socket is closed 3745160814Ssimon prematurely by EGD/PRNGD. 3746160814Ssimon [Darren Tucker <dtucker@zip.com.au> via Lutz J�nicke, resolves #1014] 3747160814Ssimon 3748160814Ssimon *) Prompt for pass phrases when appropriate for PKCS12 input format. 3749160814Ssimon [Steve Henson] 3750160814Ssimon 3751160814Ssimon *) Back-port of selected performance improvements from development 3752160814Ssimon branch, as well as improved support for PowerPC platforms. 3753160814Ssimon [Andy Polyakov] 3754160814Ssimon 3755160814Ssimon *) Add lots of checks for memory allocation failure, error codes to indicate 3756160814Ssimon failure and freeing up memory if a failure occurs. 3757160814Ssimon [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 3758160814Ssimon 3759160814Ssimon *) Add new -passin argument to dgst. 3760160814Ssimon [Steve Henson] 3761160814Ssimon 3762160814Ssimon *) Perform some character comparisons of different types in X509_NAME_cmp: 3763160814Ssimon this is needed for some certificates that reencode DNs into UTF8Strings 3764160814Ssimon (in violation of RFC3280) and can't or wont issue name rollover 3765160814Ssimon certificates. 3766160814Ssimon [Steve Henson] 3767160814Ssimon 3768160814Ssimon *) Make an explicit check during certificate validation to see that 3769160814Ssimon the CA setting in each certificate on the chain is correct. As a 3770160814Ssimon side effect always do the following basic checks on extensions, 3771160814Ssimon not just when there's an associated purpose to the check: 3772160814Ssimon 3773160814Ssimon - if there is an unhandled critical extension (unless the user 3774160814Ssimon has chosen to ignore this fault) 3775160814Ssimon - if the path length has been exceeded (if one is set at all) 3776160814Ssimon - that certain extensions fit the associated purpose (if one has 3777160814Ssimon been given) 3778160814Ssimon [Richard Levitte] 3779160814Ssimon 3780142425Snectar Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 3781142425Snectar 3782142425Snectar *) Avoid a race condition when CRLs are checked in a multi threaded 3783142425Snectar environment. This would happen due to the reordering of the revoked 3784142425Snectar entries during signature checking and serial number lookup. Now the 3785142425Snectar encoding is cached and the serial number sort performed under a lock. 3786142425Snectar Add new STACK function sk_is_sorted(). 3787142425Snectar [Steve Henson] 3788142425Snectar 3789142425Snectar *) Add Delta CRL to the extension code. 3790142425Snectar [Steve Henson] 3791142425Snectar 3792142425Snectar *) Various fixes to s3_pkt.c so alerts are sent properly. 3793142425Snectar [David Holmes <d.holmes@f5.com>] 3794142425Snectar 3795142425Snectar *) Reduce the chances of duplicate issuer name and serial numbers (in 3796142425Snectar violation of RFC3280) using the OpenSSL certificate creation utilities. 3797142425Snectar This is done by creating a random 64 bit value for the initial serial 3798142425Snectar number when a serial number file is created or when a self signed 3799142425Snectar certificate is created using 'openssl req -x509'. The initial serial 3800142425Snectar number file is created using 'openssl x509 -next_serial' in CA.pl 3801142425Snectar rather than being initialized to 1. 3802142425Snectar [Steve Henson] 3803142425Snectar 3804127128Snectar Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 3805127128Snectar 3806127128Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 3807160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 3808127128Snectar [Joe Orton, Steve Henson] 3809127128Snectar 3810127128Snectar *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 3811160814Ssimon (CVE-2004-0112) 3812127128Snectar [Joe Orton, Steve Henson] 3813127128Snectar 3814127128Snectar *) Make it possible to have multiple active certificates with the same 3815127128Snectar subject in the CA index file. This is done only if the keyword 3816127128Snectar 'unique_subject' is set to 'no' in the main CA section (default 3817127128Snectar if 'CA_default') of the configuration file. The value is saved 3818127128Snectar with the database itself in a separate index attribute file, 3819127128Snectar named like the index file with '.attr' appended to the name. 3820127128Snectar [Richard Levitte] 3821127128Snectar 3822127128Snectar *) X509 verify fixes. Disable broken certificate workarounds when 3823127128Snectar X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 3824127128Snectar keyUsage extension present. Don't accept CRLs with unhandled critical 3825127128Snectar extensions: since verify currently doesn't process CRL extensions this 3826127128Snectar rejects a CRL with *any* critical extensions. Add new verify error codes 3827127128Snectar for these cases. 3828127128Snectar [Steve Henson] 3829127128Snectar 3830127128Snectar *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 3831127128Snectar A clarification of RFC2560 will require the use of OCTET STRINGs and 3832127128Snectar some implementations cannot handle the current raw format. Since OpenSSL 3833127128Snectar copies and compares OCSP nonces as opaque blobs without any attempt at 3834127128Snectar parsing them this should not create any compatibility issues. 3835127128Snectar [Steve Henson] 3836127128Snectar 3837127128Snectar *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 3838127128Snectar calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 3839127128Snectar this HMAC (and other) operations are several times slower than OpenSSL 3840127128Snectar < 0.9.7. 3841127128Snectar [Steve Henson] 3842127128Snectar 3843127128Snectar *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 3844127128Snectar [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 3845127128Snectar 3846127128Snectar *) Use the correct content when signing type "other". 3847127128Snectar [Steve Henson] 3848127128Snectar 3849120631Snectar Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 3850120631Snectar 3851120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 3852120631Snectar 3853120631Snectar Stop out of bounds reads in the ASN1 code when presented with 3854160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 3855120631Snectar 3856160814Ssimon Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 3857120631Snectar 3858120631Snectar If verify callback ignores invalid public key errors don't try to check 3859120631Snectar certificate signature with the NULL public key. 3860120631Snectar 3861120631Snectar [Steve Henson] 3862120631Snectar 3863120631Snectar *) New -ignore_err option in ocsp application to stop the server 3864120631Snectar exiting on the first error in a request. 3865120631Snectar [Steve Henson] 3866120631Snectar 3867120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 3868120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 3869120631Snectar specifications. 3870120631Snectar [Steve Henson] 3871120631Snectar 3872120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 3873120631Snectar extra data after the compression methods not only for TLS 1.0 3874120631Snectar but also for SSL 3.0 (as required by the specification). 3875120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 3876120631Snectar 3877120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 3878120631Snectar when it's 512 *bits* long, not 512 bytes. 3879120631Snectar [Richard Levitte] 3880120631Snectar 3881120631Snectar *) Change AES_cbc_encrypt() so it outputs exact multiple of 3882120631Snectar blocks during encryption. 3883120631Snectar [Richard Levitte] 3884120631Snectar 3885120631Snectar *) Various fixes to base64 BIO and non blocking I/O. On write 3886120631Snectar flushes were not handled properly if the BIO retried. On read 3887120631Snectar data was not being buffered properly and had various logic bugs. 3888120631Snectar This also affects blocking I/O when the data being decoded is a 3889120631Snectar certain size. 3890120631Snectar [Steve Henson] 3891120631Snectar 3892120631Snectar *) Various S/MIME bugfixes and compatibility changes: 3893120631Snectar output correct application/pkcs7 MIME type if 3894120631Snectar PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 3895120631Snectar Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 3896120631Snectar of files as .eml work). Correctly handle very long lines in MIME 3897120631Snectar parser. 3898120631Snectar [Steve Henson] 3899120631Snectar 3900120631Snectar Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 3901120631Snectar 3902120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 3903120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 3904120631Snectar a protocol version number mismatch like a decryption error 3905120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 3906120631Snectar [Bodo Moeller] 3907120631Snectar 3908120631Snectar *) Turn on RSA blinding by default in the default implementation 3909120631Snectar to avoid a timing attack. Applications that don't want it can call 3910120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 3911120631Snectar They would be ill-advised to do so in most cases. 3912120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 3913120631Snectar 3914120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 3915120631Snectar seeded (in this case, the secret RSA exponent is abused as 3916120631Snectar an unpredictable seed -- if it is not unpredictable, there 3917120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 3918120631Snectar by remembering the creator's thread ID in rsa->blinding and 3919120631Snectar having all other threads use local one-time blinding factors 3920120631Snectar (this requires more computation than sharing rsa->blinding, but 3921120631Snectar avoids excessive locking; and if an RSA object is not shared 3922120631Snectar between threads, blinding will still be very fast). 3923120631Snectar [Bodo Moeller] 3924120631Snectar 3925120631Snectar *) Fixed a typo bug that would cause ENGINE_set_default() to set an 3926120631Snectar ENGINE as defaults for all supported algorithms irrespective of 3927120631Snectar the 'flags' parameter. 'flags' is now honoured, so applications 3928120631Snectar should make sure they are passing it correctly. 3929120631Snectar [Geoff Thorpe] 3930120631Snectar 3931120631Snectar *) Target "mingw" now allows native Windows code to be generated in 3932120631Snectar the Cygwin environment as well as with the MinGW compiler. 3933120631Snectar [Ulf Moeller] 3934120631Snectar 3935111147Snectar Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 3936111147Snectar 3937111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 3938111147Snectar via timing by performing a MAC computation even if incorrrect 3939111147Snectar block cipher padding has been found. This is a countermeasure 3940111147Snectar against active attacks where the attacker has to distinguish 3941160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 3942111147Snectar 3943111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 3944111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 3945111147Snectar Martin Vuagnoux (EPFL, Ilion)] 3946111147Snectar 3947111147Snectar *) Make the no-err option work as intended. The intention with no-err 3948111147Snectar is not to have the whole error stack handling routines removed from 3949111147Snectar libcrypto, it's only intended to remove all the function name and 3950111147Snectar reason texts, thereby removing some of the footprint that may not 3951111147Snectar be interesting if those errors aren't displayed anyway. 3952111147Snectar 3953111147Snectar NOTE: it's still possible for any application or module to have it's 3954111147Snectar own set of error texts inserted. The routines are there, just not 3955111147Snectar used by default when no-err is given. 3956111147Snectar [Richard Levitte] 3957111147Snectar 3958111147Snectar *) Add support for FreeBSD on IA64. 3959111147Snectar [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 3960111147Snectar 3961111147Snectar *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 3962111147Snectar Kerberos function mit_des_cbc_cksum(). Before this change, 3963111147Snectar the value returned by DES_cbc_cksum() was like the one from 3964111147Snectar mit_des_cbc_cksum(), except the bytes were swapped. 3965111147Snectar [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 3966111147Snectar 3967111147Snectar *) Allow an application to disable the automatic SSL chain building. 3968111147Snectar Before this a rather primitive chain build was always performed in 3969111147Snectar ssl3_output_cert_chain(): an application had no way to send the 3970111147Snectar correct chain if the automatic operation produced an incorrect result. 3971111147Snectar 3972111147Snectar Now the chain builder is disabled if either: 3973111147Snectar 3974111147Snectar 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 3975111147Snectar 3976111147Snectar 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 3977111147Snectar 3978111147Snectar The reasoning behind this is that an application would not want the 3979111147Snectar auto chain building to take place if extra chain certificates are 3980111147Snectar present and it might also want a means of sending no additional 3981111147Snectar certificates (for example the chain has two certificates and the 3982111147Snectar root is omitted). 3983111147Snectar [Steve Henson] 3984111147Snectar 3985111147Snectar *) Add the possibility to build without the ENGINE framework. 3986111147Snectar [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 3987111147Snectar 3988111147Snectar *) Under Win32 gmtime() can return NULL: check return value in 3989111147Snectar OPENSSL_gmtime(). Add error code for case where gmtime() fails. 3990111147Snectar [Steve Henson] 3991111147Snectar 3992111147Snectar *) DSA routines: under certain error conditions uninitialized BN objects 3993111147Snectar could be freed. Solution: make sure initialization is performed early 3994111147Snectar enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 3995111147Snectar Nils Larsch <nla@trustcenter.de> via PR#459) 3996111147Snectar [Lutz Jaenicke] 3997111147Snectar 3998111147Snectar *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 3999111147Snectar checked on reconnect on the client side, therefore session resumption 4000111147Snectar could still fail with a "ssl session id is different" error. This 4001111147Snectar behaviour is masked when SSL_OP_ALL is used due to 4002111147Snectar SSL_OP_MICROSOFT_SESS_ID_BUG being set. 4003111147Snectar Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4004111147Snectar followup to PR #377. 4005111147Snectar [Lutz Jaenicke] 4006111147Snectar 4007111147Snectar *) IA-32 assembler support enhancements: unified ELF targets, support 4008111147Snectar for SCO/Caldera platforms, fix for Cygwin shared build. 4009111147Snectar [Andy Polyakov] 4010111147Snectar 4011111147Snectar *) Add support for FreeBSD on sparc64. As a consequence, support for 4012111147Snectar FreeBSD on non-x86 processors is separate from x86 processors on 4013111147Snectar the config script, much like the NetBSD support. 4014111147Snectar [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 4015111147Snectar 4016109998Smarkm Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 4017109998Smarkm 4018120631Snectar [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 4019120631Snectar OpenSSL 0.9.7.] 4020120631Snectar 4021109998Smarkm *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 4022109998Smarkm code (06) was taken as the first octet of the session ID and the last 4023109998Smarkm octet was ignored consequently. As a result SSLv2 client side session 4024109998Smarkm caching could not have worked due to the session ID mismatch between 4025109998Smarkm client and server. 4026109998Smarkm Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4027109998Smarkm PR #377. 4028109998Smarkm [Lutz Jaenicke] 4029109998Smarkm 4030109998Smarkm *) Change the declaration of needed Kerberos libraries to use EX_LIBS 4031109998Smarkm instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 4032109998Smarkm removed entirely. 4033109998Smarkm [Richard Levitte] 4034109998Smarkm 4035109998Smarkm *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 4036109998Smarkm seems that in spite of existing for more than a year, many application 4037109998Smarkm author have done nothing to provide the necessary callbacks, which 4038109998Smarkm means that this particular engine will not work properly anywhere. 4039109998Smarkm This is a very unfortunate situation which forces us, in the name 4040109998Smarkm of usability, to give the hw_ncipher.c a static lock, which is part 4041109998Smarkm of libcrypto. 4042109998Smarkm NOTE: This is for the 0.9.7 series ONLY. This hack will never 4043109998Smarkm appear in 0.9.8 or later. We EXPECT application authors to have 4044109998Smarkm dealt properly with this when 0.9.8 is released (unless we actually 4045109998Smarkm make such changes in the libcrypto locking code that changes will 4046109998Smarkm have to be made anyway). 4047109998Smarkm [Richard Levitte] 4048109998Smarkm 4049109998Smarkm *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 4050109998Smarkm octets have been read, EOF or an error occurs. Without this change 4051109998Smarkm some truncated ASN1 structures will not produce an error. 4052109998Smarkm [Steve Henson] 4053109998Smarkm 4054109998Smarkm *) Disable Heimdal support, since it hasn't been fully implemented. 4055109998Smarkm Still give the possibility to force the use of Heimdal, but with 4056109998Smarkm warnings and a request that patches get sent to openssl-dev. 4057109998Smarkm [Richard Levitte] 4058109998Smarkm 4059109998Smarkm *) Add the VC-CE target, introduce the WINCE sysname, and add 4060109998Smarkm INSTALL.WCE and appropriate conditionals to make it build. 4061109998Smarkm [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 4062109998Smarkm 4063109998Smarkm *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 4064109998Smarkm cygssl-x.y.z.dll, where x, y and z are the major, minor and 4065109998Smarkm edit numbers of the version. 4066109998Smarkm [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 4067109998Smarkm 4068109998Smarkm *) Introduce safe string copy and catenation functions 4069109998Smarkm (BUF_strlcpy() and BUF_strlcat()). 4070109998Smarkm [Ben Laurie (CHATS) and Richard Levitte] 4071109998Smarkm 4072109998Smarkm *) Avoid using fixed-size buffers for one-line DNs. 4073109998Smarkm [Ben Laurie (CHATS)] 4074109998Smarkm 4075109998Smarkm *) Add BUF_MEM_grow_clean() to avoid information leakage when 4076109998Smarkm resizing buffers containing secrets, and use where appropriate. 4077109998Smarkm [Ben Laurie (CHATS)] 4078109998Smarkm 4079109998Smarkm *) Avoid using fixed size buffers for configuration file location. 4080109998Smarkm [Ben Laurie (CHATS)] 4081109998Smarkm 4082109998Smarkm *) Avoid filename truncation for various CA files. 4083109998Smarkm [Ben Laurie (CHATS)] 4084109998Smarkm 4085109998Smarkm *) Use sizeof in preference to magic numbers. 4086109998Smarkm [Ben Laurie (CHATS)] 4087109998Smarkm 4088109998Smarkm *) Avoid filename truncation in cert requests. 4089109998Smarkm [Ben Laurie (CHATS)] 4090109998Smarkm 4091109998Smarkm *) Add assertions to check for (supposedly impossible) buffer 4092109998Smarkm overflows. 4093109998Smarkm [Ben Laurie (CHATS)] 4094109998Smarkm 4095109998Smarkm *) Don't cache truncated DNS entries in the local cache (this could 4096109998Smarkm potentially lead to a spoofing attack). 4097109998Smarkm [Ben Laurie (CHATS)] 4098109998Smarkm 4099109998Smarkm *) Fix various buffers to be large enough for hex/decimal 4100109998Smarkm representations in a platform independent manner. 4101109998Smarkm [Ben Laurie (CHATS)] 4102109998Smarkm 4103109998Smarkm *) Add CRYPTO_realloc_clean() to avoid information leakage when 4104109998Smarkm resizing buffers containing secrets, and use where appropriate. 4105109998Smarkm [Ben Laurie (CHATS)] 4106109998Smarkm 4107109998Smarkm *) Add BIO_indent() to avoid much slightly worrying code to do 4108109998Smarkm indents. 4109109998Smarkm [Ben Laurie (CHATS)] 4110109998Smarkm 4111109998Smarkm *) Convert sprintf()/BIO_puts() to BIO_printf(). 4112109998Smarkm [Ben Laurie (CHATS)] 4113109998Smarkm 4114109998Smarkm *) buffer_gets() could terminate with the buffer only half 4115109998Smarkm full. Fixed. 4116109998Smarkm [Ben Laurie (CHATS)] 4117109998Smarkm 4118109998Smarkm *) Add assertions to prevent user-supplied crypto functions from 4119109998Smarkm overflowing internal buffers by having large block sizes, etc. 4120109998Smarkm [Ben Laurie (CHATS)] 4121109998Smarkm 4122109998Smarkm *) New OPENSSL_assert() macro (similar to assert(), but enabled 4123109998Smarkm unconditionally). 4124109998Smarkm [Ben Laurie (CHATS)] 4125109998Smarkm 4126109998Smarkm *) Eliminate unused copy of key in RC4. 4127109998Smarkm [Ben Laurie (CHATS)] 4128109998Smarkm 4129109998Smarkm *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 4130109998Smarkm [Ben Laurie (CHATS)] 4131109998Smarkm 4132109998Smarkm *) Fix off-by-one error in EGD path. 4133109998Smarkm [Ben Laurie (CHATS)] 4134109998Smarkm 4135109998Smarkm *) If RANDFILE path is too long, ignore instead of truncating. 4136109998Smarkm [Ben Laurie (CHATS)] 4137109998Smarkm 4138109998Smarkm *) Eliminate unused and incorrectly sized X.509 structure 4139109998Smarkm CBCParameter. 4140109998Smarkm [Ben Laurie (CHATS)] 4141109998Smarkm 4142109998Smarkm *) Eliminate unused and dangerous function knumber(). 4143109998Smarkm [Ben Laurie (CHATS)] 4144109998Smarkm 4145109998Smarkm *) Eliminate unused and dangerous structure, KSSL_ERR. 4146109998Smarkm [Ben Laurie (CHATS)] 4147109998Smarkm 4148109998Smarkm *) Protect against overlong session ID context length in an encoded 4149109998Smarkm session object. Since these are local, this does not appear to be 4150109998Smarkm exploitable. 4151109998Smarkm [Ben Laurie (CHATS)] 4152109998Smarkm 4153109998Smarkm *) Change from security patch (see 0.9.6e below) that did not affect 4154109998Smarkm the 0.9.6 release series: 4155109998Smarkm 4156109998Smarkm Remote buffer overflow in SSL3 protocol - an attacker could 4157109998Smarkm supply an oversized master key in Kerberos-enabled versions. 4158160814Ssimon (CVE-2002-0657) 4159109998Smarkm [Ben Laurie (CHATS)] 4160109998Smarkm 4161109998Smarkm *) Change the SSL kerb5 codes to match RFC 2712. 4162109998Smarkm [Richard Levitte] 4163109998Smarkm 4164109998Smarkm *) Make -nameopt work fully for req and add -reqopt switch. 4165109998Smarkm [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 4166109998Smarkm 4167109998Smarkm *) The "block size" for block ciphers in CFB and OFB mode should be 1. 4168109998Smarkm [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 4169109998Smarkm 4170109998Smarkm *) Make sure tests can be performed even if the corresponding algorithms 4171109998Smarkm have been removed entirely. This was also the last step to make 4172109998Smarkm OpenSSL compilable with DJGPP under all reasonable conditions. 4173109998Smarkm [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 4174109998Smarkm 4175109998Smarkm *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 4176109998Smarkm to allow version independent disabling of normally unselected ciphers, 4177109998Smarkm which may be activated as a side-effect of selecting a single cipher. 4178109998Smarkm 4179109998Smarkm (E.g., cipher list string "RSA" enables ciphersuites that are left 4180109998Smarkm out of "ALL" because they do not provide symmetric encryption. 4181109998Smarkm "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 4182109998Smarkm [Lutz Jaenicke, Bodo Moeller] 4183109998Smarkm 4184109998Smarkm *) Add appropriate support for separate platform-dependent build 4185109998Smarkm directories. The recommended way to make a platform-dependent 4186109998Smarkm build directory is the following (tested on Linux), maybe with 4187109998Smarkm some local tweaks: 4188109998Smarkm 4189109998Smarkm # Place yourself outside of the OpenSSL source tree. In 4190109998Smarkm # this example, the environment variable OPENSSL_SOURCE 4191109998Smarkm # is assumed to contain the absolute OpenSSL source directory. 4192109998Smarkm mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 4193109998Smarkm cd objtree/"`uname -s`-`uname -r`-`uname -m`" 4194111147Snectar (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 4195109998Smarkm mkdir -p `dirname $F` 4196109998Smarkm ln -s $OPENSSL_SOURCE/$F $F 4197109998Smarkm done 4198109998Smarkm 4199109998Smarkm To be absolutely sure not to disturb the source tree, a "make clean" 4200109998Smarkm is a good thing. If it isn't successfull, don't worry about it, 4201109998Smarkm it probably means the source directory is very clean. 4202109998Smarkm [Richard Levitte] 4203109998Smarkm 4204109998Smarkm *) Make sure any ENGINE control commands make local copies of string 4205109998Smarkm pointers passed to them whenever necessary. Otherwise it is possible 4206109998Smarkm the caller may have overwritten (or deallocated) the original string 4207109998Smarkm data when a later ENGINE operation tries to use the stored values. 4208109998Smarkm [G�tz Babin-Ebell <babinebell@trustcenter.de>] 4209109998Smarkm 4210109998Smarkm *) Improve diagnostics in file reading and command-line digests. 4211109998Smarkm [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 4212109998Smarkm 4213109998Smarkm *) Add AES modes CFB and OFB to the object database. Correct an 4214109998Smarkm error in AES-CFB decryption. 4215109998Smarkm [Richard Levitte] 4216109998Smarkm 4217109998Smarkm *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 4218109998Smarkm allows existing EVP_CIPHER_CTX structures to be reused after 4219109998Smarkm calling EVP_*Final(). This behaviour is used by encryption 4220109998Smarkm BIOs and some applications. This has the side effect that 4221109998Smarkm applications must explicitly clean up cipher contexts with 4222109998Smarkm EVP_CIPHER_CTX_cleanup() or they will leak memory. 4223109998Smarkm [Steve Henson] 4224109998Smarkm 4225109998Smarkm *) Check the values of dna and dnb in bn_mul_recursive before calling 4226109998Smarkm bn_mul_comba (a non zero value means the a or b arrays do not contain 4227109998Smarkm n2 elements) and fallback to bn_mul_normal if either is not zero. 4228109998Smarkm [Steve Henson] 4229109998Smarkm 4230109998Smarkm *) Fix escaping of non-ASCII characters when using the -subj option 4231109998Smarkm of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 4232109998Smarkm [Lutz Jaenicke] 4233109998Smarkm 4234109998Smarkm *) Make object definitions compliant to LDAP (RFC2256): SN is the short 4235109998Smarkm form for "surname", serialNumber has no short form. 4236109998Smarkm Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 4237109998Smarkm therefore remove "mail" short name for "internet 7". 4238109998Smarkm The OID for unique identifiers in X509 certificates is 4239109998Smarkm x500UniqueIdentifier, not uniqueIdentifier. 4240109998Smarkm Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 4241109998Smarkm [Lutz Jaenicke] 4242109998Smarkm 4243109998Smarkm *) Add an "init" command to the ENGINE config module and auto initialize 4244109998Smarkm ENGINEs. Without any "init" command the ENGINE will be initialized 4245109998Smarkm after all ctrl commands have been executed on it. If init=1 the 4246109998Smarkm ENGINE is initailized at that point (ctrls before that point are run 4247109998Smarkm on the uninitialized ENGINE and after on the initialized one). If 4248109998Smarkm init=0 then the ENGINE will not be iniatialized at all. 4249109998Smarkm [Steve Henson] 4250109998Smarkm 4251109998Smarkm *) Fix the 'app_verify_callback' interface so that the user-defined 4252109998Smarkm argument is actually passed to the callback: In the 4253109998Smarkm SSL_CTX_set_cert_verify_callback() prototype, the callback 4254109998Smarkm declaration has been changed from 4255109998Smarkm int (*cb)() 4256109998Smarkm into 4257109998Smarkm int (*cb)(X509_STORE_CTX *,void *); 4258109998Smarkm in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 4259109998Smarkm i=s->ctx->app_verify_callback(&ctx) 4260109998Smarkm has been changed into 4261109998Smarkm i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 4262109998Smarkm 4263109998Smarkm To update applications using SSL_CTX_set_cert_verify_callback(), 4264109998Smarkm a dummy argument can be added to their callback functions. 4265109998Smarkm [D. K. Smetters <smetters@parc.xerox.com>] 4266109998Smarkm 4267109998Smarkm *) Added the '4758cca' ENGINE to support IBM 4758 cards. 4268109998Smarkm [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 4269109998Smarkm 4270109998Smarkm *) Add and OPENSSL_LOAD_CONF define which will cause 4271109998Smarkm OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 4272109998Smarkm This allows older applications to transparently support certain 4273109998Smarkm OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 4274109998Smarkm Two new functions OPENSSL_add_all_algorithms_noconf() which will never 4275109998Smarkm load the config file and OPENSSL_add_all_algorithms_conf() which will 4276109998Smarkm always load it have also been added. 4277109998Smarkm [Steve Henson] 4278109998Smarkm 4279109998Smarkm *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 4280109998Smarkm Adjust NIDs and EVP layer. 4281109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4282109998Smarkm 4283109998Smarkm *) Config modules support in openssl utility. 4284109998Smarkm 4285109998Smarkm Most commands now load modules from the config file, 4286109998Smarkm though in a few (such as version) this isn't done 4287109998Smarkm because it couldn't be used for anything. 4288109998Smarkm 4289109998Smarkm In the case of ca and req the config file used is 4290109998Smarkm the same as the utility itself: that is the -config 4291109998Smarkm command line option can be used to specify an 4292109998Smarkm alternative file. 4293109998Smarkm [Steve Henson] 4294109998Smarkm 4295109998Smarkm *) Move default behaviour from OPENSSL_config(). If appname is NULL 4296109998Smarkm use "openssl_conf" if filename is NULL use default openssl config file. 4297109998Smarkm [Steve Henson] 4298109998Smarkm 4299109998Smarkm *) Add an argument to OPENSSL_config() to allow the use of an alternative 4300109998Smarkm config section name. Add a new flag to tolerate a missing config file 4301109998Smarkm and move code to CONF_modules_load_file(). 4302109998Smarkm [Steve Henson] 4303109998Smarkm 4304109998Smarkm *) Support for crypto accelerator cards from Accelerated Encryption 4305109998Smarkm Processing, www.aep.ie. (Use engine 'aep') 4306109998Smarkm The support was copied from 0.9.6c [engine] and adapted/corrected 4307109998Smarkm to work with the new engine framework. 4308109998Smarkm [AEP Inc. and Richard Levitte] 4309109998Smarkm 4310109998Smarkm *) Support for SureWare crypto accelerator cards from Baltimore 4311109998Smarkm Technologies. (Use engine 'sureware') 4312109998Smarkm The support was copied from 0.9.6c [engine] and adapted 4313109998Smarkm to work with the new engine framework. 4314109998Smarkm [Richard Levitte] 4315109998Smarkm 4316109998Smarkm *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 4317109998Smarkm make the newer ENGINE framework commands for the CHIL engine work. 4318109998Smarkm [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 4319109998Smarkm 4320109998Smarkm *) Make it possible to produce shared libraries on ReliantUNIX. 4321109998Smarkm [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 4322109998Smarkm 4323109998Smarkm *) Add the configuration target debug-linux-ppro. 4324109998Smarkm Make 'openssl rsa' use the general key loading routines 4325109998Smarkm implemented in apps.c, and make those routines able to 4326109998Smarkm handle the key format FORMAT_NETSCAPE and the variant 4327109998Smarkm FORMAT_IISSGC. 4328109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4329109998Smarkm 4330109998Smarkm *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 4331109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4332109998Smarkm 4333109998Smarkm *) Add -keyform to rsautl, and document -engine. 4334109998Smarkm [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 4335109998Smarkm 4336109998Smarkm *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 4337109998Smarkm BIO_R_NO_SUCH_FILE error code rather than the generic 4338109998Smarkm ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 4339109998Smarkm [Ben Laurie] 4340109998Smarkm 4341109998Smarkm *) Add new functions 4342109998Smarkm ERR_peek_last_error 4343109998Smarkm ERR_peek_last_error_line 4344109998Smarkm ERR_peek_last_error_line_data. 4345109998Smarkm These are similar to 4346109998Smarkm ERR_peek_error 4347109998Smarkm ERR_peek_error_line 4348109998Smarkm ERR_peek_error_line_data, 4349109998Smarkm but report on the latest error recorded rather than the first one 4350109998Smarkm still in the error queue. 4351109998Smarkm [Ben Laurie, Bodo Moeller] 4352109998Smarkm 4353109998Smarkm *) default_algorithms option in ENGINE config module. This allows things 4354109998Smarkm like: 4355109998Smarkm default_algorithms = ALL 4356109998Smarkm default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 4357109998Smarkm [Steve Henson] 4358109998Smarkm 4359109998Smarkm *) Prelminary ENGINE config module. 4360109998Smarkm [Steve Henson] 4361109998Smarkm 4362109998Smarkm *) New experimental application configuration code. 4363109998Smarkm [Steve Henson] 4364109998Smarkm 4365109998Smarkm *) Change the AES code to follow the same name structure as all other 4366109998Smarkm symmetric ciphers, and behave the same way. Move everything to 4367109998Smarkm the directory crypto/aes, thereby obsoleting crypto/rijndael. 4368109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4369109998Smarkm 4370109998Smarkm *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 4371109998Smarkm [Ben Laurie and Theo de Raadt] 4372109998Smarkm 4373109998Smarkm *) Add option to output public keys in req command. 4374109998Smarkm [Massimiliano Pala madwolf@openca.org] 4375109998Smarkm 4376109998Smarkm *) Use wNAFs in EC_POINTs_mul() for improved efficiency 4377109998Smarkm (up to about 10% better than before for P-192 and P-224). 4378109998Smarkm [Bodo Moeller] 4379109998Smarkm 4380109998Smarkm *) New functions/macros 4381109998Smarkm 4382109998Smarkm SSL_CTX_set_msg_callback(ctx, cb) 4383109998Smarkm SSL_CTX_set_msg_callback_arg(ctx, arg) 4384109998Smarkm SSL_set_msg_callback(ssl, cb) 4385109998Smarkm SSL_set_msg_callback_arg(ssl, arg) 4386109998Smarkm 4387109998Smarkm to request calling a callback function 4388109998Smarkm 4389109998Smarkm void cb(int write_p, int version, int content_type, 4390109998Smarkm const void *buf, size_t len, SSL *ssl, void *arg) 4391109998Smarkm 4392109998Smarkm whenever a protocol message has been completely received 4393109998Smarkm (write_p == 0) or sent (write_p == 1). Here 'version' is the 4394109998Smarkm protocol version according to which the SSL library interprets 4395109998Smarkm the current protocol message (SSL2_VERSION, SSL3_VERSION, or 4396109998Smarkm TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 4397109998Smarkm the content type as defined in the SSL 3.0/TLS 1.0 protocol 4398109998Smarkm specification (change_cipher_spec(20), alert(21), handshake(22)). 4399109998Smarkm 'buf' and 'len' point to the actual message, 'ssl' to the 4400109998Smarkm SSL object, and 'arg' is the application-defined value set by 4401109998Smarkm SSL[_CTX]_set_msg_callback_arg(). 4402109998Smarkm 4403109998Smarkm 'openssl s_client' and 'openssl s_server' have new '-msg' options 4404109998Smarkm to enable a callback that displays all protocol messages. 4405109998Smarkm [Bodo Moeller] 4406109998Smarkm 4407109998Smarkm *) Change the shared library support so shared libraries are built as 4408109998Smarkm soon as the corresponding static library is finished, and thereby get 4409109998Smarkm openssl and the test programs linked against the shared library. 4410109998Smarkm This still only happens when the keyword "shard" has been given to 4411109998Smarkm the configuration scripts. 4412109998Smarkm 4413109998Smarkm NOTE: shared library support is still an experimental thing, and 4414109998Smarkm backward binary compatibility is still not guaranteed. 4415109998Smarkm ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 4416109998Smarkm 4417109998Smarkm *) Add support for Subject Information Access extension. 4418109998Smarkm [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 4419109998Smarkm 4420109998Smarkm *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 4421109998Smarkm additional bytes when new memory had to be allocated, not just 4422109998Smarkm when reusing an existing buffer. 4423109998Smarkm [Bodo Moeller] 4424109998Smarkm 4425109998Smarkm *) New command line and configuration option 'utf8' for the req command. 4426109998Smarkm This allows field values to be specified as UTF8 strings. 4427109998Smarkm [Steve Henson] 4428109998Smarkm 4429109998Smarkm *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 4430109998Smarkm runs for the former and machine-readable output for the latter. 4431109998Smarkm [Ben Laurie] 4432109998Smarkm 4433109998Smarkm *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 4434109998Smarkm of the e-mail address in the DN (i.e., it will go into a certificate 4435109998Smarkm extension only). The new configuration file option 'email_in_dn = no' 4436109998Smarkm has the same effect. 4437109998Smarkm [Massimiliano Pala madwolf@openca.org] 4438109998Smarkm 4439109998Smarkm *) Change all functions with names starting with des_ to be starting 4440109998Smarkm with DES_ instead. Add wrappers that are compatible with libdes, 4441109998Smarkm but are named _ossl_old_des_*. Finally, add macros that map the 4442109998Smarkm des_* symbols to the corresponding _ossl_old_des_* if libdes 4443109998Smarkm compatibility is desired. If OpenSSL 0.9.6c compatibility is 4444109998Smarkm desired, the des_* symbols will be mapped to DES_*, with one 4445109998Smarkm exception. 4446109998Smarkm 4447109998Smarkm Since we provide two compatibility mappings, the user needs to 4448109998Smarkm define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 4449109998Smarkm compatibility is desired. The default (i.e., when that macro 4450109998Smarkm isn't defined) is OpenSSL 0.9.6c compatibility. 4451109998Smarkm 4452109998Smarkm There are also macros that enable and disable the support of old 4453109998Smarkm des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 4454109998Smarkm and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 4455109998Smarkm are defined, the default will apply: to support the old des routines. 4456109998Smarkm 4457109998Smarkm In either case, one must include openssl/des.h to get the correct 4458109998Smarkm definitions. Do not try to just include openssl/des_old.h, that 4459109998Smarkm won't work. 4460109998Smarkm 4461109998Smarkm NOTE: This is a major break of an old API into a new one. Software 4462109998Smarkm authors are encouraged to switch to the DES_ style functions. Some 4463109998Smarkm time in the future, des_old.h and the libdes compatibility functions 4464109998Smarkm will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 4465109998Smarkm default), and then completely removed. 4466109998Smarkm [Richard Levitte] 4467109998Smarkm 4468109998Smarkm *) Test for certificates which contain unsupported critical extensions. 4469109998Smarkm If such a certificate is found during a verify operation it is 4470109998Smarkm rejected by default: this behaviour can be overridden by either 4471109998Smarkm handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 4472109998Smarkm by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 4473109998Smarkm X509_supported_extension() has also been added which returns 1 if a 4474109998Smarkm particular extension is supported. 4475109998Smarkm [Steve Henson] 4476109998Smarkm 4477109998Smarkm *) Modify the behaviour of EVP cipher functions in similar way to digests 4478109998Smarkm to retain compatibility with existing code. 4479109998Smarkm [Steve Henson] 4480109998Smarkm 4481109998Smarkm *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 4482109998Smarkm compatibility with existing code. In particular the 'ctx' parameter does 4483109998Smarkm not have to be to be initialized before the call to EVP_DigestInit() and 4484109998Smarkm it is tidied up after a call to EVP_DigestFinal(). New function 4485109998Smarkm EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 4486109998Smarkm EVP_MD_CTX_copy() changed to not require the destination to be 4487109998Smarkm initialized valid and new function EVP_MD_CTX_copy_ex() added which 4488109998Smarkm requires the destination to be valid. 4489109998Smarkm 4490109998Smarkm Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 4491109998Smarkm EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 4492109998Smarkm [Steve Henson] 4493109998Smarkm 4494109998Smarkm *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 4495109998Smarkm so that complete 'Handshake' protocol structures are kept in memory 4496109998Smarkm instead of overwriting 'msg_type' and 'length' with 'body' data. 4497109998Smarkm [Bodo Moeller] 4498109998Smarkm 4499109998Smarkm *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 4500109998Smarkm [Massimo Santin via Richard Levitte] 4501109998Smarkm 4502109998Smarkm *) Major restructuring to the underlying ENGINE code. This includes 4503109998Smarkm reduction of linker bloat, separation of pure "ENGINE" manipulation 4504109998Smarkm (initialisation, etc) from functionality dealing with implementations 4505109998Smarkm of specific crypto iterfaces. This change also introduces integrated 4506109998Smarkm support for symmetric ciphers and digest implementations - so ENGINEs 4507109998Smarkm can now accelerate these by providing EVP_CIPHER and EVP_MD 4508109998Smarkm implementations of their own. This is detailed in crypto/engine/README 4509109998Smarkm as it couldn't be adequately described here. However, there are a few 4510109998Smarkm API changes worth noting - some RSA, DSA, DH, and RAND functions that 4511109998Smarkm were changed in the original introduction of ENGINE code have now 4512109998Smarkm reverted back - the hooking from this code to ENGINE is now a good 4513109998Smarkm deal more passive and at run-time, operations deal directly with 4514109998Smarkm RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 4515109998Smarkm dereferencing through an ENGINE pointer any more. Also, the ENGINE 4516109998Smarkm functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 4517109998Smarkm they were not being used by the framework as there is no concept of a 4518109998Smarkm BIGNUM_METHOD and they could not be generalised to the new 4519109998Smarkm 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 4520109998Smarkm ENGINE_cpy() has been removed as it cannot be consistently defined in 4521109998Smarkm the new code. 4522109998Smarkm [Geoff Thorpe] 4523109998Smarkm 4524109998Smarkm *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 4525109998Smarkm [Steve Henson] 4526109998Smarkm 4527109998Smarkm *) Change mkdef.pl to sort symbols that get the same entry number, 4528109998Smarkm and make sure the automatically generated functions ERR_load_* 4529109998Smarkm become part of libeay.num as well. 4530109998Smarkm [Richard Levitte] 4531109998Smarkm 4532109998Smarkm *) New function SSL_renegotiate_pending(). This returns true once 4533109998Smarkm renegotiation has been requested (either SSL_renegotiate() call 4534109998Smarkm or HelloRequest/ClientHello receveived from the peer) and becomes 4535109998Smarkm false once a handshake has been completed. 4536109998Smarkm (For servers, SSL_renegotiate() followed by SSL_do_handshake() 4537109998Smarkm sends a HelloRequest, but does not ensure that a handshake takes 4538109998Smarkm place. SSL_renegotiate_pending() is useful for checking if the 4539109998Smarkm client has followed the request.) 4540109998Smarkm [Bodo Moeller] 4541109998Smarkm 4542109998Smarkm *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 4543109998Smarkm By default, clients may request session resumption even during 4544109998Smarkm renegotiation (if session ID contexts permit); with this option, 4545109998Smarkm session resumption is possible only in the first handshake. 4546109998Smarkm 4547109998Smarkm SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 4548109998Smarkm more bits available for options that should not be part of 4549109998Smarkm SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 4550109998Smarkm [Bodo Moeller] 4551109998Smarkm 4552109998Smarkm *) Add some demos for certificate and certificate request creation. 4553109998Smarkm [Steve Henson] 4554109998Smarkm 4555109998Smarkm *) Make maximum certificate chain size accepted from the peer application 4556109998Smarkm settable (SSL*_get/set_max_cert_list()), as proposed by 4557109998Smarkm "Douglas E. Engert" <deengert@anl.gov>. 4558109998Smarkm [Lutz Jaenicke] 4559109998Smarkm 4560109998Smarkm *) Add support for shared libraries for Unixware-7 4561109998Smarkm (Boyd Lynn Gerber <gerberb@zenez.com>). 4562109998Smarkm [Lutz Jaenicke] 4563109998Smarkm 4564109998Smarkm *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 4565109998Smarkm be done prior to destruction. Use this to unload error strings from 4566109998Smarkm ENGINEs that load their own error strings. NB: This adds two new API 4567109998Smarkm functions to "get" and "set" this destroy handler in an ENGINE. 4568109998Smarkm [Geoff Thorpe] 4569109998Smarkm 4570109998Smarkm *) Alter all existing ENGINE implementations (except "openssl" and 4571109998Smarkm "openbsd") to dynamically instantiate their own error strings. This 4572109998Smarkm makes them more flexible to be built both as statically-linked ENGINEs 4573109998Smarkm and self-contained shared-libraries loadable via the "dynamic" ENGINE. 4574109998Smarkm Also, add stub code to each that makes building them as self-contained 4575109998Smarkm shared-libraries easier (see README.ENGINE). 4576109998Smarkm [Geoff Thorpe] 4577109998Smarkm 4578109998Smarkm *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 4579109998Smarkm implementations into applications that are completely implemented in 4580109998Smarkm self-contained shared-libraries. The "dynamic" ENGINE exposes control 4581109998Smarkm commands that can be used to configure what shared-library to load and 4582109998Smarkm to control aspects of the way it is handled. Also, made an update to 4583109998Smarkm the README.ENGINE file that brings its information up-to-date and 4584109998Smarkm provides some information and instructions on the "dynamic" ENGINE 4585109998Smarkm (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 4586109998Smarkm [Geoff Thorpe] 4587109998Smarkm 4588109998Smarkm *) Make it possible to unload ranges of ERR strings with a new 4589109998Smarkm "ERR_unload_strings" function. 4590109998Smarkm [Geoff Thorpe] 4591109998Smarkm 4592109998Smarkm *) Add a copy() function to EVP_MD. 4593109998Smarkm [Ben Laurie] 4594109998Smarkm 4595109998Smarkm *) Make EVP_MD routines take a context pointer instead of just the 4596109998Smarkm md_data void pointer. 4597109998Smarkm [Ben Laurie] 4598109998Smarkm 4599109998Smarkm *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 4600109998Smarkm that the digest can only process a single chunk of data 4601109998Smarkm (typically because it is provided by a piece of 4602109998Smarkm hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 4603109998Smarkm is only going to provide a single chunk of data, and hence the 4604109998Smarkm framework needn't accumulate the data for oneshot drivers. 4605109998Smarkm [Ben Laurie] 4606109998Smarkm 4607109998Smarkm *) As with "ERR", make it possible to replace the underlying "ex_data" 4608109998Smarkm functions. This change also alters the storage and management of global 4609109998Smarkm ex_data state - it's now all inside ex_data.c and all "class" code (eg. 4610109998Smarkm RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 4611109998Smarkm index counters. The API functions that use this state have been changed 4612109998Smarkm to take a "class_index" rather than pointers to the class's local STACK 4613109998Smarkm and counter, and there is now an API function to dynamically create new 4614109998Smarkm classes. This centralisation allows us to (a) plug a lot of the 4615109998Smarkm thread-safety problems that existed, and (b) makes it possible to clean 4616109998Smarkm up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 4617109998Smarkm such data would previously have always leaked in application code and 4618109998Smarkm workarounds were in place to make the memory debugging turn a blind eye 4619109998Smarkm to it. Application code that doesn't use this new function will still 4620109998Smarkm leak as before, but their memory debugging output will announce it now 4621109998Smarkm rather than letting it slide. 4622109998Smarkm 4623109998Smarkm Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 4624109998Smarkm induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 4625109998Smarkm has a return value to indicate success or failure. 4626109998Smarkm [Geoff Thorpe] 4627109998Smarkm 4628109998Smarkm *) Make it possible to replace the underlying "ERR" functions such that the 4629109998Smarkm global state (2 LHASH tables and 2 locks) is only used by the "default" 4630109998Smarkm implementation. This change also adds two functions to "get" and "set" 4631109998Smarkm the implementation prior to it being automatically set the first time 4632109998Smarkm any other ERR function takes place. Ie. an application can call "get", 4633109998Smarkm pass the return value to a module it has just loaded, and that module 4634109998Smarkm can call its own "set" function using that value. This means the 4635109998Smarkm module's "ERR" operations will use (and modify) the error state in the 4636109998Smarkm application and not in its own statically linked copy of OpenSSL code. 4637109998Smarkm [Geoff Thorpe] 4638109998Smarkm 4639109998Smarkm *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 4640109998Smarkm reference counts. This performs normal REF_PRINT/REF_CHECK macros on 4641109998Smarkm the operation, and provides a more encapsulated way for external code 4642109998Smarkm (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 4643109998Smarkm to use these functions rather than manually incrementing the counts. 4644109998Smarkm 4645109998Smarkm Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 4646109998Smarkm [Geoff Thorpe] 4647109998Smarkm 4648109998Smarkm *) Add EVP test program. 4649109998Smarkm [Ben Laurie] 4650109998Smarkm 4651109998Smarkm *) Add symmetric cipher support to ENGINE. Expect the API to change! 4652109998Smarkm [Ben Laurie] 4653109998Smarkm 4654109998Smarkm *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 4655109998Smarkm X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 4656109998Smarkm X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 4657109998Smarkm These allow a CRL to be built without having to access X509_CRL fields 4658109998Smarkm directly. Modify 'ca' application to use new functions. 4659109998Smarkm [Steve Henson] 4660109998Smarkm 4661109998Smarkm *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 4662109998Smarkm bug workarounds. Rollback attack detection is a security feature. 4663109998Smarkm The problem will only arise on OpenSSL servers when TLSv1 is not 4664109998Smarkm available (sslv3_server_method() or SSL_OP_NO_TLSv1). 4665109998Smarkm Software authors not wanting to support TLSv1 will have special reasons 4666109998Smarkm for their choice and can explicitly enable this option. 4667109998Smarkm [Bodo Moeller, Lutz Jaenicke] 4668109998Smarkm 4669109998Smarkm *) Rationalise EVP so it can be extended: don't include a union of 4670109998Smarkm cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 4671109998Smarkm (similar to those existing for EVP_CIPHER_CTX). 4672109998Smarkm Usage example: 4673109998Smarkm 4674109998Smarkm EVP_MD_CTX md; 4675109998Smarkm 4676109998Smarkm EVP_MD_CTX_init(&md); /* new function call */ 4677109998Smarkm EVP_DigestInit(&md, EVP_sha1()); 4678109998Smarkm EVP_DigestUpdate(&md, in, len); 4679109998Smarkm EVP_DigestFinal(&md, out, NULL); 4680109998Smarkm EVP_MD_CTX_cleanup(&md); /* new function call */ 4681109998Smarkm 4682109998Smarkm [Ben Laurie] 4683109998Smarkm 4684109998Smarkm *) Make DES key schedule conform to the usual scheme, as well as 4685109998Smarkm correcting its structure. This means that calls to DES functions 4686109998Smarkm now have to pass a pointer to a des_key_schedule instead of a 4687109998Smarkm plain des_key_schedule (which was actually always a pointer 4688109998Smarkm anyway): E.g., 4689109998Smarkm 4690109998Smarkm des_key_schedule ks; 4691109998Smarkm 4692109998Smarkm des_set_key_checked(..., &ks); 4693109998Smarkm des_ncbc_encrypt(..., &ks, ...); 4694109998Smarkm 4695109998Smarkm (Note that a later change renames 'des_...' into 'DES_...'.) 4696109998Smarkm [Ben Laurie] 4697109998Smarkm 4698109998Smarkm *) Initial reduction of linker bloat: the use of some functions, such as 4699109998Smarkm PEM causes large amounts of unused functions to be linked in due to 4700109998Smarkm poor organisation. For example pem_all.c contains every PEM function 4701109998Smarkm which has a knock on effect of linking in large amounts of (unused) 4702109998Smarkm ASN1 code. Grouping together similar functions and splitting unrelated 4703109998Smarkm functions prevents this. 4704109998Smarkm [Steve Henson] 4705109998Smarkm 4706109998Smarkm *) Cleanup of EVP macros. 4707109998Smarkm [Ben Laurie] 4708109998Smarkm 4709109998Smarkm *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 4710109998Smarkm correct _ecb suffix. 4711109998Smarkm [Ben Laurie] 4712109998Smarkm 4713109998Smarkm *) Add initial OCSP responder support to ocsp application. The 4714109998Smarkm revocation information is handled using the text based index 4715109998Smarkm use by the ca application. The responder can either handle 4716109998Smarkm requests generated internally, supplied in files (for example 4717109998Smarkm via a CGI script) or using an internal minimal server. 4718109998Smarkm [Steve Henson] 4719109998Smarkm 4720109998Smarkm *) Add configuration choices to get zlib compression for TLS. 4721109998Smarkm [Richard Levitte] 4722109998Smarkm 4723109998Smarkm *) Changes to Kerberos SSL for RFC 2712 compliance: 4724109998Smarkm 1. Implemented real KerberosWrapper, instead of just using 4725109998Smarkm KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 4726109998Smarkm 2. Implemented optional authenticator field of KerberosWrapper. 4727109998Smarkm 4728109998Smarkm Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 4729109998Smarkm and authenticator structs; see crypto/krb5/. 4730109998Smarkm 4731109998Smarkm Generalized Kerberos calls to support multiple Kerberos libraries. 4732109998Smarkm [Vern Staats <staatsvr@asc.hpc.mil>, 4733109998Smarkm Jeffrey Altman <jaltman@columbia.edu> 4734109998Smarkm via Richard Levitte] 4735109998Smarkm 4736109998Smarkm *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 4737109998Smarkm already does with RSA. testdsa.h now has 'priv_key/pub_key' 4738109998Smarkm values for each of the key sizes rather than having just 4739109998Smarkm parameters (and 'speed' generating keys each time). 4740109998Smarkm [Geoff Thorpe] 4741109998Smarkm 4742109998Smarkm *) Speed up EVP routines. 4743109998Smarkm Before: 4744109998Smarkmencrypt 4745109998Smarkmtype 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 4746109998Smarkmdes-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 4747109998Smarkmdes-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 4748109998Smarkmdes-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 4749109998Smarkmdecrypt 4750109998Smarkmdes-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 4751109998Smarkmdes-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 4752109998Smarkmdes-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 4753109998Smarkm After: 4754109998Smarkmencrypt 4755109998Smarkmdes-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 4756109998Smarkmdecrypt 4757109998Smarkmdes-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 4758109998Smarkm [Ben Laurie] 4759109998Smarkm 4760109998Smarkm *) Added the OS2-EMX target. 4761109998Smarkm ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 4762109998Smarkm 4763109998Smarkm *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 4764109998Smarkm to support NCONF routines in extension code. New function CONF_set_nconf() 4765109998Smarkm to allow functions which take an NCONF to also handle the old LHASH 4766109998Smarkm structure: this means that the old CONF compatible routines can be 4767109998Smarkm retained (in particular wrt extensions) without having to duplicate the 4768109998Smarkm code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 4769109998Smarkm [Steve Henson] 4770109998Smarkm 4771109998Smarkm *) Enhance the general user interface with mechanisms for inner control 4772109998Smarkm and with possibilities to have yes/no kind of prompts. 4773109998Smarkm [Richard Levitte] 4774109998Smarkm 4775109998Smarkm *) Change all calls to low level digest routines in the library and 4776109998Smarkm applications to use EVP. Add missing calls to HMAC_cleanup() and 4777109998Smarkm don't assume HMAC_CTX can be copied using memcpy(). 4778109998Smarkm [Verdon Walker <VWalker@novell.com>, Steve Henson] 4779109998Smarkm 4780109998Smarkm *) Add the possibility to control engines through control names but with 4781109998Smarkm arbitrary arguments instead of just a string. 4782109998Smarkm Change the key loaders to take a UI_METHOD instead of a callback 4783109998Smarkm function pointer. NOTE: this breaks binary compatibility with earlier 4784109998Smarkm versions of OpenSSL [engine]. 4785109998Smarkm Adapt the nCipher code for these new conditions and add a card insertion 4786109998Smarkm callback. 4787109998Smarkm [Richard Levitte] 4788109998Smarkm 4789109998Smarkm *) Enhance the general user interface with mechanisms to better support 4790109998Smarkm dialog box interfaces, application-defined prompts, the possibility 4791109998Smarkm to use defaults (for example default passwords from somewhere else) 4792109998Smarkm and interrupts/cancellations. 4793109998Smarkm [Richard Levitte] 4794109998Smarkm 4795109998Smarkm *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 4796109998Smarkm attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 4797109998Smarkm [Steve Henson] 4798109998Smarkm 4799109998Smarkm *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 4800109998Smarkm tidy up some unnecessarily weird code in 'sk_new()'). 4801109998Smarkm [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 4802109998Smarkm 4803109998Smarkm *) Change the key loading routines for ENGINEs to use the same kind 4804109998Smarkm callback (pem_password_cb) as all other routines that need this 4805109998Smarkm kind of callback. 4806109998Smarkm [Richard Levitte] 4807109998Smarkm 4808109998Smarkm *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 4809109998Smarkm 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 4810109998Smarkm than this minimum value is recommended. 4811109998Smarkm [Lutz Jaenicke] 4812109998Smarkm 4813109998Smarkm *) New random seeder for OpenVMS, using the system process statistics 4814109998Smarkm that are easily reachable. 4815109998Smarkm [Richard Levitte] 4816109998Smarkm 4817109998Smarkm *) Windows apparently can't transparently handle global 4818109998Smarkm variables defined in DLLs. Initialisations such as: 4819109998Smarkm 4820109998Smarkm const ASN1_ITEM *it = &ASN1_INTEGER_it; 4821109998Smarkm 4822109998Smarkm wont compile. This is used by the any applications that need to 4823109998Smarkm declare their own ASN1 modules. This was fixed by adding the option 4824109998Smarkm EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 4825109998Smarkm needed for static libraries under Win32. 4826109998Smarkm [Steve Henson] 4827109998Smarkm 4828109998Smarkm *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 4829109998Smarkm setting of purpose and trust fields. New X509_STORE trust and 4830109998Smarkm purpose functions and tidy up setting in other SSL functions. 4831109998Smarkm [Steve Henson] 4832109998Smarkm 4833109998Smarkm *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 4834109998Smarkm structure. These are inherited by X509_STORE_CTX when it is 4835109998Smarkm initialised. This allows various defaults to be set in the 4836109998Smarkm X509_STORE structure (such as flags for CRL checking and custom 4837109998Smarkm purpose or trust settings) for functions which only use X509_STORE_CTX 4838109998Smarkm internally such as S/MIME. 4839109998Smarkm 4840109998Smarkm Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 4841109998Smarkm trust settings if they are not set in X509_STORE. This allows X509_STORE 4842109998Smarkm purposes and trust (in S/MIME for example) to override any set by default. 4843109998Smarkm 4844109998Smarkm Add command line options for CRL checking to smime, s_client and s_server 4845109998Smarkm applications. 4846109998Smarkm [Steve Henson] 4847109998Smarkm 4848109998Smarkm *) Initial CRL based revocation checking. If the CRL checking flag(s) 4849109998Smarkm are set then the CRL is looked up in the X509_STORE structure and 4850109998Smarkm its validity and signature checked, then if the certificate is found 4851109998Smarkm in the CRL the verify fails with a revoked error. 4852109998Smarkm 4853109998Smarkm Various new CRL related callbacks added to X509_STORE_CTX structure. 4854109998Smarkm 4855109998Smarkm Command line options added to 'verify' application to support this. 4856109998Smarkm 4857109998Smarkm This needs some additional work, such as being able to handle multiple 4858109998Smarkm CRLs with different times, extension based lookup (rather than just 4859109998Smarkm by subject name) and ultimately more complete V2 CRL extension 4860109998Smarkm handling. 4861109998Smarkm [Steve Henson] 4862109998Smarkm 4863109998Smarkm *) Add a general user interface API (crypto/ui/). This is designed 4864109998Smarkm to replace things like des_read_password and friends (backward 4865109998Smarkm compatibility functions using this new API are provided). 4866109998Smarkm The purpose is to remove prompting functions from the DES code 4867109998Smarkm section as well as provide for prompting through dialog boxes in 4868109998Smarkm a window system and the like. 4869109998Smarkm [Richard Levitte] 4870109998Smarkm 4871109998Smarkm *) Add "ex_data" support to ENGINE so implementations can add state at a 4872109998Smarkm per-structure level rather than having to store it globally. 4873109998Smarkm [Geoff] 4874109998Smarkm 4875109998Smarkm *) Make it possible for ENGINE structures to be copied when retrieved by 4876109998Smarkm ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 4877109998Smarkm This causes the "original" ENGINE structure to act like a template, 4878109998Smarkm analogous to the RSA vs. RSA_METHOD type of separation. Because of this 4879109998Smarkm operational state can be localised to each ENGINE structure, despite the 4880109998Smarkm fact they all share the same "methods". New ENGINE structures returned in 4881109998Smarkm this case have no functional references and the return value is the single 4882109998Smarkm structural reference. This matches the single structural reference returned 4883109998Smarkm by ENGINE_by_id() normally, when it is incremented on the pre-existing 4884109998Smarkm ENGINE structure. 4885109998Smarkm [Geoff] 4886109998Smarkm 4887109998Smarkm *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 4888109998Smarkm needs to match any other type at all we need to manually clear the 4889109998Smarkm tag cache. 4890109998Smarkm [Steve Henson] 4891109998Smarkm 4892109998Smarkm *) Changes to the "openssl engine" utility to include; 4893109998Smarkm - verbosity levels ('-v', '-vv', and '-vvv') that provide information 4894109998Smarkm about an ENGINE's available control commands. 4895109998Smarkm - executing control commands from command line arguments using the 4896109998Smarkm '-pre' and '-post' switches. '-post' is only used if '-t' is 4897109998Smarkm specified and the ENGINE is successfully initialised. The syntax for 4898109998Smarkm the individual commands are colon-separated, for example; 4899109998Smarkm openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 4900109998Smarkm [Geoff] 4901109998Smarkm 4902109998Smarkm *) New dynamic control command support for ENGINEs. ENGINEs can now 4903109998Smarkm declare their own commands (numbers), names (strings), descriptions, 4904109998Smarkm and input types for run-time discovery by calling applications. A 4905109998Smarkm subset of these commands are implicitly classed as "executable" 4906109998Smarkm depending on their input type, and only these can be invoked through 4907109998Smarkm the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 4908109998Smarkm can be based on user input, config files, etc). The distinction is 4909109998Smarkm that "executable" commands cannot return anything other than a boolean 4910109998Smarkm result and can only support numeric or string input, whereas some 4911109998Smarkm discoverable commands may only be for direct use through 4912109998Smarkm ENGINE_ctrl(), eg. supporting the exchange of binary data, function 4913109998Smarkm pointers, or other custom uses. The "executable" commands are to 4914109998Smarkm support parameterisations of ENGINE behaviour that can be 4915109998Smarkm unambiguously defined by ENGINEs and used consistently across any 4916109998Smarkm OpenSSL-based application. Commands have been added to all the 4917109998Smarkm existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 4918109998Smarkm control over shared-library paths without source code alterations. 4919109998Smarkm [Geoff] 4920109998Smarkm 4921109998Smarkm *) Changed all ENGINE implementations to dynamically allocate their 4922109998Smarkm ENGINEs rather than declaring them statically. Apart from this being 4923109998Smarkm necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 4924109998Smarkm this also allows the implementations to compile without using the 4925109998Smarkm internal engine_int.h header. 4926109998Smarkm [Geoff] 4927109998Smarkm 4928109998Smarkm *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 4929109998Smarkm 'const' value. Any code that should be able to modify a RAND_METHOD 4930109998Smarkm should already have non-const pointers to it (ie. they should only 4931109998Smarkm modify their own ones). 4932109998Smarkm [Geoff] 4933109998Smarkm 4934109998Smarkm *) Made a variety of little tweaks to the ENGINE code. 4935109998Smarkm - "atalla" and "ubsec" string definitions were moved from header files 4936109998Smarkm to C code. "nuron" string definitions were placed in variables 4937109998Smarkm rather than hard-coded - allowing parameterisation of these values 4938109998Smarkm later on via ctrl() commands. 4939109998Smarkm - Removed unused "#if 0"'d code. 4940109998Smarkm - Fixed engine list iteration code so it uses ENGINE_free() to release 4941109998Smarkm structural references. 4942109998Smarkm - Constified the RAND_METHOD element of ENGINE structures. 4943109998Smarkm - Constified various get/set functions as appropriate and added 4944109998Smarkm missing functions (including a catch-all ENGINE_cpy that duplicates 4945109998Smarkm all ENGINE values onto a new ENGINE except reference counts/state). 4946109998Smarkm - Removed NULL parameter checks in get/set functions. Setting a method 4947109998Smarkm or function to NULL is a way of cancelling out a previously set 4948109998Smarkm value. Passing a NULL ENGINE parameter is just plain stupid anyway 4949109998Smarkm and doesn't justify the extra error symbols and code. 4950109998Smarkm - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 4951109998Smarkm flags from engine_int.h to engine.h. 4952109998Smarkm - Changed prototypes for ENGINE handler functions (init(), finish(), 4953109998Smarkm ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 4954109998Smarkm [Geoff] 4955109998Smarkm 4956109998Smarkm *) Implement binary inversion algorithm for BN_mod_inverse in addition 4957109998Smarkm to the algorithm using long division. The binary algorithm can be 4958109998Smarkm used only if the modulus is odd. On 32-bit systems, it is faster 4959109998Smarkm only for relatively small moduli (roughly 20-30% for 128-bit moduli, 4960109998Smarkm roughly 5-15% for 256-bit moduli), so we use it only for moduli 4961109998Smarkm up to 450 bits. In 64-bit environments, the binary algorithm 4962109998Smarkm appears to be advantageous for much longer moduli; here we use it 4963109998Smarkm for moduli up to 2048 bits. 4964109998Smarkm [Bodo Moeller] 4965109998Smarkm 4966109998Smarkm *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 4967109998Smarkm could not support the combine flag in choice fields. 4968109998Smarkm [Steve Henson] 4969109998Smarkm 4970109998Smarkm *) Add a 'copy_extensions' option to the 'ca' utility. This copies 4971109998Smarkm extensions from a certificate request to the certificate. 4972109998Smarkm [Steve Henson] 4973109998Smarkm 4974109998Smarkm *) Allow multiple 'certopt' and 'nameopt' options to be separated 4975109998Smarkm by commas. Add 'namopt' and 'certopt' options to the 'ca' config 4976109998Smarkm file: this allows the display of the certificate about to be 4977109998Smarkm signed to be customised, to allow certain fields to be included 4978109998Smarkm or excluded and extension details. The old system didn't display 4979109998Smarkm multicharacter strings properly, omitted fields not in the policy 4980109998Smarkm and couldn't display additional details such as extensions. 4981109998Smarkm [Steve Henson] 4982109998Smarkm 4983109998Smarkm *) Function EC_POINTs_mul for multiple scalar multiplication 4984109998Smarkm of an arbitrary number of elliptic curve points 4985109998Smarkm \sum scalars[i]*points[i], 4986109998Smarkm optionally including the generator defined for the EC_GROUP: 4987109998Smarkm scalar*generator + \sum scalars[i]*points[i]. 4988109998Smarkm 4989109998Smarkm EC_POINT_mul is a simple wrapper function for the typical case 4990109998Smarkm that the point list has just one item (besides the optional 4991109998Smarkm generator). 4992109998Smarkm [Bodo Moeller] 4993109998Smarkm 4994109998Smarkm *) First EC_METHODs for curves over GF(p): 4995109998Smarkm 4996109998Smarkm EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 4997109998Smarkm operations and provides various method functions that can also 4998109998Smarkm operate with faster implementations of modular arithmetic. 4999109998Smarkm 5000109998Smarkm EC_GFp_mont_method() reuses most functions that are part of 5001109998Smarkm EC_GFp_simple_method, but uses Montgomery arithmetic. 5002109998Smarkm 5003109998Smarkm [Bodo Moeller; point addition and point doubling 5004109998Smarkm implementation directly derived from source code provided by 5005109998Smarkm Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 5006109998Smarkm 5007109998Smarkm *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 5008109998Smarkm crypto/ec/ec_lib.c): 5009109998Smarkm 5010109998Smarkm Curves are EC_GROUP objects (with an optional group generator) 5011109998Smarkm based on EC_METHODs that are built into the library. 5012109998Smarkm 5013109998Smarkm Points are EC_POINT objects based on EC_GROUP objects. 5014109998Smarkm 5015109998Smarkm Most of the framework would be able to handle curves over arbitrary 5016109998Smarkm finite fields, but as there are no obvious types for fields other 5017109998Smarkm than GF(p), some functions are limited to that for now. 5018109998Smarkm [Bodo Moeller] 5019109998Smarkm 5020109998Smarkm *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 5021109998Smarkm that the file contains a complete HTTP response. 5022109998Smarkm [Richard Levitte] 5023109998Smarkm 5024109998Smarkm *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 5025109998Smarkm change the def and num file printf format specifier from "%-40sXXX" 5026109998Smarkm to "%-39s XXX". The latter will always guarantee a space after the 5027109998Smarkm field while the former will cause them to run together if the field 5028109998Smarkm is 40 of more characters long. 5029109998Smarkm [Steve Henson] 5030109998Smarkm 5031109998Smarkm *) Constify the cipher and digest 'method' functions and structures 5032109998Smarkm and modify related functions to take constant EVP_MD and EVP_CIPHER 5033109998Smarkm pointers. 5034109998Smarkm [Steve Henson] 5035109998Smarkm 5036109998Smarkm *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 5037109998Smarkm in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 5038109998Smarkm [Bodo Moeller] 5039109998Smarkm 5040109998Smarkm *) Modify EVP_Digest*() routines so they now return values. Although the 5041109998Smarkm internal software routines can never fail additional hardware versions 5042109998Smarkm might. 5043109998Smarkm [Steve Henson] 5044109998Smarkm 5045109998Smarkm *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 5046109998Smarkm 5047109998Smarkm Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 5048109998Smarkm (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 5049109998Smarkm 5050109998Smarkm ASN1 error codes 5051109998Smarkm ERR_R_NESTED_ASN1_ERROR 5052109998Smarkm ... 5053109998Smarkm ERR_R_MISSING_ASN1_EOS 5054109998Smarkm were 4 .. 9, conflicting with 5055109998Smarkm ERR_LIB_RSA (= ERR_R_RSA_LIB) 5056109998Smarkm ... 5057109998Smarkm ERR_LIB_PEM (= ERR_R_PEM_LIB). 5058109998Smarkm They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 5059109998Smarkm 5060109998Smarkm Add new error code 'ERR_R_INTERNAL_ERROR'. 5061109998Smarkm [Bodo Moeller] 5062109998Smarkm 5063109998Smarkm *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 5064109998Smarkm suffices. 5065109998Smarkm [Bodo Moeller] 5066109998Smarkm 5067109998Smarkm *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 5068109998Smarkm sets the subject name for a new request or supersedes the 5069109998Smarkm subject name in a given request. Formats that can be parsed are 5070109998Smarkm 'CN=Some Name, OU=myOU, C=IT' 5071109998Smarkm and 5072109998Smarkm 'CN=Some Name/OU=myOU/C=IT'. 5073109998Smarkm 5074109998Smarkm Add options '-batch' and '-verbose' to 'openssl req'. 5075109998Smarkm [Massimiliano Pala <madwolf@hackmasters.net>] 5076109998Smarkm 5077109998Smarkm *) Introduce the possibility to access global variables through 5078109998Smarkm functions on platform were that's the best way to handle exporting 5079109998Smarkm global variables in shared libraries. To enable this functionality, 5080109998Smarkm one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 5081109998Smarkm "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 5082109998Smarkm is normally done by Configure or something similar). 5083109998Smarkm 5084109998Smarkm To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 5085109998Smarkm in the source file (foo.c) like this: 5086109998Smarkm 5087109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 5088109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(double,bar); 5089109998Smarkm 5090109998Smarkm To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 5091109998Smarkm and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 5092109998Smarkm 5093109998Smarkm OPENSSL_DECLARE_GLOBAL(int,foo); 5094109998Smarkm #define foo OPENSSL_GLOBAL_REF(foo) 5095109998Smarkm OPENSSL_DECLARE_GLOBAL(double,bar); 5096109998Smarkm #define bar OPENSSL_GLOBAL_REF(bar) 5097109998Smarkm 5098109998Smarkm The #defines are very important, and therefore so is including the 5099109998Smarkm header file everywhere where the defined globals are used. 5100109998Smarkm 5101109998Smarkm The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 5102109998Smarkm of ASN.1 items, but that structure is a bit different. 5103109998Smarkm 5104109998Smarkm The largest change is in util/mkdef.pl which has been enhanced with 5105109998Smarkm better and easier to understand logic to choose which symbols should 5106109998Smarkm go into the Windows .def files as well as a number of fixes and code 5107109998Smarkm cleanup (among others, algorithm keywords are now sorted 5108109998Smarkm lexicographically to avoid constant rewrites). 5109109998Smarkm [Richard Levitte] 5110109998Smarkm 5111109998Smarkm *) In BN_div() keep a copy of the sign of 'num' before writing the 5112109998Smarkm result to 'rm' because if rm==num the value will be overwritten 5113109998Smarkm and produce the wrong result if 'num' is negative: this caused 5114109998Smarkm problems with BN_mod() and BN_nnmod(). 5115109998Smarkm [Steve Henson] 5116109998Smarkm 5117109998Smarkm *) Function OCSP_request_verify(). This checks the signature on an 5118109998Smarkm OCSP request and verifies the signer certificate. The signer 5119109998Smarkm certificate is just checked for a generic purpose and OCSP request 5120109998Smarkm trust settings. 5121109998Smarkm [Steve Henson] 5122109998Smarkm 5123109998Smarkm *) Add OCSP_check_validity() function to check the validity of OCSP 5124109998Smarkm responses. OCSP responses are prepared in real time and may only 5125109998Smarkm be a few seconds old. Simply checking that the current time lies 5126109998Smarkm between thisUpdate and nextUpdate max reject otherwise valid responses 5127109998Smarkm caused by either OCSP responder or client clock inaccuracy. Instead 5128109998Smarkm we allow thisUpdate and nextUpdate to fall within a certain period of 5129109998Smarkm the current time. The age of the response can also optionally be 5130109998Smarkm checked. Two new options -validity_period and -status_age added to 5131109998Smarkm ocsp utility. 5132109998Smarkm [Steve Henson] 5133109998Smarkm 5134109998Smarkm *) If signature or public key algorithm is unrecognized print out its 5135109998Smarkm OID rather that just UNKNOWN. 5136109998Smarkm [Steve Henson] 5137109998Smarkm 5138109998Smarkm *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 5139109998Smarkm OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 5140109998Smarkm ID to be generated from the issuer certificate alone which can then be 5141109998Smarkm passed to OCSP_id_issuer_cmp(). 5142109998Smarkm [Steve Henson] 5143109998Smarkm 5144109998Smarkm *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 5145109998Smarkm ASN1 modules to export functions returning ASN1_ITEM pointers 5146109998Smarkm instead of the ASN1_ITEM structures themselves. This adds several 5147109998Smarkm new macros which allow the underlying ASN1 function/structure to 5148109998Smarkm be accessed transparently. As a result code should not use ASN1_ITEM 5149109998Smarkm references directly (such as &X509_it) but instead use the relevant 5150109998Smarkm macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 5151109998Smarkm use of the new ASN1 code on platforms where exporting structures 5152109998Smarkm is problematical (for example in shared libraries) but exporting 5153109998Smarkm functions returning pointers to structures is not. 5154109998Smarkm [Steve Henson] 5155109998Smarkm 5156109998Smarkm *) Add support for overriding the generation of SSL/TLS session IDs. 5157109998Smarkm These callbacks can be registered either in an SSL_CTX or per SSL. 5158109998Smarkm The purpose of this is to allow applications to control, if they wish, 5159109998Smarkm the arbitrary values chosen for use as session IDs, particularly as it 5160109998Smarkm can be useful for session caching in multiple-server environments. A 5161109998Smarkm command-line switch for testing this (and any client code that wishes 5162109998Smarkm to use such a feature) has been added to "s_server". 5163109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 5164109998Smarkm 5165109998Smarkm *) Modify mkdef.pl to recognise and parse preprocessor conditionals 5166109998Smarkm of the form '#if defined(...) || defined(...) || ...' and 5167109998Smarkm '#if !defined(...) && !defined(...) && ...'. This also avoids 5168109998Smarkm the growing number of special cases it was previously handling. 5169109998Smarkm [Richard Levitte] 5170109998Smarkm 5171109998Smarkm *) Make all configuration macros available for application by making 5172109998Smarkm sure they are available in opensslconf.h, by giving them names starting 5173109998Smarkm with "OPENSSL_" to avoid conflicts with other packages and by making 5174109998Smarkm sure e_os2.h will cover all platform-specific cases together with 5175109998Smarkm opensslconf.h. 5176109998Smarkm Additionally, it is now possible to define configuration/platform- 5177109998Smarkm specific names (called "system identities"). In the C code, these 5178109998Smarkm are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 5179109998Smarkm macro with the name beginning with "OPENSSL_SYS_", which is determined 5180109998Smarkm from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 5181109998Smarkm what is available. 5182109998Smarkm [Richard Levitte] 5183109998Smarkm 5184109998Smarkm *) New option -set_serial to 'req' and 'x509' this allows the serial 5185109998Smarkm number to use to be specified on the command line. Previously self 5186109998Smarkm signed certificates were hard coded with serial number 0 and the 5187109998Smarkm CA options of 'x509' had to use a serial number in a file which was 5188109998Smarkm auto incremented. 5189109998Smarkm [Steve Henson] 5190109998Smarkm 5191109998Smarkm *) New options to 'ca' utility to support V2 CRL entry extensions. 5192109998Smarkm Currently CRL reason, invalidity date and hold instruction are 5193109998Smarkm supported. Add new CRL extensions to V3 code and some new objects. 5194109998Smarkm [Steve Henson] 5195109998Smarkm 5196109998Smarkm *) New function EVP_CIPHER_CTX_set_padding() this is used to 5197109998Smarkm disable standard block padding (aka PKCS#5 padding) in the EVP 5198109998Smarkm API, which was previously mandatory. This means that the data is 5199109998Smarkm not padded in any way and so the total length much be a multiple 5200109998Smarkm of the block size, otherwise an error occurs. 5201109998Smarkm [Steve Henson] 5202109998Smarkm 5203109998Smarkm *) Initial (incomplete) OCSP SSL support. 5204109998Smarkm [Steve Henson] 5205109998Smarkm 5206109998Smarkm *) New function OCSP_parse_url(). This splits up a URL into its host, 5207109998Smarkm port and path components: primarily to parse OCSP URLs. New -url 5208109998Smarkm option to ocsp utility. 5209109998Smarkm [Steve Henson] 5210109998Smarkm 5211109998Smarkm *) New nonce behavior. The return value of OCSP_check_nonce() now 5212109998Smarkm reflects the various checks performed. Applications can decide 5213109998Smarkm whether to tolerate certain situations such as an absent nonce 5214109998Smarkm in a response when one was present in a request: the ocsp application 5215109998Smarkm just prints out a warning. New function OCSP_add1_basic_nonce() 5216109998Smarkm this is to allow responders to include a nonce in a response even if 5217109998Smarkm the request is nonce-less. 5218109998Smarkm [Steve Henson] 5219109998Smarkm 5220109998Smarkm *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 5221109998Smarkm skipped when using openssl x509 multiple times on a single input file, 5222109998Smarkm e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 5223109998Smarkm [Bodo Moeller] 5224109998Smarkm 5225109998Smarkm *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 5226109998Smarkm set string type: to handle setting ASN1_TIME structures. Fix ca 5227109998Smarkm utility to correctly initialize revocation date of CRLs. 5228109998Smarkm [Steve Henson] 5229109998Smarkm 5230109998Smarkm *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 5231109998Smarkm the clients preferred ciphersuites and rather use its own preferences. 5232109998Smarkm Should help to work around M$ SGC (Server Gated Cryptography) bug in 5233109998Smarkm Internet Explorer by ensuring unchanged hash method during stepup. 5234109998Smarkm (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 5235109998Smarkm [Lutz Jaenicke] 5236109998Smarkm 5237109998Smarkm *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 5238109998Smarkm to aes and add a new 'exist' option to print out symbols that don't 5239109998Smarkm appear to exist. 5240109998Smarkm [Steve Henson] 5241109998Smarkm 5242109998Smarkm *) Additional options to ocsp utility to allow flags to be set and 5243109998Smarkm additional certificates supplied. 5244109998Smarkm [Steve Henson] 5245109998Smarkm 5246109998Smarkm *) Add the option -VAfile to 'openssl ocsp', so the user can give the 5247109998Smarkm OCSP client a number of certificate to only verify the response 5248109998Smarkm signature against. 5249109998Smarkm [Richard Levitte] 5250109998Smarkm 5251109998Smarkm *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 5252109998Smarkm handle the new API. Currently only ECB, CBC modes supported. Add new 5253109998Smarkm AES OIDs. 5254109998Smarkm 5255109998Smarkm Add TLS AES ciphersuites as described in RFC3268, "Advanced 5256109998Smarkm Encryption Standard (AES) Ciphersuites for Transport Layer 5257109998Smarkm Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 5258109998Smarkm not enabled by default and were not part of the "ALL" ciphersuite 5259109998Smarkm alias because they were not yet official; they could be 5260109998Smarkm explicitly requested by specifying the "AESdraft" ciphersuite 5261109998Smarkm group alias. In the final release of OpenSSL 0.9.7, the group 5262109998Smarkm alias is called "AES" and is part of "ALL".) 5263109998Smarkm [Ben Laurie, Steve Henson, Bodo Moeller] 5264109998Smarkm 5265109998Smarkm *) New function OCSP_copy_nonce() to copy nonce value (if present) from 5266109998Smarkm request to response. 5267109998Smarkm [Steve Henson] 5268109998Smarkm 5269109998Smarkm *) Functions for OCSP responders. OCSP_request_onereq_count(), 5270109998Smarkm OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 5271109998Smarkm extract information from a certificate request. OCSP_response_create() 5272109998Smarkm creates a response and optionally adds a basic response structure. 5273109998Smarkm OCSP_basic_add1_status() adds a complete single response to a basic 5274109998Smarkm response and returns the OCSP_SINGLERESP structure just added (to allow 5275109998Smarkm extensions to be included for example). OCSP_basic_add1_cert() adds a 5276109998Smarkm certificate to a basic response and OCSP_basic_sign() signs a basic 5277109998Smarkm response with various flags. New helper functions ASN1_TIME_check() 5278109998Smarkm (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 5279109998Smarkm (converts ASN1_TIME to GeneralizedTime). 5280109998Smarkm [Steve Henson] 5281109998Smarkm 5282109998Smarkm *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 5283109998Smarkm in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 5284109998Smarkm structure from a certificate. X509_pubkey_digest() digests the public_key 5285109998Smarkm contents: this is used in various key identifiers. 5286109998Smarkm [Steve Henson] 5287109998Smarkm 5288109998Smarkm *) Make sk_sort() tolerate a NULL argument. 5289109998Smarkm [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 5290109998Smarkm 5291109998Smarkm *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 5292109998Smarkm passed by the function are trusted implicitly. If any of them signed the 5293109998Smarkm response then it is assumed to be valid and is not verified. 5294109998Smarkm [Steve Henson] 5295109998Smarkm 5296109998Smarkm *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 5297109998Smarkm to data. This was previously part of the PKCS7 ASN1 code. This 5298109998Smarkm was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 5299109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 5300109998Smarkm <support@securenetterm.com>] 5301109998Smarkm 5302109998Smarkm *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 5303109998Smarkm routines: without these tracing memory leaks is very painful. 5304109998Smarkm Fix leaks in PKCS12 and PKCS7 routines. 5305109998Smarkm [Steve Henson] 5306109998Smarkm 5307109998Smarkm *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 5308109998Smarkm Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 5309109998Smarkm effectively meant GeneralizedTime would never be used. Now it 5310109998Smarkm is initialised to -1 but X509_time_adj() now has to check the value 5311109998Smarkm and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 5312109998Smarkm V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 5313109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 5314109998Smarkm <support@securenetterm.com>] 5315109998Smarkm 5316109998Smarkm *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 5317109998Smarkm result in a zero length in the ASN1_INTEGER structure which was 5318109998Smarkm not consistent with the structure when d2i_ASN1_INTEGER() was used 5319109998Smarkm and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 5320109998Smarkm to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 5321109998Smarkm where it did not print out a minus for negative ASN1_INTEGER. 5322109998Smarkm [Steve Henson] 5323109998Smarkm 5324109998Smarkm *) Add summary printout to ocsp utility. The various functions which 5325109998Smarkm convert status values to strings have been renamed to: 5326109998Smarkm OCSP_response_status_str(), OCSP_cert_status_str() and 5327109998Smarkm OCSP_crl_reason_str() and are no longer static. New options 5328109998Smarkm to verify nonce values and to disable verification. OCSP response 5329109998Smarkm printout format cleaned up. 5330109998Smarkm [Steve Henson] 5331109998Smarkm 5332109998Smarkm *) Add additional OCSP certificate checks. These are those specified 5333109998Smarkm in RFC2560. This consists of two separate checks: the CA of the 5334109998Smarkm certificate being checked must either be the OCSP signer certificate 5335109998Smarkm or the issuer of the OCSP signer certificate. In the latter case the 5336109998Smarkm OCSP signer certificate must contain the OCSP signing extended key 5337109998Smarkm usage. This check is performed by attempting to match the OCSP 5338109998Smarkm signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 5339109998Smarkm in the OCSP_CERTID structures of the response. 5340109998Smarkm [Steve Henson] 5341109998Smarkm 5342109998Smarkm *) Initial OCSP certificate verification added to OCSP_basic_verify() 5343109998Smarkm and related routines. This uses the standard OpenSSL certificate 5344109998Smarkm verify routines to perform initial checks (just CA validity) and 5345109998Smarkm to obtain the certificate chain. Then additional checks will be 5346109998Smarkm performed on the chain. Currently the root CA is checked to see 5347109998Smarkm if it is explicitly trusted for OCSP signing. This is used to set 5348109998Smarkm a root CA as a global signing root: that is any certificate that 5349109998Smarkm chains to that CA is an acceptable OCSP signing certificate. 5350109998Smarkm [Steve Henson] 5351109998Smarkm 5352109998Smarkm *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 5353109998Smarkm extensions from a separate configuration file. 5354109998Smarkm As when reading extensions from the main configuration file, 5355109998Smarkm the '-extensions ...' option may be used for specifying the 5356109998Smarkm section to use. 5357109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 5358109998Smarkm 5359109998Smarkm *) New OCSP utility. Allows OCSP requests to be generated or 5360109998Smarkm read. The request can be sent to a responder and the output 5361109998Smarkm parsed, outputed or printed in text form. Not complete yet: 5362109998Smarkm still needs to check the OCSP response validity. 5363109998Smarkm [Steve Henson] 5364109998Smarkm 5365109998Smarkm *) New subcommands for 'openssl ca': 5366109998Smarkm 'openssl ca -status <serial>' prints the status of the cert with 5367109998Smarkm the given serial number (according to the index file). 5368109998Smarkm 'openssl ca -updatedb' updates the expiry status of certificates 5369109998Smarkm in the index file. 5370109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 5371109998Smarkm 5372109998Smarkm *) New '-newreq-nodes' command option to CA.pl. This is like 5373109998Smarkm '-newreq', but calls 'openssl req' with the '-nodes' option 5374109998Smarkm so that the resulting key is not encrypted. 5375109998Smarkm [Damien Miller <djm@mindrot.org>] 5376109998Smarkm 5377109998Smarkm *) New configuration for the GNU Hurd. 5378109998Smarkm [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 5379109998Smarkm 5380109998Smarkm *) Initial code to implement OCSP basic response verify. This 5381109998Smarkm is currently incomplete. Currently just finds the signer's 5382109998Smarkm certificate and verifies the signature on the response. 5383109998Smarkm [Steve Henson] 5384109998Smarkm 5385109998Smarkm *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 5386109998Smarkm value of OPENSSLDIR. This is available via the new '-d' option 5387109998Smarkm to 'openssl version', and is also included in 'openssl version -a'. 5388109998Smarkm [Bodo Moeller] 5389109998Smarkm 5390109998Smarkm *) Allowing defining memory allocation callbacks that will be given 5391109998Smarkm file name and line number information in additional arguments 5392109998Smarkm (a const char* and an int). The basic functionality remains, as 5393109998Smarkm well as the original possibility to just replace malloc(), 5394109998Smarkm realloc() and free() by functions that do not know about these 5395109998Smarkm additional arguments. To register and find out the current 5396109998Smarkm settings for extended allocation functions, the following 5397109998Smarkm functions are provided: 5398109998Smarkm 5399109998Smarkm CRYPTO_set_mem_ex_functions 5400109998Smarkm CRYPTO_set_locked_mem_ex_functions 5401109998Smarkm CRYPTO_get_mem_ex_functions 5402109998Smarkm CRYPTO_get_locked_mem_ex_functions 5403109998Smarkm 5404109998Smarkm These work the same way as CRYPTO_set_mem_functions and friends. 5405109998Smarkm CRYPTO_get_[locked_]mem_functions now writes 0 where such an 5406109998Smarkm extended allocation function is enabled. 5407109998Smarkm Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 5408109998Smarkm a conventional allocation function is enabled. 5409109998Smarkm [Richard Levitte, Bodo Moeller] 5410109998Smarkm 5411109998Smarkm *) Finish off removing the remaining LHASH function pointer casts. 5412109998Smarkm There should no longer be any prototype-casting required when using 5413109998Smarkm the LHASH abstraction, and any casts that remain are "bugs". See 5414109998Smarkm the callback types and macros at the head of lhash.h for details 5415109998Smarkm (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 5416109998Smarkm [Geoff Thorpe] 5417109998Smarkm 5418109998Smarkm *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 5419109998Smarkm If /dev/[u]random devices are not available or do not return enough 5420109998Smarkm entropy, EGD style sockets (served by EGD or PRNGD) will automatically 5421109998Smarkm be queried. 5422109998Smarkm The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 5423109998Smarkm /etc/entropy will be queried once each in this sequence, quering stops 5424109998Smarkm when enough entropy was collected without querying more sockets. 5425109998Smarkm [Lutz Jaenicke] 5426109998Smarkm 5427109998Smarkm *) Change the Unix RAND_poll() variant to be able to poll several 5428109998Smarkm random devices, as specified by DEVRANDOM, until a sufficient amount 5429109998Smarkm of data has been collected. We spend at most 10 ms on each file 5430109998Smarkm (select timeout) and read in non-blocking mode. DEVRANDOM now 5431109998Smarkm defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 5432109998Smarkm (previously it was just the string "/dev/urandom"), so on typical 5433109998Smarkm platforms the 10 ms delay will never occur. 5434109998Smarkm Also separate out the Unix variant to its own file, rand_unix.c. 5435109998Smarkm For VMS, there's a currently-empty rand_vms.c. 5436109998Smarkm [Richard Levitte] 5437109998Smarkm 5438109998Smarkm *) Move OCSP client related routines to ocsp_cl.c. These 5439109998Smarkm provide utility functions which an application needing 5440109998Smarkm to issue a request to an OCSP responder and analyse the 5441109998Smarkm response will typically need: as opposed to those which an 5442109998Smarkm OCSP responder itself would need which will be added later. 5443109998Smarkm 5444109998Smarkm OCSP_request_sign() signs an OCSP request with an API similar 5445109998Smarkm to PKCS7_sign(). OCSP_response_status() returns status of OCSP 5446109998Smarkm response. OCSP_response_get1_basic() extracts basic response 5447109998Smarkm from response. OCSP_resp_find_status(): finds and extracts status 5448109998Smarkm information from an OCSP_CERTID structure (which will be created 5449109998Smarkm when the request structure is built). These are built from lower 5450109998Smarkm level functions which work on OCSP_SINGLERESP structures but 5451109998Smarkm wont normally be used unless the application wishes to examine 5452109998Smarkm extensions in the OCSP response for example. 5453109998Smarkm 5454109998Smarkm Replace nonce routines with a pair of functions. 5455109998Smarkm OCSP_request_add1_nonce() adds a nonce value and optionally 5456109998Smarkm generates a random value. OCSP_check_nonce() checks the 5457109998Smarkm validity of the nonce in an OCSP response. 5458109998Smarkm [Steve Henson] 5459109998Smarkm 5460109998Smarkm *) Change function OCSP_request_add() to OCSP_request_add0_id(). 5461109998Smarkm This doesn't copy the supplied OCSP_CERTID and avoids the 5462109998Smarkm need to free up the newly created id. Change return type 5463109998Smarkm to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 5464109998Smarkm This can then be used to add extensions to the request. 5465109998Smarkm Deleted OCSP_request_new(), since most of its functionality 5466109998Smarkm is now in OCSP_REQUEST_new() (and the case insensitive name 5467109998Smarkm clash) apart from the ability to set the request name which 5468109998Smarkm will be added elsewhere. 5469109998Smarkm [Steve Henson] 5470109998Smarkm 5471109998Smarkm *) Update OCSP API. Remove obsolete extensions argument from 5472109998Smarkm various functions. Extensions are now handled using the new 5473109998Smarkm OCSP extension code. New simple OCSP HTTP function which 5474109998Smarkm can be used to send requests and parse the response. 5475109998Smarkm [Steve Henson] 5476109998Smarkm 5477109998Smarkm *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 5478109998Smarkm ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 5479109998Smarkm uses the special reorder version of SET OF to sort the attributes 5480109998Smarkm and reorder them to match the encoded order. This resolves a long 5481109998Smarkm standing problem: a verify on a PKCS7 structure just after signing 5482109998Smarkm it used to fail because the attribute order did not match the 5483109998Smarkm encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 5484109998Smarkm it uses the received order. This is necessary to tolerate some broken 5485109998Smarkm software that does not order SET OF. This is handled by encoding 5486109998Smarkm as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 5487109998Smarkm to produce the required SET OF. 5488109998Smarkm [Steve Henson] 5489109998Smarkm 5490109998Smarkm *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 5491109998Smarkm OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 5492109998Smarkm files to get correct declarations of the ASN.1 item variables. 5493109998Smarkm [Richard Levitte] 5494109998Smarkm 5495109998Smarkm *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 5496109998Smarkm PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 5497109998Smarkm asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 5498109998Smarkm NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 5499109998Smarkm New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 5500109998Smarkm ASN1_ITEM and no wrapper functions. 5501109998Smarkm [Steve Henson] 5502109998Smarkm 5503109998Smarkm *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 5504109998Smarkm replace the old function pointer based I/O routines. Change most of 5505109998Smarkm the *_d2i_bio() and *_d2i_fp() functions to use these. 5506109998Smarkm [Steve Henson] 5507109998Smarkm 5508109998Smarkm *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 5509109998Smarkm lines, recognice more "algorithms" that can be deselected, and make 5510109998Smarkm it complain about algorithm deselection that isn't recognised. 5511109998Smarkm [Richard Levitte] 5512109998Smarkm 5513109998Smarkm *) New ASN1 functions to handle dup, sign, verify, digest, pack and 5514109998Smarkm unpack operations in terms of ASN1_ITEM. Modify existing wrappers 5515109998Smarkm to use new functions. Add NO_ASN1_OLD which can be set to remove 5516109998Smarkm some old style ASN1 functions: this can be used to determine if old 5517109998Smarkm code will still work when these eventually go away. 5518109998Smarkm [Steve Henson] 5519109998Smarkm 5520109998Smarkm *) New extension functions for OCSP structures, these follow the 5521109998Smarkm same conventions as certificates and CRLs. 5522109998Smarkm [Steve Henson] 5523109998Smarkm 5524109998Smarkm *) New function X509V3_add1_i2d(). This automatically encodes and 5525109998Smarkm adds an extension. Its behaviour can be customised with various 5526109998Smarkm flags to append, replace or delete. Various wrappers added for 5527109998Smarkm certifcates and CRLs. 5528109998Smarkm [Steve Henson] 5529109998Smarkm 5530109998Smarkm *) Fix to avoid calling the underlying ASN1 print routine when 5531109998Smarkm an extension cannot be parsed. Correct a typo in the 5532109998Smarkm OCSP_SERVICELOC extension. Tidy up print OCSP format. 5533109998Smarkm [Steve Henson] 5534109998Smarkm 5535109998Smarkm *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 5536109998Smarkm entries for variables. 5537109998Smarkm [Steve Henson] 5538109998Smarkm 5539109998Smarkm *) Add functionality to apps/openssl.c for detecting locking 5540109998Smarkm problems: As the program is single-threaded, all we have 5541109998Smarkm to do is register a locking callback using an array for 5542109998Smarkm storing which locks are currently held by the program. 5543109998Smarkm [Bodo Moeller] 5544109998Smarkm 5545109998Smarkm *) Use a lock around the call to CRYPTO_get_ex_new_index() in 5546109998Smarkm SSL_get_ex_data_X509_STORE_idx(), which is used in 5547109998Smarkm ssl_verify_cert_chain() and thus can be called at any time 5548109998Smarkm during TLS/SSL handshakes so that thread-safety is essential. 5549109998Smarkm Unfortunately, the ex_data design is not at all suited 5550109998Smarkm for multi-threaded use, so it probably should be abolished. 5551109998Smarkm [Bodo Moeller] 5552109998Smarkm 5553109998Smarkm *) Added Broadcom "ubsec" ENGINE to OpenSSL. 5554109998Smarkm [Broadcom, tweaked and integrated by Geoff Thorpe] 5555109998Smarkm 5556109998Smarkm *) Move common extension printing code to new function 5557109998Smarkm X509V3_print_extensions(). Reorganise OCSP print routines and 5558109998Smarkm implement some needed OCSP ASN1 functions. Add OCSP extensions. 5559109998Smarkm [Steve Henson] 5560109998Smarkm 5561109998Smarkm *) New function X509_signature_print() to remove duplication in some 5562109998Smarkm print routines. 5563109998Smarkm [Steve Henson] 5564109998Smarkm 5565109998Smarkm *) Add a special meaning when SET OF and SEQUENCE OF flags are both 5566109998Smarkm set (this was treated exactly the same as SET OF previously). This 5567109998Smarkm is used to reorder the STACK representing the structure to match the 5568109998Smarkm encoding. This will be used to get round a problem where a PKCS7 5569109998Smarkm structure which was signed could not be verified because the STACK 5570109998Smarkm order did not reflect the encoded order. 5571109998Smarkm [Steve Henson] 5572109998Smarkm 5573109998Smarkm *) Reimplement the OCSP ASN1 module using the new code. 5574109998Smarkm [Steve Henson] 5575109998Smarkm 5576109998Smarkm *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 5577109998Smarkm for its ASN1 operations. The old style function pointers still exist 5578109998Smarkm for now but they will eventually go away. 5579109998Smarkm [Steve Henson] 5580109998Smarkm 5581109998Smarkm *) Merge in replacement ASN1 code from the ASN1 branch. This almost 5582109998Smarkm completely replaces the old ASN1 functionality with a table driven 5583109998Smarkm encoder and decoder which interprets an ASN1_ITEM structure describing 5584109998Smarkm the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 5585109998Smarkm largely maintained. Almost all of the old asn1_mac.h macro based ASN1 5586109998Smarkm has also been converted to the new form. 5587109998Smarkm [Steve Henson] 5588109998Smarkm 5589109998Smarkm *) Change BN_mod_exp_recp so that negative moduli are tolerated 5590109998Smarkm (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 5591109998Smarkm so that BN_mod_exp_mont and BN_mod_exp_mont_word work 5592109998Smarkm for negative moduli. 5593109998Smarkm [Bodo Moeller] 5594109998Smarkm 5595109998Smarkm *) Fix BN_uadd and BN_usub: Always return non-negative results instead 5596109998Smarkm of not touching the result's sign bit. 5597109998Smarkm [Bodo Moeller] 5598109998Smarkm 5599109998Smarkm *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 5600109998Smarkm set. 5601109998Smarkm [Bodo Moeller] 5602109998Smarkm 5603109998Smarkm *) Changed the LHASH code to use prototypes for callbacks, and created 5604109998Smarkm macros to declare and implement thin (optionally static) functions 5605109998Smarkm that provide type-safety and avoid function pointer casting for the 5606109998Smarkm type-specific callbacks. 5607109998Smarkm [Geoff Thorpe] 5608109998Smarkm 5609109998Smarkm *) Added Kerberos Cipher Suites to be used with TLS, as written in 5610109998Smarkm RFC 2712. 5611109998Smarkm [Veers Staats <staatsvr@asc.hpc.mil>, 5612109998Smarkm Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 5613109998Smarkm 5614109998Smarkm *) Reformat the FAQ so the different questions and answers can be divided 5615109998Smarkm in sections depending on the subject. 5616109998Smarkm [Richard Levitte] 5617109998Smarkm 5618109998Smarkm *) Have the zlib compression code load ZLIB.DLL dynamically under 5619109998Smarkm Windows. 5620109998Smarkm [Richard Levitte] 5621109998Smarkm 5622109998Smarkm *) New function BN_mod_sqrt for computing square roots modulo a prime 5623109998Smarkm (using the probabilistic Tonelli-Shanks algorithm unless 5624109998Smarkm p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 5625109998Smarkm be handled deterministically). 5626109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5627109998Smarkm 5628109998Smarkm *) Make BN_mod_inverse faster by explicitly handling small quotients 5629109998Smarkm in the Euclid loop. (Speed gain about 20% for small moduli [256 or 5630109998Smarkm 512 bits], about 30% for larger ones [1024 or 2048 bits].) 5631109998Smarkm [Bodo Moeller] 5632109998Smarkm 5633109998Smarkm *) New function BN_kronecker. 5634109998Smarkm [Bodo Moeller] 5635109998Smarkm 5636109998Smarkm *) Fix BN_gcd so that it works on negative inputs; the result is 5637109998Smarkm positive unless both parameters are zero. 5638109998Smarkm Previously something reasonably close to an infinite loop was 5639109998Smarkm possible because numbers could be growing instead of shrinking 5640109998Smarkm in the implementation of Euclid's algorithm. 5641109998Smarkm [Bodo Moeller] 5642109998Smarkm 5643109998Smarkm *) Fix BN_is_word() and BN_is_one() macros to take into account the 5644109998Smarkm sign of the number in question. 5645109998Smarkm 5646109998Smarkm Fix BN_is_word(a,w) to work correctly for w == 0. 5647109998Smarkm 5648109998Smarkm The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 5649109998Smarkm because its test if the absolute value of 'a' equals 'w'. 5650109998Smarkm Note that BN_abs_is_word does *not* handle w == 0 reliably; 5651109998Smarkm it exists mostly for use in the implementations of BN_is_zero(), 5652109998Smarkm BN_is_one(), and BN_is_word(). 5653109998Smarkm [Bodo Moeller] 5654109998Smarkm 5655109998Smarkm *) New function BN_swap. 5656109998Smarkm [Bodo Moeller] 5657109998Smarkm 5658109998Smarkm *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 5659109998Smarkm the exponentiation functions are more likely to produce reasonable 5660109998Smarkm results on negative inputs. 5661109998Smarkm [Bodo Moeller] 5662109998Smarkm 5663109998Smarkm *) Change BN_mod_mul so that the result is always non-negative. 5664109998Smarkm Previously, it could be negative if one of the factors was negative; 5665109998Smarkm I don't think anyone really wanted that behaviour. 5666109998Smarkm [Bodo Moeller] 5667109998Smarkm 5668109998Smarkm *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 5669109998Smarkm (except for exponentiation, which stays in crypto/bn/bn_exp.c, 5670109998Smarkm and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 5671109998Smarkm and add new functions: 5672109998Smarkm 5673109998Smarkm BN_nnmod 5674109998Smarkm BN_mod_sqr 5675109998Smarkm BN_mod_add 5676109998Smarkm BN_mod_add_quick 5677109998Smarkm BN_mod_sub 5678109998Smarkm BN_mod_sub_quick 5679109998Smarkm BN_mod_lshift1 5680109998Smarkm BN_mod_lshift1_quick 5681109998Smarkm BN_mod_lshift 5682109998Smarkm BN_mod_lshift_quick 5683109998Smarkm 5684109998Smarkm These functions always generate non-negative results. 5685109998Smarkm 5686109998Smarkm BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 5687109998Smarkm such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 5688109998Smarkm 5689109998Smarkm BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 5690109998Smarkm BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 5691109998Smarkm be reduced modulo m. 5692109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 5693109998Smarkm 5694111147Snectar#if 0 5695111147Snectar The following entry accidentily appeared in the CHANGES file 5696111147Snectar distributed with OpenSSL 0.9.7. The modifications described in 5697111147Snectar it do *not* apply to OpenSSL 0.9.7. 5698111147Snectar 5699109998Smarkm *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 5700109998Smarkm was actually never needed) and in BN_mul(). The removal in BN_mul() 5701109998Smarkm required a small change in bn_mul_part_recursive() and the addition 5702109998Smarkm of the functions bn_cmp_part_words(), bn_sub_part_words() and 5703109998Smarkm bn_add_part_words(), which do the same thing as bn_cmp_words(), 5704109998Smarkm bn_sub_words() and bn_add_words() except they take arrays with 5705109998Smarkm differing sizes. 5706109998Smarkm [Richard Levitte] 5707111147Snectar#endif 5708109998Smarkm 5709109998Smarkm *) In 'openssl passwd', verify passwords read from the terminal 5710109998Smarkm unless the '-salt' option is used (which usually means that 5711109998Smarkm verification would just waste user's time since the resulting 5712109998Smarkm hash is going to be compared with some given password hash) 5713109998Smarkm or the new '-noverify' option is used. 5714109998Smarkm 5715109998Smarkm This is an incompatible change, but it does not affect 5716109998Smarkm non-interactive use of 'openssl passwd' (passwords on the command 5717109998Smarkm line, '-stdin' option, '-in ...' option) and thus should not 5718109998Smarkm cause any problems. 5719109998Smarkm [Bodo Moeller] 5720109998Smarkm 5721109998Smarkm *) Remove all references to RSAref, since there's no more need for it. 5722109998Smarkm [Richard Levitte] 5723109998Smarkm 5724109998Smarkm *) Make DSO load along a path given through an environment variable 5725109998Smarkm (SHLIB_PATH) with shl_load(). 5726109998Smarkm [Richard Levitte] 5727109998Smarkm 5728109998Smarkm *) Constify the ENGINE code as a result of BIGNUM constification. 5729109998Smarkm Also constify the RSA code and most things related to it. In a 5730109998Smarkm few places, most notable in the depth of the ASN.1 code, ugly 5731109998Smarkm casts back to non-const were required (to be solved at a later 5732109998Smarkm time) 5733109998Smarkm [Richard Levitte] 5734109998Smarkm 5735109998Smarkm *) Make it so the openssl application has all engines loaded by default. 5736109998Smarkm [Richard Levitte] 5737109998Smarkm 5738109998Smarkm *) Constify the BIGNUM routines a little more. 5739109998Smarkm [Richard Levitte] 5740109998Smarkm 5741109998Smarkm *) Add the following functions: 5742109998Smarkm 5743109998Smarkm ENGINE_load_cswift() 5744109998Smarkm ENGINE_load_chil() 5745109998Smarkm ENGINE_load_atalla() 5746109998Smarkm ENGINE_load_nuron() 5747109998Smarkm ENGINE_load_builtin_engines() 5748109998Smarkm 5749109998Smarkm That way, an application can itself choose if external engines that 5750109998Smarkm are built-in in OpenSSL shall ever be used or not. The benefit is 5751109998Smarkm that applications won't have to be linked with libdl or other dso 5752109998Smarkm libraries unless it's really needed. 5753109998Smarkm 5754109998Smarkm Changed 'openssl engine' to load all engines on demand. 5755109998Smarkm Changed the engine header files to avoid the duplication of some 5756109998Smarkm declarations (they differed!). 5757109998Smarkm [Richard Levitte] 5758109998Smarkm 5759109998Smarkm *) 'openssl engine' can now list capabilities. 5760109998Smarkm [Richard Levitte] 5761109998Smarkm 5762109998Smarkm *) Better error reporting in 'openssl engine'. 5763109998Smarkm [Richard Levitte] 5764109998Smarkm 5765109998Smarkm *) Never call load_dh_param(NULL) in s_server. 5766109998Smarkm [Bodo Moeller] 5767109998Smarkm 5768109998Smarkm *) Add engine application. It can currently list engines by name and 5769109998Smarkm identity, and test if they are actually available. 5770109998Smarkm [Richard Levitte] 5771109998Smarkm 5772109998Smarkm *) Improve RPM specification file by forcing symbolic linking and making 5773109998Smarkm sure the installed documentation is also owned by root.root. 5774109998Smarkm [Damien Miller <djm@mindrot.org>] 5775109998Smarkm 5776109998Smarkm *) Give the OpenSSL applications more possibilities to make use of 5777109998Smarkm keys (public as well as private) handled by engines. 5778109998Smarkm [Richard Levitte] 5779109998Smarkm 5780109998Smarkm *) Add OCSP code that comes from CertCo. 5781109998Smarkm [Richard Levitte] 5782109998Smarkm 5783109998Smarkm *) Add VMS support for the Rijndael code. 5784109998Smarkm [Richard Levitte] 5785109998Smarkm 5786109998Smarkm *) Added untested support for Nuron crypto accelerator. 5787109998Smarkm [Ben Laurie] 5788109998Smarkm 5789109998Smarkm *) Add support for external cryptographic devices. This code was 5790109998Smarkm previously distributed separately as the "engine" branch. 5791109998Smarkm [Geoff Thorpe, Richard Levitte] 5792109998Smarkm 5793109998Smarkm *) Rework the filename-translation in the DSO code. It is now possible to 5794109998Smarkm have far greater control over how a "name" is turned into a filename 5795109998Smarkm depending on the operating environment and any oddities about the 5796109998Smarkm different shared library filenames on each system. 5797109998Smarkm [Geoff Thorpe] 5798109998Smarkm 5799109998Smarkm *) Support threads on FreeBSD-elf in Configure. 5800109998Smarkm [Richard Levitte] 5801109998Smarkm 5802109998Smarkm *) Fix for SHA1 assembly problem with MASM: it produces 5803109998Smarkm warnings about corrupt line number information when assembling 5804109998Smarkm with debugging information. This is caused by the overlapping 5805109998Smarkm of two sections. 5806109998Smarkm [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 5807109998Smarkm 5808109998Smarkm *) NCONF changes. 5809109998Smarkm NCONF_get_number() has no error checking at all. As a replacement, 5810109998Smarkm NCONF_get_number_e() is defined (_e for "error checking") and is 5811109998Smarkm promoted strongly. The old NCONF_get_number is kept around for 5812109998Smarkm binary backward compatibility. 5813109998Smarkm Make it possible for methods to load from something other than a BIO, 5814109998Smarkm by providing a function pointer that is given a name instead of a BIO. 5815109998Smarkm For example, this could be used to load configuration data from an 5816109998Smarkm LDAP server. 5817109998Smarkm [Richard Levitte] 5818109998Smarkm 5819109998Smarkm *) Fix for non blocking accept BIOs. Added new I/O special reason 5820109998Smarkm BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 5821109998Smarkm with non blocking I/O was not possible because no retry code was 5822109998Smarkm implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 5823109998Smarkm this case. 5824109998Smarkm [Steve Henson] 5825109998Smarkm 5826109998Smarkm *) Added the beginnings of Rijndael support. 5827109998Smarkm [Ben Laurie] 5828109998Smarkm 5829109998Smarkm *) Fix for bug in DirectoryString mask setting. Add support for 5830109998Smarkm X509_NAME_print_ex() in 'req' and X509_print_ex() function 5831109998Smarkm to allow certificate printing to more controllable, additional 5832109998Smarkm 'certopt' option to 'x509' to allow new printing options to be 5833109998Smarkm set. 5834109998Smarkm [Steve Henson] 5835109998Smarkm 5836109998Smarkm *) Clean old EAY MD5 hack from e_os.h. 5837109998Smarkm [Richard Levitte] 5838109998Smarkm 5839142425Snectar Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 5840142425Snectar 5841142425Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 5842160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 5843142425Snectar [Joe Orton, Steve Henson] 5844142425Snectar 5845142425Snectar Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 5846142425Snectar 5847142425Snectar *) Fix additional bug revealed by the NISCC test suite: 5848142425Snectar 5849142425Snectar Stop bug triggering large recursion when presented with 5850160814Ssimon certain ASN.1 tags (CVE-2003-0851) 5851142425Snectar [Steve Henson] 5852142425Snectar 5853120631Snectar Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 5854120631Snectar 5855120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 5856120631Snectar 5857120631Snectar Stop out of bounds reads in the ASN1 code when presented with 5858160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 5859120631Snectar 5860120631Snectar If verify callback ignores invalid public key errors don't try to check 5861120631Snectar certificate signature with the NULL public key. 5862120631Snectar 5863120631Snectar [Steve Henson] 5864120631Snectar 5865120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 5866120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 5867120631Snectar specifications. 5868120631Snectar [Steve Henson] 5869120631Snectar 5870120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 5871120631Snectar extra data after the compression methods not only for TLS 1.0 5872120631Snectar but also for SSL 3.0 (as required by the specification). 5873120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 5874120631Snectar 5875120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 5876120631Snectar when it's 512 *bits* long, not 512 bytes. 5877120631Snectar [Richard Levitte] 5878120631Snectar 5879120631Snectar Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 5880120631Snectar 5881120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 5882120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 5883120631Snectar a protocol version number mismatch like a decryption error 5884120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 5885120631Snectar [Bodo Moeller] 5886120631Snectar 5887120631Snectar *) Turn on RSA blinding by default in the default implementation 5888120631Snectar to avoid a timing attack. Applications that don't want it can call 5889120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 5890120631Snectar They would be ill-advised to do so in most cases. 5891120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 5892120631Snectar 5893120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 5894120631Snectar seeded (in this case, the secret RSA exponent is abused as 5895120631Snectar an unpredictable seed -- if it is not unpredictable, there 5896120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 5897120631Snectar by remembering the creator's thread ID in rsa->blinding and 5898120631Snectar having all other threads use local one-time blinding factors 5899120631Snectar (this requires more computation than sharing rsa->blinding, but 5900120631Snectar avoids excessive locking; and if an RSA object is not shared 5901120631Snectar between threads, blinding will still be very fast). 5902120631Snectar [Bodo Moeller] 5903120631Snectar 5904111147Snectar Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 5905111147Snectar 5906111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 5907111147Snectar via timing by performing a MAC computation even if incorrrect 5908111147Snectar block cipher padding has been found. This is a countermeasure 5909111147Snectar against active attacks where the attacker has to distinguish 5910160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 5911111147Snectar 5912111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 5913111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 5914111147Snectar Martin Vuagnoux (EPFL, Ilion)] 5915111147Snectar 5916109998Smarkm Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 5917109998Smarkm 5918109998Smarkm *) New function OPENSSL_cleanse(), which is used to cleanse a section of 5919109998Smarkm memory from it's contents. This is done with a counter that will 5920109998Smarkm place alternating values in each byte. This can be used to solve 5921109998Smarkm two issues: 1) the removal of calls to memset() by highly optimizing 5922109998Smarkm compilers, and 2) cleansing with other values than 0, since those can 5923109998Smarkm be read through on certain media, for example a swap space on disk. 5924109998Smarkm [Geoff Thorpe] 5925109998Smarkm 5926109998Smarkm *) Bugfix: client side session caching did not work with external caching, 5927109998Smarkm because the session->cipher setting was not restored when reloading 5928109998Smarkm from the external cache. This problem was masked, when 5929109998Smarkm SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 5930109998Smarkm (Found by Steve Haslam <steve@araqnid.ddts.net>.) 5931109998Smarkm [Lutz Jaenicke] 5932109998Smarkm 5933109998Smarkm *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 5934109998Smarkm length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 5935109998Smarkm [Zeev Lieber <zeev-l@yahoo.com>] 5936109998Smarkm 5937109998Smarkm *) Undo an undocumented change introduced in 0.9.6e which caused 5938109998Smarkm repeated calls to OpenSSL_add_all_ciphers() and 5939109998Smarkm OpenSSL_add_all_digests() to be ignored, even after calling 5940109998Smarkm EVP_cleanup(). 5941109998Smarkm [Richard Levitte] 5942109998Smarkm 5943109998Smarkm *) Change the default configuration reader to deal with last line not 5944109998Smarkm being properly terminated. 5945109998Smarkm [Richard Levitte] 5946109998Smarkm 5947109998Smarkm *) Change X509_NAME_cmp() so it applies the special rules on handling 5948109998Smarkm DN values that are of type PrintableString, as well as RDNs of type 5949109998Smarkm emailAddress where the value has the type ia5String. 5950109998Smarkm [stefank@valicert.com via Richard Levitte] 5951109998Smarkm 5952109998Smarkm *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 5953109998Smarkm the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 5954109998Smarkm doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 5955109998Smarkm the bitwise-OR of the two for use by the majority of applications 5956109998Smarkm wanting this behaviour, and update the docs. The documented 5957109998Smarkm behaviour and actual behaviour were inconsistent and had been 5958109998Smarkm changing anyway, so this is more a bug-fix than a behavioural 5959109998Smarkm change. 5960109998Smarkm [Geoff Thorpe, diagnosed by Nadav Har'El] 5961109998Smarkm 5962109998Smarkm *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 5963109998Smarkm (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 5964109998Smarkm [Bodo Moeller] 5965109998Smarkm 5966109998Smarkm *) Fix initialization code race conditions in 5967109998Smarkm SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 5968109998Smarkm SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 5969109998Smarkm SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 5970109998Smarkm TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 5971109998Smarkm ssl2_get_cipher_by_char(), 5972109998Smarkm ssl3_get_cipher_by_char(). 5973109998Smarkm [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 5974109998Smarkm 5975109998Smarkm *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 5976109998Smarkm the cached sessions are flushed, as the remove_cb() might use ex_data 5977109998Smarkm contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 5978109998Smarkm (see [openssl.org #212]). 5979109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 5980109998Smarkm 5981109998Smarkm *) Fix typo in OBJ_txt2obj which incorrectly passed the content 5982109998Smarkm length, instead of the encoding length to d2i_ASN1_OBJECT. 5983109998Smarkm [Steve Henson] 5984109998Smarkm 5985101618Snectar Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 5986101618Snectar 5987101618Snectar *) [In 0.9.6g-engine release:] 5988101618Snectar Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 5989101618Snectar [Lynn Gazis <lgazis@rainbow.com>] 5990101618Snectar 5991101613Snectar Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 5992101613Snectar 5993101613Snectar *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 5994101613Snectar and get fix the header length calculation. 5995101613Snectar [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 5996101613Snectar Alon Kantor <alonk@checkpoint.com> (and others), 5997101613Snectar Steve Henson] 5998101613Snectar 5999101613Snectar *) Use proper error handling instead of 'assertions' in buffer 6000101613Snectar overflow checks added in 0.9.6e. This prevents DoS (the 6001101613Snectar assertions could call abort()). 6002101613Snectar [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 6003101613Snectar 6004100936Snectar Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 6005100936Snectar 6006109998Smarkm *) Add various sanity checks to asn1_get_length() to reject 6007109998Smarkm the ASN1 length bytes if they exceed sizeof(long), will appear 6008109998Smarkm negative or the content length exceeds the length of the 6009109998Smarkm supplied buffer. 6010109998Smarkm [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6011109998Smarkm 6012100936Snectar *) Fix cipher selection routines: ciphers without encryption had no flags 6013100936Snectar for the cipher strength set and where therefore not handled correctly 6014100936Snectar by the selection routines (PR #130). 6015100936Snectar [Lutz Jaenicke] 6016100936Snectar 6017100936Snectar *) Fix EVP_dsa_sha macro. 6018100936Snectar [Nils Larsch] 6019100936Snectar 6020100936Snectar *) New option 6021100936Snectar SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 6022100936Snectar for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 6023100936Snectar that was added in OpenSSL 0.9.6d. 6024100936Snectar 6025100936Snectar As the countermeasure turned out to be incompatible with some 6026100936Snectar broken SSL implementations, the new option is part of SSL_OP_ALL. 6027100936Snectar SSL_OP_ALL is usually employed when compatibility with weird SSL 6028100936Snectar implementations is desired (e.g. '-bugs' option to 's_client' and 6029100936Snectar 's_server'), so the new option is automatically set in many 6030100936Snectar applications. 6031100936Snectar [Bodo Moeller] 6032100936Snectar 6033100936Snectar *) Changes in security patch: 6034100936Snectar 6035100936Snectar Changes marked "(CHATS)" were sponsored by the Defense Advanced 6036100936Snectar Research Projects Agency (DARPA) and Air Force Research Laboratory, 6037100936Snectar Air Force Materiel Command, USAF, under agreement number 6038100936Snectar F30602-01-2-0537. 6039100936Snectar 6040100936Snectar *) Add various sanity checks to asn1_get_length() to reject 6041100936Snectar the ASN1 length bytes if they exceed sizeof(long), will appear 6042100936Snectar negative or the content length exceeds the length of the 6043160814Ssimon supplied buffer. (CVE-2002-0659) 6044100936Snectar [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6045100936Snectar 6046100936Snectar *) Assertions for various potential buffer overflows, not known to 6047100936Snectar happen in practice. 6048100936Snectar [Ben Laurie (CHATS)] 6049100936Snectar 6050100936Snectar *) Various temporary buffers to hold ASCII versions of integers were 6051160814Ssimon too small for 64 bit platforms. (CVE-2002-0655) 6052100936Snectar [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 6053100936Snectar 6054100936Snectar *) Remote buffer overflow in SSL3 protocol - an attacker could 6055160814Ssimon supply an oversized session ID to a client. (CVE-2002-0656) 6056100936Snectar [Ben Laurie (CHATS)] 6057100936Snectar 6058100936Snectar *) Remote buffer overflow in SSL2 protocol - an attacker could 6059160814Ssimon supply an oversized client master key. (CVE-2002-0656) 6060100936Snectar [Ben Laurie (CHATS)] 6061100936Snectar 6062100928Snectar Changes between 0.9.6c and 0.9.6d [9 May 2002] 6063100928Snectar 6064100928Snectar *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 6065100928Snectar encoded as NULL) with id-dsa-with-sha1. 6066100928Snectar [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 6067100928Snectar 6068100928Snectar *) Check various X509_...() return values in apps/req.c. 6069100928Snectar [Nils Larsch <nla@trustcenter.de>] 6070100928Snectar 6071100928Snectar *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 6072100928Snectar an end-of-file condition would erronously be flagged, when the CRLF 6073100928Snectar was just at the end of a processed block. The bug was discovered when 6074100928Snectar processing data through a buffering memory BIO handing the data to a 6075100928Snectar BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 6076100928Snectar <ptsekov@syntrex.com> and Nedelcho Stanev. 6077100928Snectar [Lutz Jaenicke] 6078100928Snectar 6079100928Snectar *) Implement a countermeasure against a vulnerability recently found 6080100928Snectar in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 6081100928Snectar before application data chunks to avoid the use of known IVs 6082100928Snectar with data potentially chosen by the attacker. 6083100928Snectar [Bodo Moeller] 6084100928Snectar 6085100928Snectar *) Fix length checks in ssl3_get_client_hello(). 6086100928Snectar [Bodo Moeller] 6087100928Snectar 6088100928Snectar *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 6089100928Snectar to prevent ssl3_read_internal() from incorrectly assuming that 6090100928Snectar ssl3_read_bytes() found application data while handshake 6091100928Snectar processing was enabled when in fact s->s3->in_read_app_data was 6092100928Snectar merely automatically cleared during the initial handshake. 6093100928Snectar [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 6094100928Snectar 6095100928Snectar *) Fix object definitions for Private and Enterprise: they were not 6096100928Snectar recognized in their shortname (=lowercase) representation. Extend 6097100928Snectar obj_dat.pl to issue an error when using undefined keywords instead 6098100928Snectar of silently ignoring the problem (Svenning Sorensen 6099100928Snectar <sss@sss.dnsalias.net>). 6100100928Snectar [Lutz Jaenicke] 6101100928Snectar 6102100928Snectar *) Fix DH_generate_parameters() so that it works for 'non-standard' 6103100928Snectar generators, i.e. generators other than 2 and 5. (Previously, the 6104100928Snectar code did not properly initialise the 'add' and 'rem' values to 6105100928Snectar BN_generate_prime().) 6106100928Snectar 6107100928Snectar In the new general case, we do not insist that 'generator' is 6108100928Snectar actually a primitive root: This requirement is rather pointless; 6109100928Snectar a generator of the order-q subgroup is just as good, if not 6110100928Snectar better. 6111100928Snectar [Bodo Moeller] 6112100928Snectar 6113100928Snectar *) Map new X509 verification errors to alerts. Discovered and submitted by 6114100928Snectar Tom Wu <tom@arcot.com>. 6115100928Snectar [Lutz Jaenicke] 6116100928Snectar 6117100928Snectar *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 6118100928Snectar returning non-zero before the data has been completely received 6119100928Snectar when using non-blocking I/O. 6120100928Snectar [Bodo Moeller; problem pointed out by John Hughes] 6121100928Snectar 6122100928Snectar *) Some of the ciphers missed the strength entry (SSL_LOW etc). 6123100928Snectar [Ben Laurie, Lutz Jaenicke] 6124100928Snectar 6125100928Snectar *) Fix bug in SSL_clear(): bad sessions were not removed (found by 6126100928Snectar Yoram Zahavi <YoramZ@gilian.com>). 6127100928Snectar [Lutz Jaenicke] 6128100928Snectar 6129100928Snectar *) Add information about CygWin 1.3 and on, and preserve proper 6130100928Snectar configuration for the versions before that. 6131100928Snectar [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 6132100928Snectar 6133100928Snectar *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 6134100928Snectar check whether we deal with a copy of a session and do not delete from 6135100928Snectar the cache in this case. Problem reported by "Izhar Shoshani Levi" 6136100928Snectar <izhar@checkpoint.com>. 6137100928Snectar [Lutz Jaenicke] 6138100928Snectar 6139100928Snectar *) Do not store session data into the internal session cache, if it 6140100928Snectar is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 6141100928Snectar flag is set). Proposed by Aslam <aslam@funk.com>. 6142100928Snectar [Lutz Jaenicke] 6143100928Snectar 6144100928Snectar *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 6145100928Snectar value is 0. 6146100928Snectar [Richard Levitte] 6147100928Snectar 6148109998Smarkm *) [In 0.9.6d-engine release:] 6149109998Smarkm Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 6150100928Snectar [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 6151100928Snectar 6152100928Snectar *) Add the configuration target linux-s390x. 6153100928Snectar [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 6154100928Snectar 6155100928Snectar *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 6156100928Snectar ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 6157100928Snectar variable as an indication that a ClientHello message has been 6158100928Snectar received. As the flag value will be lost between multiple 6159100928Snectar invocations of ssl3_accept when using non-blocking I/O, the 6160100928Snectar function may not be aware that a handshake has actually taken 6161100928Snectar place, thus preventing a new session from being added to the 6162100928Snectar session cache. 6163100928Snectar 6164100928Snectar To avoid this problem, we now set s->new_session to 2 instead of 6165100928Snectar using a local variable. 6166100928Snectar [Lutz Jaenicke, Bodo Moeller] 6167100928Snectar 6168100928Snectar *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 6169100928Snectar if the SSL_R_LENGTH_MISMATCH error is detected. 6170100928Snectar [Geoff Thorpe, Bodo Moeller] 6171100928Snectar 6172100928Snectar *) New 'shared_ldflag' column in Configure platform table. 6173100928Snectar [Richard Levitte] 6174100928Snectar 6175100928Snectar *) Fix EVP_CIPHER_mode macro. 6176100928Snectar ["Dan S. Camper" <dan@bti.net>] 6177100928Snectar 6178100928Snectar *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 6179100928Snectar type, we must throw them away by setting rr->length to 0. 6180100928Snectar [D P Chang <dpc@qualys.com>] 6181100928Snectar 618289837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 618389837Skris 618489837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 618589837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 618689837Skris worked incorrectly for those cases where range = 10..._2 and 618789837Skris 3*range is two bits longer than range.) 618889837Skris [Bodo Moeller] 618989837Skris 619089837Skris *) Only add signing time to PKCS7 structures if it is not already 619189837Skris present. 619289837Skris [Steve Henson] 619389837Skris 619489837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 619589837Skris OBJ_ld_ce should be OBJ_id_ce. 619689837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 619789837Skris incorrect (cf. RFC 3039). 619889837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 619989837Skris 620089837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 620189837Skris returns early because it has nothing to do. 620289837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 620389837Skris 620489837Skris *) [In 0.9.6c-engine release:] 620589837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 620689837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 620789837Skris 620889837Skris *) [In 0.9.6c-engine release:] 620989837Skris Add support for Cryptographic Appliance's keyserver technology. 621089837Skris (Use engine 'keyclient') 621189837Skris [Cryptographic Appliances and Geoff Thorpe] 621289837Skris 621389837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 621489837Skris is called via tools/c89.sh because arguments have to be 621589837Skris rearranged (all '-L' options must appear before the first object 621689837Skris modules). 621789837Skris [Richard Shapiro <rshapiro@abinitio.com>] 621889837Skris 621989837Skris *) [In 0.9.6c-engine release:] 622089837Skris Add support for Broadcom crypto accelerator cards, backported 622189837Skris from 0.9.7. 622289837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 622389837Skris 622489837Skris *) [In 0.9.6c-engine release:] 622589837Skris Add support for SureWare crypto accelerator cards from 622689837Skris Baltimore Technologies. (Use engine 'sureware') 622789837Skris [Baltimore Technologies and Mark Cox] 622889837Skris 622989837Skris *) [In 0.9.6c-engine release:] 623089837Skris Add support for crypto accelerator cards from Accelerated 623189837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 623289837Skris [AEP Inc. and Mark Cox] 623389837Skris 623489837Skris *) Add a configuration entry for gcc on UnixWare. 623589837Skris [Gary Benson <gbenson@redhat.com>] 623689837Skris 623789837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 623889837Skris messages are stored in a single piece (fixed-length part and 623989837Skris variable-length part combined) and fix various bugs found on the way. 624089837Skris [Bodo Moeller] 624189837Skris 624289837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 624389837Skris instead. BIO_gethostbyname() does not know what timeouts are 624489837Skris appropriate, so entries would stay in cache even when they have 624589837Skris become invalid. 624689837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 624789837Skris 624889837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 624989837Skris faced with a pathologically small ClientHello fragment that does 625089837Skris not contain client_version: Instead of aborting with an error, 625189837Skris simply choose the highest available protocol version (i.e., 625289837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 625389837Skris messages are never sent like this, but this change gives us 625489837Skris strictly correct behaviour at least for TLS. 625589837Skris [Bodo Moeller] 625689837Skris 625789837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 625889837Skris never resets s->method to s->ctx->method when called from within 625989837Skris one of the SSL handshake functions. 626089837Skris [Bodo Moeller; problem pointed out by Niko Baric] 626189837Skris 626289837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 626389837Skris (sent using the client's version number) if client_version is 626489837Skris smaller than the protocol version in use. Also change 626589837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 626689837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 626789837Skris the client will at least see that alert. 626889837Skris [Bodo Moeller] 626989837Skris 627089837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 627189837Skris correctly. 627289837Skris [Bodo Moeller] 627389837Skris 627489837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 627589837Skris client receives HelloRequest while in a handshake. 627689837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 627789837Skris 627889837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 627989837Skris should end in 'break', not 'goto end' which circuments various 628089837Skris cleanups done in state SSL_ST_OK. But session related stuff 628189837Skris must be disabled for SSL_ST_OK in the case that we just sent a 628289837Skris HelloRequest. 628389837Skris 628489837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 628589837Skris before just sending a HelloRequest. 628689837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 628789837Skris 628889837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 628989837Skris reveal whether illegal block cipher padding was found or a MAC 629089837Skris verification error occured. (Neither SSLerr() codes nor alerts 629189837Skris are directly visible to potential attackers, but the information 629289837Skris may leak via logfiles.) 629389837Skris 629489837Skris Similar changes are not required for the SSL 2.0 implementation 629589837Skris because the number of padding bytes is sent in clear for SSL 2.0, 629689837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 629789837Skris failed to verify that the purported number of padding bytes is in 629889837Skris the legal range. 629989837Skris [Bodo Moeller] 630089837Skris 630189837Skris *) Add OpenUNIX-8 support including shared libraries 630289837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 630389837Skris [Lutz Jaenicke] 630489837Skris 630589837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 630689837Skris 'wristwatch attack' using huge encoding parameters (cf. 630789837Skris James H. Manger's CRYPTO 2001 paper). Note that the 630889837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 630989837Skris encoding parameters and hence was not vulnerable. 631089837Skris [Bodo Moeller] 631189837Skris 631289837Skris *) BN_sqr() bug fix. 631389837Skris [Ulf M�ller, reported by Jim Ellis <jim.ellis@cavium.com>] 631489837Skris 631589837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 631689837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 631789837Skris followed by modular reduction. 631889837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 631989837Skris 632089837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 632189837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 632289837Skris [Bodo Moeller] 632389837Skris 632489837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 632589837Skris This function was broken, as the check for a new client hello message 632689837Skris to handle SGC did not allow these large messages. 632789837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 632889837Skris [Lutz Jaenicke] 632989837Skris 633089837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 633189837Skris [Lutz Jaenicke] 633289837Skris 633389837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 633489837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 633589837Skris [Lutz Jaenicke] 633689837Skris 633789837Skris *) Rework the configuration and shared library support for Tru64 Unix. 633889837Skris The configuration part makes use of modern compiler features and 633989837Skris still retains old compiler behavior for those that run older versions 634089837Skris of the OS. The shared library support part includes a variant that 634189837Skris uses the RPATH feature, and is available through the special 634289837Skris configuration target "alpha-cc-rpath", which will never be selected 634389837Skris automatically. 634489837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 634589837Skris 634689837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 634789837Skris with the same message size as in ssl3_get_certificate_request(). 634889837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 634989837Skris messages might inadvertently be reject as too long. 635089837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 635189837Skris 635289837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 635389837Skris [Andy Polyakov] 635489837Skris 635589837Skris *) Modified SSL library such that the verify_callback that has been set 635689837Skris specificly for an SSL object with SSL_set_verify() is actually being 635789837Skris used. Before the change, a verify_callback set with this function was 635889837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 635989837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 636089837Skris to allow the necessary settings. 636189837Skris [Lutz Jaenicke] 636289837Skris 636389837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 636489837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 636589837Skris done automatically (in contradiction to the requirements of the C 636689837Skris standard). This made problems when used from OpenSSH. 636789837Skris [Lutz Jaenicke] 636889837Skris 636989837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 637089837Skris dh->length and always used 637189837Skris 637289837Skris BN_rand_range(priv_key, dh->p). 637389837Skris 637489837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 637589837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 637689837Skris dh->length (recommended exponent length) is much smaller than the 637789837Skris length of dh->p. We could use BN_rand_range() if the order of 637889837Skris the subgroup was stored in the DH structure, but we only have 637989837Skris dh->length. 638089837Skris 638189837Skris So switch back to 638289837Skris 638389837Skris BN_rand(priv_key, l, ...) 638489837Skris 638589837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 638689837Skris otherwise. 638789837Skris [Bodo Moeller] 638889837Skris 638989837Skris *) In 639089837Skris 639189837Skris RSA_eay_public_encrypt 639289837Skris RSA_eay_private_decrypt 639389837Skris RSA_eay_private_encrypt (signing) 639489837Skris RSA_eay_public_decrypt (signature verification) 639589837Skris 639689837Skris (default implementations for RSA_public_encrypt, 639789837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 639889837Skris always reject numbers >= n. 639989837Skris [Bodo Moeller] 640089837Skris 640189837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 640289837Skris to synchronize access to 'locking_thread'. This is necessary on 640389837Skris systems where access to 'locking_thread' (an 'unsigned long' 640489837Skris variable) is not atomic. 640589837Skris [Bodo Moeller] 640689837Skris 640789837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 640889837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 640989837Skris a race condition if 0 is a valid thread ID. 641089837Skris [Travis Vitek <vitek@roguewave.com>] 641189837Skris 641289837Skris *) Add support for shared libraries under Irix. 641389837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 641489837Skris 641589837Skris *) Add configuration option to build on Linux on both big-endian and 641689837Skris little-endian MIPS. 641789837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 641889837Skris 641989837Skris *) Add the possibility to create shared libraries on HP-UX. 642089837Skris [Richard Levitte] 642189837Skris 642279998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 642379998Skris 642479998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 642579998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 642679998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 642779998Skris PRNG state recovery was possible based on the output of 642879998Skris one PRNG request appropriately sized to gain knowledge on 642979998Skris 'md' followed by enough consecutive 1-byte PRNG requests 643079998Skris to traverse all of 'state'. 643179998Skris 643279998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 643379998Skris during PRNG output generation, hash all of the previous 643479998Skris 'md_local' value, not just the half used for PRNG output. 643579998Skris 643679998Skris 2. Make the number of bytes from 'state' included into the hash 643779998Skris independent from the number of PRNG bytes requested. 643879998Skris 643979998Skris The first measure alone would be sufficient to avoid 644079998Skris Markku-Juhani's attack. (Actually it had never occurred 644179998Skris to me that the half of 'md_local' used for chaining was the 644279998Skris half from which PRNG output bytes were taken -- I had always 644379998Skris assumed that the secret half would be used.) The second 644479998Skris measure makes sure that additional data from 'state' is never 644579998Skris mixed into 'md_local' in small portions; this heuristically 644679998Skris further strengthens the PRNG. 644779998Skris [Bodo Moeller] 644879998Skris 644979998Skris *) Fix crypto/bn/asm/mips3.s. 645079998Skris [Andy Polyakov] 645179998Skris 645279998Skris *) When only the key is given to "enc", the IV is undefined. Print out 645379998Skris an error message in this case. 645479998Skris [Lutz Jaenicke] 645579998Skris 645679998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 645779998Skris [Steve Henson] 645879998Skris 645979998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 646079998Skris positive and less than q. 646179998Skris [Bodo Moeller] 646279998Skris 646379998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 646479998Skris used: it isn't thread safe and the add_lock_callback should handle 646579998Skris that itself. 646679998Skris [Paul Rose <Paul.Rose@bridge.com>] 646779998Skris 646879998Skris *) Verify that incoming data obeys the block size in 646979998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 647079998Skris [Bodo Moeller] 647179998Skris 647279998Skris *) Fix OAEP check. 647379998Skris [Ulf M�ller, Bodo M�ller] 647479998Skris 647579998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 647689837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 647779998Skris when fixing the server behaviour for backwards-compatible 'client 647879998Skris hello' messages. (Note that the attack is impractical against 647979998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 648079998Skris means that the probability of guessing a valid ciphertext is 648179998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 648279998Skris paper.) 648379998Skris 648479998Skris Before 0.9.5, the countermeasure (hide the error by generating a 648579998Skris random 'decryption result') did not work properly because 648679998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 648779998Skris detect the supposedly ignored error. 648879998Skris 648979998Skris Both problems are now fixed. 649079998Skris [Bodo Moeller] 649179998Skris 649279998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 649379998Skris (previously it was 1024). 649479998Skris [Bodo Moeller] 649579998Skris 649679998Skris *) Fix for compatibility mode trust settings: ignore trust settings 649779998Skris unless some valid trust or reject settings are present. 649879998Skris [Steve Henson] 649979998Skris 650079998Skris *) Fix for blowfish EVP: its a variable length cipher. 650179998Skris [Steve Henson] 650279998Skris 650379998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 650479998Skris parameters in DSA public key structures and return an error in the 650579998Skris DSA routines if parameters are absent. 650679998Skris [Steve Henson] 650779998Skris 650879998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 650979998Skris in the current directory if neither $RANDFILE nor $HOME was set. 651079998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 651179998Skris caused some confusion to Windows users who haven't defined $HOME. 651279998Skris Thus RAND_file_name() is changed again: e_os.h can define a 651379998Skris DEFAULT_HOME, which will be used if $HOME is not set. 651479998Skris For Windows, we use "C:"; on other platforms, we still require 651579998Skris environment variables. 651679998Skris 651779998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 651879998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 651979998Skris having multiple threads call RAND_poll() concurrently. 652079998Skris [Bodo Moeller] 652179998Skris 652279998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 652379998Skris combination of a flag and a thread ID variable. 652479998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 652579998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 652689837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 652779998Skris that they do not hold after the first thread unsets add_do_not_lock). 652879998Skris [Bodo Moeller] 652979998Skris 653079998Skris *) Change bctest again: '-x' expressions are not available in all 653179998Skris versions of 'test'. 653279998Skris [Bodo Moeller] 653379998Skris 653476866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 653572613Skris 653676866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 653776866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 653876866Skris 653976866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 654076866Skris the default extension for executables, if any. Also, make the perl 654176866Skris scripts that use symlink() to test if it really exists and use "cp" 654276866Skris if it doesn't. All this made OpenSSL compilable and installable in 654376866Skris CygWin. 654476866Skris [Richard Levitte] 654576866Skris 654676866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 654776866Skris If SEQUENCE is length is indefinite just set c->slen to the total 654876866Skris amount of data available. 654976866Skris [Steve Henson, reported by shige@FreeBSD.org] 655076866Skris [This change does not apply to 0.9.7.] 655176866Skris 655276866Skris *) Change bctest to avoid here-documents inside command substitution 655376866Skris (workaround for FreeBSD /bin/sh bug). 655476866Skris For compatibility with Ultrix, avoid shell functions (introduced 655576866Skris in the bctest version that searches along $PATH). 655676866Skris [Bodo Moeller] 655776866Skris 655876866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 655976866Skris with des_encrypt() defined on some operating systems, like Solaris 656076866Skris and UnixWare. 656176866Skris [Richard Levitte] 656276866Skris 656376866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 656476866Skris On the Importance of Eliminating Errors in Cryptographic 656576866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 656676866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 656776866Skris [Ulf Moeller] 656876866Skris 656976866Skris *) MIPS assembler BIGNUM division bug fix. 657076866Skris [Andy Polyakov] 657176866Skris 657276866Skris *) Disabled incorrect Alpha assembler code. 657376866Skris [Richard Levitte] 657476866Skris 657576866Skris *) Fix PKCS#7 decode routines so they correctly update the length 657676866Skris after reading an EOC for the EXPLICIT tag. 657776866Skris [Steve Henson] 657876866Skris [This change does not apply to 0.9.7.] 657976866Skris 658076866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 658176866Skris if a 3DES key was generated with a 0 initial byte. Include 658276866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 658376866Skris (but broken) behaviour. 658476866Skris [Steve Henson] 658576866Skris 658676866Skris *) Enhance bctest to search for a working bc along $PATH and print 658776866Skris it when found. 658876866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 658976866Skris 659076866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 659176866Skris don't write to the wrong index in ERR_set_error_data. 659276866Skris [Bodo Moeller] 659376866Skris 659476866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 659576866Skris did not exist. 659676866Skris [Bodo Moeller] 659776866Skris 659876866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 659976866Skris [Jeremy Cooper <jeremy@baymoo.org>] 660076866Skris 660176866Skris *) Make it possible to reuse SSLv2 sessions. 660276866Skris [Richard Levitte] 660376866Skris 660476866Skris *) In copy_email() check for >= 0 as a return value for 660576866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 660676866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 660776866Skris 660876866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 660976866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 661076866Skris PKCS7_verify() fails with non detached data. 661176866Skris [Steve Henson] 661276866Skris 661376866Skris *) Don't use getenv in library functions when run as setuid/setgid. 661476866Skris New function OPENSSL_issetugid(). 661576866Skris [Ulf Moeller] 661676866Skris 661776866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 661876866Skris due to incorrect handling of multi-threading: 661976866Skris 662076866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 662176866Skris 662276866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 662376866Skris 662476866Skris 3. Count how many times MemCheck_off() has been called so that 662576866Skris nested use can be treated correctly. This also avoids 662676866Skris inband-signalling in the previous code (which relied on the 662776866Skris assumption that thread ID 0 is impossible). 662876866Skris [Bodo Moeller] 662976866Skris 663076866Skris *) Add "-rand" option also to s_client and s_server. 663176866Skris [Lutz Jaenicke] 663276866Skris 663376866Skris *) Fix CPU detection on Irix 6.x. 663476866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 663576866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 663676866Skris 663776866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 663876866Skris was empty. 663976866Skris [Steve Henson] 664076866Skris [This change does not apply to 0.9.7.] 664176866Skris 664276866Skris *) Use the cached encoding of an X509_NAME structure rather than 664376866Skris copying it. This is apparently the reason for the libsafe "errors" 664476866Skris but the code is actually correct. 664576866Skris [Steve Henson] 664676866Skris 664772613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 664872613Skris Bleichenbacher's DSA attack. 664976866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 665076866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 665176866Skris and leaves the highest bit random. 665276866Skris [Ulf Moeller, Bodo Moeller] 665372613Skris 665472613Skris *) In the NCONF_...-based implementations for CONF_... queries 665572613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 665672613Skris a temporary CONF structure with the data component set to NULL 665772613Skris (which gives segmentation faults in lh_retrieve). 665872613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 665972613Skris CONF_get_number (which may use environment variables) and directly 666072613Skris return NULL from CONF_get_section. 666172613Skris [Bodo Moeller] 666272613Skris 666372613Skris *) Fix potential buffer overrun for EBCDIC. 666472613Skris [Ulf Moeller] 666572613Skris 666672613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 666772613Skris keyUsage if basicConstraints absent for a CA. 666872613Skris [Steve Henson] 666972613Skris 667072613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 667172613Skris is more generally accepted (no spaces before the semicolon), since 667272613Skris some programs can't parse those values properly otherwise. Also make 667372613Skris sure BIO's that break lines after each write do not create invalid 667472613Skris headers. 667572613Skris [Richard Levitte] 667672613Skris 667772613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 667872613Skris macros previously used would not encode an empty SEQUENCE OF 667972613Skris and break the signature. 668072613Skris [Steve Henson] 668176866Skris [This change does not apply to 0.9.7.] 668272613Skris 668372613Skris *) Zero the premaster secret after deriving the master secret in 668472613Skris DH ciphersuites. 668572613Skris [Steve Henson] 668672613Skris 668772613Skris *) Add some EVP_add_digest_alias registrations (as found in 668872613Skris OpenSSL_add_all_digests()) to SSL_library_init() 668972613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 669072613Skris compatibility with peers using X.509 certificates 669172613Skris with unconventional AlgorithmIdentifier OIDs. 669272613Skris [Bodo Moeller] 669372613Skris 669472613Skris *) Fix for Irix with NO_ASM. 669572613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 669672613Skris 669772613Skris *) ./config script fixes. 669872613Skris [Ulf Moeller, Richard Levitte] 669972613Skris 670072613Skris *) Fix 'openssl passwd -1'. 670172613Skris [Bodo Moeller] 670272613Skris 670372613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 670472613Skris terminated strings whose length is passed in the passlen 670572613Skris parameter, for example from PEM callbacks. This was done 670672613Skris by adding an extra length parameter to asc2uni(). 670772613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 670872613Skris 670972613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 671072613Skris call failed, free the DSA structure. 671172613Skris [Bodo Moeller] 671272613Skris 671372613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 671472613Skris These are present in some PKCS#12 files. 671572613Skris [Steve Henson] 671672613Skris 671772613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 671872613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 671972613Skris when writing a 32767 byte record. 672072613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 672172613Skris 672272613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 672372613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 672472613Skris 672572613Skris (RSA objects have a reference count access to which is protected 672672613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 672772613Skris so they are meant to be shared between threads.) 672872613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 672972613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 673072613Skris 673172613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 673272613Skris [Bodo Moeller] 673372613Skris 673476866Skris *) Use better test patterns in bntest. 673576866Skris [Ulf M�ller] 673676866Skris 673772613Skris *) rand_win.c fix for Borland C. 673872613Skris [Ulf M�ller] 673972613Skris 674072613Skris *) BN_rshift bugfix for n == 0. 674172613Skris [Bodo Moeller] 674272613Skris 674376866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 674476866Skris so that 'make test' does not abort just because 'bc' is broken. 674576866Skris [Bodo Moeller] 674676866Skris 674772613Skris *) Store verify_result within SSL_SESSION also for client side to 674872613Skris avoid potential security hole. (Re-used sessions on the client side 674972613Skris always resulted in verify_result==X509_V_OK, not using the original 675072613Skris result of the server certificate verification.) 675172613Skris [Lutz Jaenicke] 675272613Skris 675372613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 675472613Skris SSL3_RT_APPLICATION_DATA, return 0. 675572613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 675672613Skris [Bodo Moeller] 675772613Skris 675872613Skris *) Fix SSL_peek: 675972613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 676072613Skris releases, have been re-implemented by renaming the previous 676172613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 676272613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 676372613Skris to them. The new ssl[23]_{read,peek} functions are calls to 676472613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 676572613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 676672613Skris does the actual work for ssl3_read_internal. 676772613Skris [Bodo Moeller] 676872613Skris 676976866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 677076866Skris the method-specific "init()" handler. Also clean up ex_data after 677176866Skris calling the method-specific "finish()" handler. Previously, this was 677276866Skris happening the other way round. 677376866Skris [Geoff Thorpe] 677476866Skris 677572613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 677672613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 677772613Skris [Bodo Moeller] 677872613Skris 677976866Skris *) Make sure that shared libraries get the internal name engine with 678076866Skris the full version number and not just 0. This should mark the 678176866Skris shared libraries as not backward compatible. Of course, this should 678276866Skris be changed again when we can guarantee backward binary compatibility. 678376866Skris [Richard Levitte] 678476866Skris 678572613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 678672613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 678772613Skris 678876866Skris *) Rework the system to generate shared libraries: 678976866Skris 679076866Skris - Make note of the expected extension for the shared libraries and 679176866Skris if there is a need for symbolic links from for example libcrypto.so.0 679276866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 679376866Skris that. 679476866Skris 679576866Skris - Make as few rebuilds of the shared libraries as possible. 679676866Skris 679776866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 679876866Skris 679976866Skris - When installing, install the shared libraries separately from the 680076866Skris static ones. 680176866Skris [Richard Levitte] 680276866Skris 680372613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 680472613Skris 680572613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 680672613Skris and not in SSL_clear because the latter is also used by the 680772613Skris accept/connect functions; previously, the settings made by 680872613Skris SSL_set_read_ahead would be lost during the handshake. 680972613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 681072613Skris 681172613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 681272613Skris Previously, it would create entries for disableed algorithms no 681372613Skris matter what. 681472613Skris [Richard Levitte] 681572613Skris 681672613Skris *) Added several new manual pages for SSL_* function. 681772613Skris [Lutz Jaenicke] 681872613Skris 681968651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 682068651Skris 682168651Skris *) In ssl23_get_client_hello, generate an error message when faced 682268651Skris with an initial SSL 3.0/TLS record that is too small to contain the 682368651Skris first two bytes of the ClientHello message, i.e. client_version. 682468651Skris (Note that this is a pathologic case that probably has never happened 682568651Skris in real life.) The previous approach was to use the version number 682668651Skris from the record header as a substitute; but our protocol choice 682768651Skris should not depend on that one because it is not authenticated 682868651Skris by the Finished messages. 682968651Skris [Bodo Moeller] 683068651Skris 683168651Skris *) More robust randomness gathering functions for Windows. 683268651Skris [Jeffrey Altman <jaltman@columbia.edu>] 683368651Skris 683468651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 683568651Skris not set then we don't setup the error code for issuer check errors 683668651Skris to avoid possibly overwriting other errors which the callback does 683768651Skris handle. If an application does set the flag then we assume it knows 683868651Skris what it is doing and can handle the new informational codes 683968651Skris appropriately. 684068651Skris [Steve Henson] 684168651Skris 684268651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 684368651Skris a general "ANY" type, as such it should be able to decode anything 684468651Skris including tagged types. However it didn't check the class so it would 684568651Skris wrongly interpret tagged types in the same way as their universal 684668651Skris counterpart and unknown types were just rejected. Changed so that the 684768651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 684868651Skris that is the encoding is stored intact. There is also a new type 684968651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 685068651Skris case we have no idea what the actual type is so we just lump them all 685168651Skris together. 685268651Skris [Steve Henson] 685368651Skris 685468651Skris *) On VMS, stdout may very well lead to a file that is written to 685568651Skris in a record-oriented fashion. That means that every write() will 685668651Skris write a separate record, which will be read separately by the 685768651Skris programs trying to read from it. This can be very confusing. 685868651Skris 685968651Skris The solution is to put a BIO filter in the way that will buffer 686068651Skris text until a linefeed is reached, and then write everything a 686168651Skris line at a time, so every record written will be an actual line, 686268651Skris not chunks of lines and not (usually doesn't happen, but I've 686368651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 686468651Skris the answer. 686568651Skris 686668651Skris Currently, it's a VMS-only method, because that's where it has 686768651Skris been tested well enough. 686868651Skris [Richard Levitte] 686968651Skris 687068651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 687168651Skris it can return incorrect results. 687268651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 687368651Skris but it was in 0.9.6-beta[12].) 687468651Skris [Bodo Moeller] 687568651Skris 687668651Skris *) Disable the check for content being present when verifying detached 687768651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 687868651Skris include zero length content when signing messages. 687968651Skris [Steve Henson] 688068651Skris 688168651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 688268651Skris BIO_ctrl (for BIO pairs). 688368651Skris [Bodo M�ller] 688468651Skris 688568651Skris *) Add DSO method for VMS. 688668651Skris [Richard Levitte] 688768651Skris 688868651Skris *) Bug fix: Montgomery multiplication could produce results with the 688968651Skris wrong sign. 689068651Skris [Ulf M�ller] 689168651Skris 689268651Skris *) Add RPM specification openssl.spec and modify it to build three 689368651Skris packages. The default package contains applications, application 689468651Skris documentation and run-time libraries. The devel package contains 689568651Skris include files, static libraries and function documentation. The 689668651Skris doc package contains the contents of the doc directory. The original 689768651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 689868651Skris [Richard Levitte] 689968651Skris 690068651Skris *) Add a large number of documentation files for many SSL routines. 690168651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 690268651Skris 690368651Skris *) Add a configuration entry for Sony News 4. 690468651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 690568651Skris 690668651Skris *) Don't set the two most significant bits to one when generating a 690768651Skris random number < q in the DSA library. 690868651Skris [Ulf M�ller] 690968651Skris 691068651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 691168651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 691268651Skris the underlying transport is blocking) if a handshake took place. 691368651Skris (The default behaviour is needed by applications such as s_client 691468651Skris and s_server that use select() to determine when to use SSL_read; 691568651Skris but for applications that know in advance when to expect data, it 691668651Skris just makes things more complicated.) 691768651Skris [Bodo Moeller] 691868651Skris 691968651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 692068651Skris from EGD. 692168651Skris [Ben Laurie] 692268651Skris 692368651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 692468651Skris work better on such systems. 692568651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 692668651Skris 692768651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 692868651Skris Update PKCS12_parse() so it copies the friendlyName and the 692968651Skris keyid to the certificates aux info. 693068651Skris [Steve Henson] 693168651Skris 693268651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 693368651Skris if there was more than one signature. 693468651Skris [Sven Uszpelkat <su@celocom.de>] 693568651Skris 693668651Skris *) Major change in util/mkdef.pl to include extra information 693768651Skris about each symbol, as well as presentig variables as well 693868651Skris as functions. This change means that there's n more need 693968651Skris to rebuild the .num files when some algorithms are excluded. 694068651Skris [Richard Levitte] 694168651Skris 694268651Skris *) Allow the verify time to be set by an application, 694368651Skris rather than always using the current time. 694468651Skris [Steve Henson] 694568651Skris 694668651Skris *) Phase 2 verify code reorganisation. The certificate 694768651Skris verify code now looks up an issuer certificate by a 694868651Skris number of criteria: subject name, authority key id 694968651Skris and key usage. It also verifies self signed certificates 695068651Skris by the same criteria. The main comparison function is 695168651Skris X509_check_issued() which performs these checks. 695268651Skris 695368651Skris Lot of changes were necessary in order to support this 695468651Skris without completely rewriting the lookup code. 695568651Skris 695668651Skris Authority and subject key identifier are now cached. 695768651Skris 695868651Skris The LHASH 'certs' is X509_STORE has now been replaced 695968651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 696068651Skris LHASH can't store or retrieve multiple objects with 696168651Skris the same hash value. 696268651Skris 696368651Skris As a result various functions (which were all internal 696468651Skris use only) have changed to handle the new X509_STORE 696568651Skris structure. This will break anything that messed round 696668651Skris with X509_STORE internally. 696768651Skris 696868651Skris The functions X509_STORE_add_cert() now checks for an 696968651Skris exact match, rather than just subject name. 697068651Skris 697168651Skris The X509_STORE API doesn't directly support the retrieval 697268651Skris of multiple certificates matching a given criteria, however 697368651Skris this can be worked round by performing a lookup first 697468651Skris (which will fill the cache with candidate certificates) 697568651Skris and then examining the cache for matches. This is probably 697668651Skris the best we can do without throwing out X509_LOOKUP 697768651Skris entirely (maybe later...). 697868651Skris 697968651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 698068651Skris 698168651Skris All certificate lookup operations now go via a get_issuer() 698268651Skris callback. Although this currently uses an X509_STORE it 698368651Skris can be replaced by custom lookups. This is a simple way 698468651Skris to bypass the X509_STORE hackery necessary to make this 698568651Skris work and makes it possible to use more efficient techniques 698668651Skris in future. A very simple version which uses a simple 698768651Skris STACK for its trusted certificate store is also provided 698868651Skris using X509_STORE_CTX_trusted_stack(). 698968651Skris 699068651Skris The verify_cb() and verify() callbacks now have equivalents 699168651Skris in the X509_STORE_CTX structure. 699268651Skris 699368651Skris X509_STORE_CTX also has a 'flags' field which can be used 699468651Skris to customise the verify behaviour. 699568651Skris [Steve Henson] 699668651Skris 699768651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 699868651Skris excludes S/MIME capabilities. 699968651Skris [Steve Henson] 700068651Skris 700168651Skris *) When a certificate request is read in keep a copy of the 700268651Skris original encoding of the signed data and use it when outputing 700368651Skris again. Signatures then use the original encoding rather than 700468651Skris a decoded, encoded version which may cause problems if the 700568651Skris request is improperly encoded. 700668651Skris [Steve Henson] 700768651Skris 700868651Skris *) For consistency with other BIO_puts implementations, call 700968651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 701068651Skris BIO_write(b, ...). 701168651Skris 701268651Skris In BIO_puts, increment b->num_write as in BIO_write. 701368651Skris [Peter.Sylvester@EdelWeb.fr] 701468651Skris 701568651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 701668651Skris BN_zero, we may not return a BIGNUM with an array consisting of 701768651Skris words set to zero.) 701868651Skris [Bodo Moeller] 701968651Skris 702068651Skris *) Avoid calling abort() from within the library when problems are 702168651Skris detected, except if preprocessor symbols have been defined 702268651Skris (such as REF_CHECK, BN_DEBUG etc.). 702368651Skris [Bodo Moeller] 702468651Skris 702568651Skris *) New openssl application 'rsautl'. This utility can be 702668651Skris used for low level RSA operations. DER public key 702768651Skris BIO/fp routines also added. 702868651Skris [Steve Henson] 702968651Skris 703068651Skris *) New Configure entry and patches for compiling on QNX 4. 703168651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 703268651Skris 703368651Skris *) A demo state-machine implementation was sponsored by 703468651Skris Nuron (http://www.nuron.com/) and is now available in 703568651Skris demos/state_machine. 703668651Skris [Ben Laurie] 703768651Skris 703868651Skris *) New options added to the 'dgst' utility for signature 703968651Skris generation and verification. 704068651Skris [Steve Henson] 704168651Skris 704268651Skris *) Unrecognized PKCS#7 content types are now handled via a 704368651Skris catch all ASN1_TYPE structure. This allows unsupported 704468651Skris types to be stored as a "blob" and an application can 704568651Skris encode and decode it manually. 704668651Skris [Steve Henson] 704768651Skris 704868651Skris *) Fix various signed/unsigned issues to make a_strex.c 704968651Skris compile under VC++. 705068651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 705168651Skris 705268651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 705368651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 705468651Skris if passed a NULL BN and its argument was negative. 705568651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 705668651Skris 705768651Skris *) Modification to PKCS#7 encoding routines to output definite 705868651Skris length encoding. Since currently the whole structures are in 705968651Skris memory there's not real point in using indefinite length 706068651Skris constructed encoding. However if OpenSSL is compiled with 706168651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 706268651Skris [Steve Henson] 706368651Skris 706468651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 706568651Skris [Richard Levitte] 706668651Skris 706768651Skris *) Added more prefixes to parse for in the the strings written 706868651Skris through a logging bio, to cover all the levels that are available 706968651Skris through syslog. The prefixes are now: 707068651Skris 707168651Skris PANIC, EMERG, EMR => LOG_EMERG 707268651Skris ALERT, ALR => LOG_ALERT 707368651Skris CRIT, CRI => LOG_CRIT 707468651Skris ERROR, ERR => LOG_ERR 707568651Skris WARNING, WARN, WAR => LOG_WARNING 707668651Skris NOTICE, NOTE, NOT => LOG_NOTICE 707768651Skris INFO, INF => LOG_INFO 707868651Skris DEBUG, DBG => LOG_DEBUG 707968651Skris 708068651Skris and as before, if none of those prefixes are present at the 708168651Skris beginning of the string, LOG_ERR is chosen. 708268651Skris 708368651Skris On Win32, the LOG_* levels are mapped according to this: 708468651Skris 708568651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 708668651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 708768651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 708868651Skris 708968651Skris [Richard Levitte] 709068651Skris 709168651Skris *) Made it possible to reconfigure with just the configuration 709268651Skris argument "reconf" or "reconfigure". The command line arguments 709368651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 709468651Skris and are retrieved from there when reconfiguring. 709568651Skris [Richard Levitte] 709668651Skris 709768651Skris *) MD4 implemented. 709868651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 709968651Skris 710068651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 710168651Skris [Richard Levitte] 710268651Skris 710368651Skris *) The obj_dat.pl script was messing up the sorting of object 710468651Skris names. The reason was that it compared the quoted version 710568651Skris of strings as a result "OCSP" > "OCSP Signing" because 710668651Skris " > SPACE. Changed script to store unquoted versions of 710768651Skris names and add quotes on output. It was also omitting some 710868651Skris names from the lookup table if they were given a default 710968651Skris value (that is if SN is missing it is given the same 711068651Skris value as LN and vice versa), these are now added on the 711168651Skris grounds that if an object has a name we should be able to 711268651Skris look it up. Finally added warning output when duplicate 711368651Skris short or long names are found. 711468651Skris [Steve Henson] 711568651Skris 711668651Skris *) Changes needed for Tandem NSK. 711768651Skris [Scott Uroff <scott@xypro.com>] 711868651Skris 711968651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 712068651Skris RSA_padding_check_SSLv23(), special padding was never detected 712168651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 712268651Skris version rollback attacks was not effective. 712368651Skris 712468651Skris In s23_clnt.c, don't use special rollback-attack detection padding 712568651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 712668651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 712768651Skris SSL 2.0 is the only protocol enabled in the server. 712868651Skris [Bodo Moeller] 712968651Skris 713068651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 713168651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 713268651Skris BIO_dump_indent() are added. 713368651Skris [Richard Levitte] 713468651Skris 713568651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 713668651Skris these print out strings and name structures based on various 713768651Skris flags including RFC2253 support and proper handling of 713868651Skris multibyte characters. Added options to the 'x509' utility 713968651Skris to allow the various flags to be set. 714068651Skris [Steve Henson] 714168651Skris 714268651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 714368651Skris Also change the functions X509_cmp_current_time() and 714468651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 714568651Skris this will enable certificates using GeneralizedTime in validity 714668651Skris dates to be checked. 714768651Skris [Steve Henson] 714868651Skris 714968651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 715068651Skris negative public key encodings) on by default, 715168651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 715268651Skris [Steve Henson] 715368651Skris 715468651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 715568651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 715668651Skris the encoding can be trivially obtained from the structure. 715768651Skris [Steve Henson] 715868651Skris 715968651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 716068651Skris not read locks (CRYPTO_r_[un]lock). 716168651Skris [Bodo Moeller] 716268651Skris 716368651Skris *) A first attempt at creating official support for shared 716468651Skris libraries through configuration. I've kept it so the 716568651Skris default is static libraries only, and the OpenSSL programs 716668651Skris are always statically linked for now, but there are 716768651Skris preparations for dynamic linking in place. 716889837Skris This has been tested on Linux and Tru64. 716968651Skris [Richard Levitte] 717068651Skris 717168651Skris *) Randomness polling function for Win9x, as described in: 717268651Skris Peter Gutmann, Software Generation of Practically Strong 717368651Skris Random Numbers. 717468651Skris [Ulf M�ller] 717568651Skris 717668651Skris *) Fix so PRNG is seeded in req if using an already existing 717768651Skris DSA key. 717868651Skris [Steve Henson] 717968651Skris 718068651Skris *) New options to smime application. -inform and -outform 718168651Skris allow alternative formats for the S/MIME message including 718268651Skris PEM and DER. The -content option allows the content to be 718368651Skris specified separately. This should allow things like Netscape 718468651Skris form signing output easier to verify. 718568651Skris [Steve Henson] 718668651Skris 718768651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 718868651Skris [Steve Henson] 718968651Skris 719068651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 719168651Skris STRING types. These convert content octets to and from the 719268651Skris underlying type. The actual tag and length octets are 719368651Skris already assumed to have been read in and checked. These 719468651Skris are needed because all other string types have virtually 719568651Skris identical handling apart from the tag. By having versions 719668651Skris of the ASN1 functions that just operate on content octets 719768651Skris IMPLICIT tagging can be handled properly. It also allows 719868651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 719968651Skris and ASN1_INTEGER are identical apart from the tag. 720068651Skris [Steve Henson] 720168651Skris 720268651Skris *) Change the handling of OID objects as follows: 720368651Skris 720468651Skris - New object identifiers are inserted in objects.txt, following 720568651Skris the syntax given in objects.README. 720668651Skris - objects.pl is used to process obj_mac.num and create a new 720768651Skris obj_mac.h. 720868651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 720968651Skris obj_mac.h. 721068651Skris 721168651Skris This is currently kind of a hack, and the perl code in objects.pl 721268651Skris isn't very elegant, but it works as I intended. The simplest way 721368651Skris to check that it worked correctly is to look in obj_dat.h and 721468651Skris check the array nid_objs and make sure the objects haven't moved 721568651Skris around (this is important!). Additions are OK, as well as 721668651Skris consistent name changes. 721768651Skris [Richard Levitte] 721868651Skris 721968651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 722068651Skris [Bodo Moeller] 722168651Skris 722268651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 722368651Skris The given file adds to whatever has already been seeded into the 722468651Skris random pool through the RANDFILE configuration file option or 722568651Skris environment variable, or the default random state file. 722668651Skris [Richard Levitte] 722768651Skris 722868651Skris *) mkstack.pl now sorts each macro group into lexical order. 722968651Skris Previously the output order depended on the order the files 723068651Skris appeared in the directory, resulting in needless rewriting 723168651Skris of safestack.h . 723268651Skris [Steve Henson] 723368651Skris 723468651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 723568651Skris work arounds for the VC++ problem that it treats func() as 723668651Skris func(void). Also stripped out the parts of mkdef.pl that 723768651Skris added extra typesafe functions: these no longer exist. 723868651Skris [Steve Henson] 723968651Skris 724068651Skris *) Reorganisation of the stack code. The macros are now all 724168651Skris collected in safestack.h . Each macro is defined in terms of 724268651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 724368651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 724468651Skris this has the advantage of retaining type safety without the 724568651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 724668651Skris then the non typesafe macros are used instead. Also modified the 724768651Skris mkstack.pl script to handle the new form. Needs testing to see 724868651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 724968651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 725068651Skris and PKCS12_STACK_OF. 725168651Skris [Steve Henson] 725268651Skris 725368651Skris *) When some versions of IIS use the 'NET' form of private key the 725468651Skris key derivation algorithm is different. Normally MD5(password) is 725568651Skris used as a 128 bit RC4 key. In the modified case 725668651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 725768651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 725868651Skris as the old Netscape_RSA functions except they have an additional 725968651Skris 'sgckey' parameter which uses the modified algorithm. Also added 726068651Skris an -sgckey command line option to the rsa utility. Thanks to 726168651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 726268651Skris algorithm to openssl-dev. 726368651Skris [Steve Henson] 726468651Skris 726568651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 726668651Skris invalid expansion on some systems (SCO 5.0.5 for example). 726768651Skris Corrected to 'c.kname'. 726868651Skris [Phillip Porch <root@theporch.com>] 726968651Skris 727068651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 727168651Skris a STACK of email addresses from a certificate or request, these look 727268651Skris in the subject name and the subject alternative name extensions and 727368651Skris omit any duplicate addresses. 727468651Skris [Steve Henson] 727568651Skris 727668651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 727768651Skris This makes DSA verification about 2 % faster. 727868651Skris [Bodo Moeller] 727968651Skris 728068651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 728168651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 728268651Skris plus overhead for 1024 bit moduli). 728368651Skris This makes exponentiations about 0.5 % faster for 1024 bit 728468651Skris exponents (as measured by "openssl speed rsa2048"). 728568651Skris [Bodo Moeller] 728668651Skris 728768651Skris *) Rename memory handling macros to avoid conflicts with other 728868651Skris software: 728968651Skris Malloc => OPENSSL_malloc 729068651Skris Malloc_locked => OPENSSL_malloc_locked 729168651Skris Realloc => OPENSSL_realloc 729268651Skris Free => OPENSSL_free 729368651Skris [Richard Levitte] 729468651Skris 729568651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 729668651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 729768651Skris [Bodo Moeller] 729868651Skris 729968651Skris *) CygWin32 support. 730068651Skris [John Jarvie <jjarvie@newsguy.com>] 730168651Skris 730268651Skris *) The type-safe stack code has been rejigged. It is now only compiled 730368651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 730468651Skris by default all type-specific stack functions are "#define"d back to 730568651Skris standard stack functions. This results in more streamlined output 730668651Skris but retains the type-safety checking possibilities of the original 730768651Skris approach. 730868651Skris [Geoff Thorpe] 730968651Skris 731068651Skris *) The STACK code has been cleaned up, and certain type declarations 731168651Skris that didn't make a lot of sense have been brought in line. This has 731268651Skris also involved a cleanup of sorts in safestack.h to more correctly 731368651Skris map type-safe stack functions onto their plain stack counterparts. 731468651Skris This work has also resulted in a variety of "const"ifications of 731568651Skris lots of the code, especially "_cmp" operations which should normally 731668651Skris be prototyped with "const" parameters anyway. 731768651Skris [Geoff Thorpe] 731868651Skris 731968651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 732068651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 732168651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 732268651Skris where all of 'md' is used each time the PRNG is used, but 'state' 732368651Skris is used only indexed by a cyclic counter. As entropy may not be 732468651Skris well distributed from the beginning, 'md' is important as a 732568651Skris chaining variable. However, the output function chains only half 732668651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 732768651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 732868651Skris in all of 'state' being rewritten, with the new values depending 732968651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 733068651Skris [Bodo Moeller] 733168651Skris 733268651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 733368651Skris the handshake is continued after ssl_verify_cert_chain(); 733468651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 733568651Skris can lead to 'unexplainable' connection aborts later. 733668651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 733768651Skris 733868651Skris *) Major EVP API cipher revision. 733968651Skris Add hooks for extra EVP features. This allows various cipher 734068651Skris parameters to be set in the EVP interface. Support added for variable 734168651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 734268651Skris setting of RC2 and RC5 parameters. 734368651Skris 734468651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 734568651Skris ciphers. 734668651Skris 734768651Skris Remove lots of duplicated code from the EVP library. For example *every* 734868651Skris cipher init() function handles the 'iv' in the same way according to the 734968651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 735068651Skris for CFB and OFB modes they zero ctx->num. 735168651Skris 735268651Skris New functionality allows removal of S/MIME code RC2 hack. 735368651Skris 735468651Skris Most of the routines have the same form and so can be declared in terms 735568651Skris of macros. 735668651Skris 735768651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 735868651Skris all individual ciphers. If the cipher wants to handle IVs or keys 735968651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 736068651Skris flags. 736168651Skris 736268651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 736368651Skris value: although software versions of the algorithms cannot fail 736468651Skris any installed hardware versions can. 736568651Skris [Steve Henson] 736668651Skris 736768651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 736868651Skris this option is set, tolerate broken clients that send the negotiated 736968651Skris protocol version number instead of the requested protocol version 737068651Skris number. 737168651Skris [Bodo Moeller] 737268651Skris 737368651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 737468651Skris i.e. non-zero for export ciphersuites, zero otherwise. 737568651Skris Previous versions had this flag inverted, inconsistent with 737668651Skris rsa_tmp_cb (..._TMP_RSA_CB). 737768651Skris [Bodo Moeller; problem reported by Amit Chopra] 737868651Skris 737968651Skris *) Add missing DSA library text string. Work around for some IIS 738068651Skris key files with invalid SEQUENCE encoding. 738168651Skris [Steve Henson] 738268651Skris 738368651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 738468651Skris and so on that are implemented in OpenSSL. 738568651Skris [Richard Levitte] 738668651Skris 738768651Skris *) Enhance c_rehash script. Old version would mishandle certificates 738868651Skris with the same subject name hash and wouldn't handle CRLs at all. 738968651Skris Added -fingerprint option to crl utility, to support new c_rehash 739068651Skris features. 739168651Skris [Steve Henson] 739268651Skris 739368651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 739468651Skris [Ulf M�ller] 739568651Skris 739668651Skris *) Fix for SSL server purpose checking. Server checking was 739768651Skris rejecting certificates which had extended key usage present 739868651Skris but no ssl client purpose. 739968651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 740068651Skris 740168651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 740268651Skris is a little unclear about how a blank password is handled. 740368651Skris Since the password in encoded as a BMPString with terminating 740468651Skris double NULL a zero length password would end up as just the 740568651Skris double NULL. However no password at all is different and is 740668651Skris handled differently in the PKCS#12 key generation code. NS 740768651Skris treats a blank password as zero length. MSIE treats it as no 740868651Skris password on export: but it will try both on import. We now do 740968651Skris the same: PKCS12_parse() tries zero length and no password if 741068651Skris the password is set to "" or NULL (NULL is now a valid password: 741168651Skris it wasn't before) as does the pkcs12 application. 741268651Skris [Steve Henson] 741368651Skris 741468651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 741568651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 741668651Skris be obtained from the error queue. 741768651Skris [Bodo Moeller] 741868651Skris 741968651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 742068651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 742168651Skris accordingly to avoid race conditions (this is necessary because 742268651Skris thread_hash is no longer constant once set). 742368651Skris [Bodo Moeller] 742468651Skris 742568651Skris *) Bugfix for linux-elf makefile.one. 742668651Skris [Ulf M�ller] 742768651Skris 742868651Skris *) RSA_get_default_method() will now cause a default 742968651Skris RSA_METHOD to be chosen if one doesn't exist already. 743068651Skris Previously this was only set during a call to RSA_new() 743168651Skris or RSA_new_method(NULL) meaning it was possible for 743268651Skris RSA_get_default_method() to return NULL. 743368651Skris [Geoff Thorpe] 743468651Skris 743568651Skris *) Added native name translation to the existing DSO code 743668651Skris that will convert (if the flag to do so is set) filenames 743768651Skris that are sufficiently small and have no path information 743868651Skris into a canonical native form. Eg. "blah" converted to 743968651Skris "libblah.so" or "blah.dll" etc. 744068651Skris [Geoff Thorpe] 744168651Skris 744268651Skris *) New function ERR_error_string_n(e, buf, len) which is like 744368651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 744468651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 744568651Skris may not be NULL. 744668651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 744768651Skris 744868651Skris *) CONF library reworked to become more general. A new CONF 744968651Skris configuration file reader "class" is implemented as well as a 745068651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 745168651Skris old CONF_* functions are still there, but are reimplemented to 745268651Skris work in terms of the new functions. Also, a set of functions 745368651Skris to handle the internal storage of the configuration data is 745468651Skris provided to make it easier to write new configuration file 745568651Skris reader "classes" (I can definitely see something reading a 745668651Skris configuration file in XML format, for example), called _CONF_*, 745768651Skris or "the configuration storage API"... 745868651Skris 745968651Skris The new configuration file reading functions are: 746068651Skris 746168651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 746268651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 746368651Skris 746468651Skris NCONF_default, NCONF_WIN32 746568651Skris 746668651Skris NCONF_dump_fp, NCONF_dump_bio 746768651Skris 746868651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 746968651Skris NCONF_new creates a new CONF object. This works in the same way 747068651Skris as other interfaces in OpenSSL, like the BIO interface. 747168651Skris NCONF_dump_* dump the internal storage of the configuration file, 747268651Skris which is useful for debugging. All other functions take the same 747368651Skris arguments as the old CONF_* functions wth the exception of the 747468651Skris first that must be a `CONF *' instead of a `LHASH *'. 747568651Skris 747668651Skris To make it easer to use the new classes with the old CONF_* functions, 747768651Skris the function CONF_set_default_method is provided. 747868651Skris [Richard Levitte] 747968651Skris 748068651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 748168651Skris mentioned in the documentation but had not been implemented. 748268651Skris (This option is not yet really useful because even the additional 748368651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 748468651Skris [Bodo Moeller] 748568651Skris 748668651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 748768651Skris OpenSSL-based applications) load shared libraries and bind to 748868651Skris them in a portable way. 748968651Skris [Geoff Thorpe, with contributions from Richard Levitte] 749068651Skris 749159191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 749259191Skris 749359191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 749459191Skris 749559191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 749659191Skris (the default implementation of RAND_status). 749759191Skris 749859191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 749959191Skris to '-clrext' (= clear extensions), as intended and documented. 750059191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 750159191Skris <attili@amaxo.com>] 750259191Skris 750359191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 750459191Skris was larger than the MD block size. 750559191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 750659191Skris 750759191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 750859191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 750959191Skris using the passed key: if the passed key was a private key the result 751059191Skris of X509_print(), for example, would be to print out all the private key 751159191Skris components. 751259191Skris [Steve Henson] 751359191Skris 751459191Skris *) des_quad_cksum() byte order bug fix. 751559191Skris [Ulf M�ller, using the problem description in krb4-0.9.7, where 751659191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 751759191Skris 751859191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 751959191Skris discouraged. 752059191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 752159191Skris 752259191Skris *) For easily testing in shell scripts whether some command 752359191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 752459191Skris returns with exit code 0 iff no command of the given name is available. 752559191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 752659191Skris the output goes to stdout and nothing is printed to stderr. 752759191Skris Additional arguments are always ignored. 752859191Skris 752959191Skris Since for each cipher there is a command of the same name, 753059191Skris the 'no-cipher' compilation switches can be tested this way. 753159191Skris 753259191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 753359191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 753459191Skris [Bodo Moeller] 753559191Skris 753659191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 753759191Skris [Bodo Moeller] 753859191Skris 753959191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 754059191Skris is set; it will be thrown away anyway because each handshake creates 754159191Skris its own key. 754259191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 754359191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 754459191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 754559191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 754659191Skris [Bodo Moeller] 754759191Skris 754859191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 754959191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 755059191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 755159191Skris does not suppress any output. 755259191Skris [Richard Levitte] 755359191Skris 755459191Skris *) Add compatibility options to the purpose and trust code. The 755559191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 755659191Skris accepts a certificate or CA, this was the previous behaviour, 755759191Skris with all the associated security issues. 755859191Skris 755959191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 756059191Skris automatically trust self signed roots in certificate store. A 756159191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 756259191Skris a purpose has no associated trust setting and it should instead 756359191Skris use the value in the default purpose. 756459191Skris [Steve Henson] 756559191Skris 756659191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 756759191Skris and fix a memory leak. 756859191Skris [Steve Henson] 756959191Skris 757059191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 757159191Skris reason strings from the previous version of the .c file, as 757259191Skris the default to have only downcase letters (and digits) in 757359191Skris automatically generated reasons codes is not always appropriate. 757459191Skris [Bodo Moeller] 757559191Skris 757659191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 757759191Skris using strerror. Previously, ERR_reason_error_string() returned 757859191Skris library names as reason strings for SYSerr; but SYSerr is a special 757959191Skris case where small numbers are errno values, not library numbers. 758059191Skris [Bodo Moeller] 758159191Skris 758259191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 758359191Skris converts DSA parameters into DH parameters. (When creating parameters, 758459191Skris DSA_generate_parameters is used.) 758559191Skris [Bodo Moeller] 758659191Skris 758759191Skris *) Include 'length' (recommended exponent length) in C code generated 758859191Skris by 'openssl dhparam -C'. 758959191Skris [Bodo Moeller] 759059191Skris 759159191Skris *) The second argument to set_label in perlasm was already being used 759259191Skris so couldn't be used as a "file scope" flag. Moved to third argument 759359191Skris which was free. 759459191Skris [Steve Henson] 759559191Skris 759659191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 759759191Skris instead of RAND_bytes for encryption IVs and salts. 759859191Skris [Bodo Moeller] 759959191Skris 760059191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 760159191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 760259191Skris RAND_set_rand_method would be impossible. 760359191Skris [Bodo Moeller] 760459191Skris 760559191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 760659191Skris number generation fails. 760759191Skris [Bodo Moeller] 760859191Skris 760959191Skris *) New 'rand' application for creating pseudo-random output. 761059191Skris [Bodo Moeller] 761159191Skris 761259191Skris *) Added configuration support for Linux/IA64 761359191Skris [Rolf Haberrecker <rolf@suse.de>] 761459191Skris 761559191Skris *) Assembler module support for Mingw32. 761659191Skris [Ulf M�ller] 761759191Skris 761859191Skris *) Shared library support for HPUX (in shlib/). 761959191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 762059191Skris 762159191Skris *) Shared library support for Solaris gcc. 762259191Skris [Lutz Behnke <behnke@trustcenter.de>] 762359191Skris 762459191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 762559191Skris 762659191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 762759191Skris were added manually and by SMIME_crlf_copy(). 762859191Skris [Steve Henson] 762959191Skris 763059191Skris *) In bntest.c don't call BN_rand with zero bits argument. 763159191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 763259191Skris 763359191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 763459191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 763559191Skris [Ulf M�ller] 763659191Skris 763759191Skris *) Add an optional second argument to the set_label() in the perl 763859191Skris assembly language builder. If this argument exists and is set 763959191Skris to 1 it signals that the assembler should use a symbol whose 764059191Skris scope is the entire file, not just the current function. This 764159191Skris is needed with MASM which uses the format label:: for this scope. 764259191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 764359191Skris 764459191Skris *) Change the ASN1 types so they are typedefs by default. Before 764559191Skris almost all types were #define'd to ASN1_STRING which was causing 764659191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 764759191Skris for example. 764859191Skris [Steve Henson] 764959191Skris 765059191Skris *) Change names of new functions to the new get1/get0 naming 765159191Skris convention: After 'get1', the caller owns a reference count 765259191Skris and has to call ..._free; 'get0' returns a pointer to some 765359191Skris data structure without incrementing reference counters. 765459191Skris (Some of the existing 'get' functions increment a reference 765559191Skris counter, some don't.) 765659191Skris Similarly, 'set1' and 'add1' functions increase reference 765759191Skris counters or duplicate objects. 765859191Skris [Steve Henson] 765959191Skris 766059191Skris *) Allow for the possibility of temp RSA key generation failure: 766159191Skris the code used to assume it always worked and crashed on failure. 766259191Skris [Steve Henson] 766359191Skris 766459191Skris *) Fix potential buffer overrun problem in BIO_printf(). 766559191Skris [Ulf M�ller, using public domain code by Patrick Powell; problem 766659191Skris pointed out by David Sacerdote <das33@cornell.edu>] 766759191Skris 766859191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 766959191Skris RAND_egd() and RAND_status(). In the command line application, 767059191Skris the EGD socket can be specified like a seed file using RANDFILE 767159191Skris or -rand. 767259191Skris [Ulf M�ller] 767359191Skris 767459191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 767559191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 767659191Skris [Steve Henson] 767759191Skris 767859191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 767959191Skris list to exclude them. This means that no special compilation option 768059191Skris is needed to use anonymous DH: it just needs to be included in the 768159191Skris cipher list. 768259191Skris [Steve Henson] 768359191Skris 768459191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 768559191Skris EVP_MD_type. The old functionality is available in a new macro called 768659191Skris EVP_MD_md(). Change code that uses it and update docs. 768759191Skris [Steve Henson] 768859191Skris 768959191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 769059191Skris where the 'void *' argument is replaced by a function pointer argument. 769159191Skris Previously 'void *' was abused to point to functions, which works on 769259191Skris many platforms, but is not correct. As these functions are usually 769359191Skris called by macros defined in OpenSSL header files, most source code 769459191Skris should work without changes. 769559191Skris [Richard Levitte] 769659191Skris 769759191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 769859191Skris sections with information on -D... compiler switches used for 769959191Skris compiling the library so that applications can see them. To enable 770059191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 770159191Skris must be defined. E.g., 770259191Skris #define OPENSSL_ALGORITHM_DEFINES 770359191Skris #include <openssl/opensslconf.h> 770459191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 770559191Skris [Richard Levitte, Ulf and Bodo M�ller] 770659191Skris 770759191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 770859191Skris record layer. 770959191Skris [Bodo Moeller] 771059191Skris 771159191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 771259191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 771359191Skris the required ASN1 format: arbitrary types determined by an OID. 771459191Skris [Steve Henson] 771559191Skris 771659191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 771759191Skris argument to 'req'. This is not because the function is newer or 771859191Skris better than others it just uses the work 'NEW' in the certificate 771959191Skris request header lines. Some software needs this. 772059191Skris [Steve Henson] 772159191Skris 772259191Skris *) Reorganise password command line arguments: now passwords can be 772359191Skris obtained from various sources. Delete the PEM_cb function and make 772459191Skris it the default behaviour: i.e. if the callback is NULL and the 772559191Skris usrdata argument is not NULL interpret it as a null terminated pass 772659191Skris phrase. If usrdata and the callback are NULL then the pass phrase 772759191Skris is prompted for as usual. 772859191Skris [Steve Henson] 772959191Skris 773059191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 773159191Skris the support is automatically enabled. The resulting binaries will 773259191Skris autodetect the card and use it if present. 773359191Skris [Ben Laurie and Compaq Inc.] 773459191Skris 773559191Skris *) Work around for Netscape hang bug. This sends certificate request 773659191Skris and server done in one record. Since this is perfectly legal in the 773759191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 773859191Skris the bugs/SSLv3 entry for more info. 773959191Skris [Steve Henson] 774059191Skris 774159191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 774259191Skris [Andy Polyakov] 774359191Skris 774459191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 774559191Skris of seed file. 774659191Skris [Steve Henson] 774759191Skris 774859191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 774959191Skris [Bodo Moeller] 775059191Skris 775159191Skris *) Add command line password options to the remaining applications. 775259191Skris [Steve Henson] 775359191Skris 775459191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 775559191Skris bits. 775659191Skris [Ulf M�ller] 775759191Skris 775859191Skris *) More tests in bntest.c, and changed test_bn output. 775959191Skris [Ulf M�ller] 776059191Skris 776159191Skris *) ./config recognizes MacOS X now. 776259191Skris [Andy Polyakov] 776359191Skris 776459191Skris *) Bug fix for BN_div() when the first words of num and divsor are 776559191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 776659191Skris [Ulf M�ller] 776759191Skris 776859191Skris *) Add support for various broken PKCS#8 formats, and command line 776959191Skris options to produce them. 777059191Skris [Steve Henson] 777159191Skris 777259191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 777359191Skris get temporary BIGNUMs from a BN_CTX. 777459191Skris [Ulf M�ller] 777559191Skris 777659191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 777759191Skris for p == 0. 777859191Skris [Ulf M�ller] 777959191Skris 778059191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 778159191Skris include a #define from the old name to the new. The original intent 778259191Skris was that statically linked binaries could for example just call 778359191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 778459191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 778559191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 778659191Skris one would link with the other. They are now in separate source files. 778759191Skris [Steve Henson] 778859191Skris 778959191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 779059191Skris [Steve Henson] 779159191Skris 779259191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 779359191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 779459191Skris loop, our standard modexp algorithms are faster). 779559191Skris [Bodo Moeller] 779659191Skris 779759191Skris *) Support for the EBCDIC character set completed. 779859191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 779959191Skris 780059191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 780159191Skris use void * instead of char * in lhash. 780259191Skris [Ulf M�ller] 780359191Skris 780459191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 780559191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 780659191Skris this the server could overwrite ephemeral keys that the client 780759191Skris has already seen). 780859191Skris [Bodo Moeller] 780959191Skris 781059191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 781159191Skris using 50 iterations of the Rabin-Miller test. 781259191Skris 781359191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 781459191Skris iterations of the Rabin-Miller test as required by the appendix 781559191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 781659191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 781759191Skris generation becomes much faster. 781859191Skris 781959191Skris This implies a change for the callback functions in DSA_is_prime 782059191Skris and DSA_generate_parameters: The callback function is called once 782159191Skris for each positive witness in the Rabin-Miller test, not just 782259191Skris occasionally in the inner loop; and the parameters to the 782359191Skris callback function now provide an iteration count for the outer 782459191Skris loop rather than for the current invocation of the inner loop. 782559191Skris DSA_generate_parameters additionally can call the callback 782659191Skris function with an 'iteration count' of -1, meaning that a 782759191Skris candidate has passed the trial division test (when q is generated 782859191Skris from an application-provided seed, trial division is skipped). 782959191Skris [Bodo Moeller] 783059191Skris 783159191Skris *) New function BN_is_prime_fasttest that optionally does trial 783259191Skris division before starting the Rabin-Miller test and has 783359191Skris an additional BN_CTX * argument (whereas BN_is_prime always 783459191Skris has to allocate at least one BN_CTX). 783559191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 783659191Skris trial division stage. 783759191Skris [Bodo Moeller] 783859191Skris 783959191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 784059191Skris as ASN1_TIME. 784159191Skris [Steve Henson] 784259191Skris 784359191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 784459191Skris [Steve Henson] 784559191Skris 784659191Skris *) New function BN_pseudo_rand(). 784759191Skris [Ulf M�ller] 784859191Skris 784959191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 785059191Skris bignum version of BN_from_montgomery() with the working code from 785159191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 785259191Skris the comments. 785359191Skris [Ulf M�ller] 785459191Skris 785559191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 785659191Skris made it impossible to use the same SSL_SESSION data structure in 785759191Skris SSL2 clients in multiple threads. 785859191Skris [Bodo Moeller] 785959191Skris 786059191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 786159191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 786259191Skris to seed the PRNG (previously an explicit byte count was required). 786359191Skris [Ulf M�ller, Bodo M�ller] 786459191Skris 786559191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 786659191Skris used (char *) instead of (void *) and had casts all over the place. 786759191Skris [Steve Henson] 786859191Skris 786959191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 787059191Skris [Ulf M�ller] 787159191Skris 787259191Skris *) Retain source code compatibility for BN_prime_checks macro: 787359191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 787459191Skris BN_prime_checks_for_size to determine the appropriate number of 787559191Skris Rabin-Miller iterations. 787659191Skris [Ulf M�ller] 787759191Skris 787859191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 787959191Skris DH_CHECK_P_NOT_SAFE_PRIME. 788059191Skris (Check if this is true? OpenPGP calls them "strong".) 788159191Skris [Ulf M�ller] 788259191Skris 788359191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 788459191Skris "dhparam". The old programs are retained for now but will handle DH keys 788559191Skris (instead of parameters) in future. 788659191Skris [Steve Henson] 788759191Skris 788859191Skris *) Make the ciphers, s_server and s_client programs check the return values 788959191Skris when a new cipher list is set. 789059191Skris [Steve Henson] 789159191Skris 789259191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 789359191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 789459191Skris wrong. 789559191Skris 789659191Skris The syntax for the cipher sorting has been extended to support sorting by 789759191Skris cipher-strength (using the strength_bits hard coded in the tables). 789859191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 789959191Skris 790059191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 790159191Skris string with an "undefined" symbol (neither command nor alphanumeric 790259191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 790359191Skris an error is flagged. 790459191Skris 790559191Skris Due to the strength-sorting extension, the code of the 790659191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 790759191Skris the readability was also increased :-) 790859191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 790959191Skris 791059191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 791159191Skris for the first serial number and places 2 in the serial number file. This 791259191Skris avoids problems when the root CA is created with serial number zero and 791359191Skris the first user certificate has the same issuer name and serial number 791459191Skris as the root CA. 791559191Skris [Steve Henson] 791659191Skris 791759191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 791859191Skris the new code. Add documentation for this stuff. 791959191Skris [Steve Henson] 792059191Skris 792159191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 792259191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 792359191Skris structures and behave in an analagous way to the X509v3 functions: 792459191Skris they shouldn't be called directly but wrapper functions should be used 792559191Skris instead. 792659191Skris 792759191Skris So we also now have some wrapper functions that call the X509at functions 792859191Skris when passed certificate requests. (TO DO: similar things can be done with 792959191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 793059191Skris things. Some of these need some d2i or i2d and print functionality 793159191Skris because they handle more complex structures.) 793259191Skris [Steve Henson] 793359191Skris 793459191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 793559191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 793659191Skris NO_RSA in ssl/s2*.c. 793759191Skris [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M�ller] 793859191Skris 793959191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 794059191Skris has a return value which indicates the quality of the random data 794159191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 794259191Skris error queue. New function RAND_pseudo_bytes() generates output that is 794359191Skris guaranteed to be unique but not unpredictable. RAND_add is like 794459191Skris RAND_seed, but takes an extra argument for an entropy estimate 794559191Skris (RAND_seed always assumes full entropy). 794659191Skris [Ulf M�ller] 794759191Skris 794859191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 794959191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 795059191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 795159191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 795259191Skris false-positive rate of at most 2^-80 for random input. 795359191Skris [Bodo Moeller] 795459191Skris 795559191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 795659191Skris [Bodo Moeller] 795759191Skris 795859191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 795959191Skris in the 0.9.5 release), this returns the chain 796059191Skris from an X509_CTX structure with a dup of the stack and all 796159191Skris the X509 reference counts upped: so the stack will exist 796259191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 796359191Skris to use this. 796459191Skris 796559191Skris Also make SSL_SESSION_print() print out the verify return 796659191Skris code. 796759191Skris [Steve Henson] 796859191Skris 796959191Skris *) Add manpage for the pkcs12 command. Also change the default 797059191Skris behaviour so MAC iteration counts are used unless the new 797159191Skris -nomaciter option is used. This improves file security and 797259191Skris only older versions of MSIE (4.0 for example) need it. 797359191Skris [Steve Henson] 797459191Skris 797559191Skris *) Honor the no-xxx Configure options when creating .DEF files. 797659191Skris [Ulf M�ller] 797759191Skris 797859191Skris *) Add PKCS#10 attributes to field table: challengePassword, 797959191Skris unstructuredName and unstructuredAddress. These are taken from 798059191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 798159191Skris international characters are used. 798259191Skris 798359191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 798459191Skris based on strings. Remove the 'loc' parameter when adding 798559191Skris attributes because these will be a SET OF encoding which is sorted 798659191Skris in ASN1 order. 798759191Skris [Steve Henson] 798859191Skris 798959191Skris *) Initial changes to the 'req' utility to allow request generation 799059191Skris automation. This will allow an application to just generate a template 799159191Skris file containing all the field values and have req construct the 799259191Skris request. 799359191Skris 799459191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 799559191Skris used all over the place including certificate requests and PKCS#7 799659191Skris structures. They are currently handled manually where necessary with 799759191Skris some primitive wrappers for PKCS#7. The new functions behave in a 799859191Skris manner analogous to the X509 extension functions: they allow 799959191Skris attributes to be looked up by NID and added. 800059191Skris 800159191Skris Later something similar to the X509V3 code would be desirable to 800259191Skris automatically handle the encoding, decoding and printing of the 800359191Skris more complex types. The string types like challengePassword can 800459191Skris be handled by the string table functions. 800559191Skris 800659191Skris Also modified the multi byte string table handling. Now there is 800759191Skris a 'global mask' which masks out certain types. The table itself 800859191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 800959191Skris is useful when for example there is only one permissible type 801059191Skris (as in countryName) and using the mask might result in no valid 801159191Skris types at all. 801259191Skris [Steve Henson] 801359191Skris 801459191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 801559191Skris SSL_get_peer_finished to allow applications to obtain the latest 801659191Skris Finished messages sent to the peer or expected from the peer, 801759191Skris respectively. (SSL_get_peer_finished is usually the Finished message 801859191Skris actually received from the peer, otherwise the protocol will be aborted.) 801959191Skris 802059191Skris As the Finished message are message digests of the complete handshake 802159191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 802259191Skris be used for external authentication procedures when the authentication 802359191Skris provided by SSL/TLS is not desired or is not enough. 802459191Skris [Bodo Moeller] 802559191Skris 802659191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 802759191Skris the host supports BWX extension and if Compaq C is present on the 802859191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 802959191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 803059191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 803159191Skris SHA1. 803259191Skris [Andy Polyakov] 803359191Skris 803459191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 803559191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 803659191Skris weak crypto and after checking the certificate is SGC a second one 803759191Skris with strong crypto. MS SGC stops the first handshake after receiving 803859191Skris the server certificate message and sends a second client hello. Since 803959191Skris a server will typically do all the time consuming operations before 804059191Skris expecting any further messages from the client (server key exchange 804159191Skris is the most expensive) there is little difference between the two. 804259191Skris 804359191Skris To get OpenSSL to support MS SGC we have to permit a second client 804459191Skris hello message after we have sent server done. In addition we have to 804559191Skris reset the MAC if we do get this second client hello. 804659191Skris [Steve Henson] 804759191Skris 804859191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 804959191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 805059191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 805159191Skris format DER encoded private key. Newer code should use PKCS#8 format which 805259191Skris has the key type encoded in the ASN1 structure. Added DER private key 805359191Skris support to pkcs8 application. 805459191Skris [Steve Henson] 805559191Skris 805659191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 805759191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 805859191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 805959191Skris is set, we interpret this as a request to violate the specification 806059191Skris (the worst that can happen is a handshake failure, and 'correct' 806159191Skris behaviour would result in a handshake failure anyway). 806259191Skris [Bodo Moeller] 806359191Skris 806459191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 806559191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 806659191Skris concurrently obtain them from an external cache). 806759191Skris The internal cache can handle only one SSL_SESSION with a given ID, 806859191Skris so if there's a conflict, we now throw out the old one to achieve 806959191Skris consistency. 807059191Skris [Bodo Moeller] 807159191Skris 807259191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 807359191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 807459191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 807559191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 807659191Skris example. 807759191Skris [Steve Henson] 807859191Skris 807959191Skris *) Simplify the trust setting structure and code. Now we just have 808059191Skris two sequences of OIDs for trusted and rejected settings. These will 808159191Skris typically have values the same as the extended key usage extension 808259191Skris and any application specific purposes. 808359191Skris 808459191Skris The trust checking code now has a default behaviour: it will just 808559191Skris check for an object with the same NID as the passed id. Functions can 808659191Skris be provided to override either the default behaviour or the behaviour 808759191Skris for a given id. SSL client, server and email already have functions 808859191Skris in place for compatibility: they check the NID and also return "trusted" 808959191Skris if the certificate is self signed. 809059191Skris [Steve Henson] 809159191Skris 809259191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 809359191Skris traditional format into an EVP_PKEY structure. 809459191Skris [Steve Henson] 809559191Skris 809659191Skris *) Add a password callback function PEM_cb() which either prompts for 809759191Skris a password if usr_data is NULL or otherwise assumes it is a null 809859191Skris terminated password. Allow passwords to be passed on command line 809959191Skris environment or config files in a few more utilities. 810059191Skris [Steve Henson] 810159191Skris 810259191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 810359191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 810459191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 810559191Skris Update documentation. 810659191Skris [Steve Henson] 810759191Skris 810859191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 810959191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 811059191Skris and produce an error if it couldn't. For compatibility we also have 811159191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 811259191Skris don't allocate anything because they don't need to. 811359191Skris [Steve Henson] 811459191Skris 811559191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 811659191Skris for details. 811759191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 811859191Skris 811959191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 812059191Skris possibly others as well. The purpose is to make an interface that 812159191Skris provide hooks so anyone can build a separate set of allocation and 812259191Skris deallocation routines to be used by OpenSSL, for example memory 812359191Skris pool implementations, or something else, which was previously hard 812459191Skris since Malloc(), Realloc() and Free() were defined as macros having 812559191Skris the values malloc, realloc and free, respectively (except for Win32 812659191Skris compilations). The same is provided for memory debugging code. 812759191Skris OpenSSL already comes with functionality to find memory leaks, but 812859191Skris this gives people a chance to debug other memory problems. 812959191Skris 813059191Skris With these changes, a new set of functions and macros have appeared: 813159191Skris 813268651Skris CRYPTO_set_mem_debug_functions() [F] 813368651Skris CRYPTO_get_mem_debug_functions() [F] 813468651Skris CRYPTO_dbg_set_options() [F] 813568651Skris CRYPTO_dbg_get_options() [F] 813668651Skris CRYPTO_malloc_debug_init() [M] 813759191Skris 813859191Skris The memory debug functions are NULL by default, unless the library 813959191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 814059191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 814159191Skris gives the standard debugging functions that come with OpenSSL) or 814259191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 814359191Skris provided by the library user) must be used. When the standard 814459191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 814559191Skris request additional information: 814659191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 814759191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 814859191Skris 814959191Skris Also, things like CRYPTO_set_mem_functions will always give the 815059191Skris expected result (the new set of functions is used for allocation 815159191Skris and deallocation) at all times, regardless of platform and compiler 815259191Skris options. 815359191Skris 815459191Skris To finish it up, some functions that were never use in any other 815559191Skris way than through macros have a new API and new semantic: 815659191Skris 815759191Skris CRYPTO_dbg_malloc() 815859191Skris CRYPTO_dbg_realloc() 815959191Skris CRYPTO_dbg_free() 816059191Skris 816159191Skris All macros of value have retained their old syntax. 816259191Skris [Richard Levitte and Bodo Moeller] 816359191Skris 816459191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 816559191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 816659191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 816759191Skris algorithm. 816859191Skris [Steve Henson] 816959191Skris 817059191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 817159191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 817259191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 817359191Skris 817459191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 817559191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 817659191Skris functionality to handle multipart/signed properly) and a utility 817759191Skris called 'smime' to call all this stuff. This is based on code I 817859191Skris originally wrote for Celo who have kindly allowed it to be 817959191Skris included in OpenSSL. 818059191Skris [Steve Henson] 818159191Skris 818259191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 818359191Skris des_set_key (aka des_key_sched). Global variable des_check_key 818459191Skris decides which of these is called by des_set_key; this way 818559191Skris des_check_key behaves as it always did, but applications and 818659191Skris the library itself, which was buggy for des_check_key == 1, 818759191Skris have a cleaner way to pick the version they need. 818859191Skris [Bodo Moeller] 818959191Skris 819059191Skris *) New function PKCS12_newpass() which changes the password of a 819159191Skris PKCS12 structure. 819259191Skris [Steve Henson] 819359191Skris 819459191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 819559191Skris dynamic mix. In both cases the ids can be used as an index into the 819659191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 819759191Skris functions so they accept a list of the field values and the 819859191Skris application doesn't need to directly manipulate the X509_TRUST 819959191Skris structure. 820059191Skris [Steve Henson] 820159191Skris 820259191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 820359191Skris need initialising. 820459191Skris [Steve Henson] 820559191Skris 820659191Skris *) Modify the way the V3 extension code looks up extensions. This now 820759191Skris works in a similar way to the object code: we have some "standard" 820859191Skris extensions in a static table which is searched with OBJ_bsearch() 820959191Skris and the application can add dynamic ones if needed. The file 821059191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 821159191Skris updated whenever a new extension is added to the core code and kept 821259191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 821359191Skris this. New extensions are not added too often so this file can readily 821459191Skris be maintained manually. 821559191Skris 821659191Skris There are two big advantages in doing things this way. The extensions 821759191Skris can be looked up immediately and no longer need to be "added" using 821859191Skris X509V3_add_standard_extensions(): this function now does nothing. 821959191Skris [Side note: I get *lots* of email saying the extension code doesn't 822059191Skris work because people forget to call this function] 822159191Skris Also no dynamic allocation is done unless new extensions are added: 822259191Skris so if we don't add custom extensions there is no need to call 822359191Skris X509V3_EXT_cleanup(). 822459191Skris [Steve Henson] 822559191Skris 822659191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 822759191Skris magic header, so unsalted files fail gracefully instead of just decrypting 822859191Skris to garbage. This is because not salting is a big security hole, so people 822959191Skris should be discouraged from doing it. 823059191Skris [Ben Laurie] 823159191Skris 823259191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 823359191Skris digest to be passed on the command line but it only used this 823459191Skris parameter when signing a certificate. Modified so all relevant 823559191Skris operations are affected by the digest parameter including the 823659191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 823759191Skris DSA key was used because it didn't fix the digest. 823859191Skris [Steve Henson] 823959191Skris 824059191Skris *) Initial certificate chain verify code. Currently tests the untrusted 824159191Skris certificates for consistency with the verify purpose (which is set 824259191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 824359191Skris 824459191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 824559191Skris this is because it will reject chains with invalid extensions whereas 824659191Skris every previous version of OpenSSL and SSLeay made no checks at all. 824759191Skris 824859191Skris Trust code: checks the root CA for the relevant trust settings. Trust 824959191Skris settings have an initial value consistent with the verify purpose: e.g. 825059191Skris if the verify purpose is for SSL client use it expects the CA to be 825159191Skris trusted for SSL client use. However the default value can be changed to 825259191Skris permit custom trust settings: one example of this would be to only trust 825359191Skris certificates from a specific "secure" set of CAs. 825459191Skris 825559191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 825659191Skris which should be used for version portability: especially since the 825759191Skris verify structure is likely to change more often now. 825859191Skris 825959191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 826059191Skris to set them. If not set then assume SSL clients will verify SSL servers 826159191Skris and vice versa. 826259191Skris 826359191Skris Two new options to the verify program: -untrusted allows a set of 826459191Skris untrusted certificates to be passed in and -purpose which sets the 826559191Skris intended purpose of the certificate. If a purpose is set then the 826659191Skris new chain verify code is used to check extension consistency. 826759191Skris [Steve Henson] 826859191Skris 826959191Skris *) Support for the authority information access extension. 827059191Skris [Steve Henson] 827159191Skris 827259191Skris *) Modify RSA and DSA PEM read routines to transparently handle 827359191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 827459191Skris public keys in a format compatible with certificate 827559191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 827659191Skris functions called *_PublicKey_* which used various odd formats so 827759191Skris these are retained for compatibility: however the DSA variants were 827859191Skris never in a public release so they have been deleted. Changed dsa/rsa 827959191Skris utilities to handle the new format: note no releases ever handled public 828059191Skris keys so we should be OK. 828159191Skris 828259191Skris The primary motivation for this change is to avoid the same fiasco 828359191Skris that dogs private keys: there are several incompatible private key 828459191Skris formats some of which are standard and some OpenSSL specific and 828559191Skris require various evil hacks to allow partial transparent handling and 828659191Skris even then it doesn't work with DER formats. Given the option anything 828759191Skris other than PKCS#8 should be dumped: but the other formats have to 828859191Skris stay in the name of compatibility. 828959191Skris 829059191Skris With public keys and the benefit of hindsight one standard format 829159191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 829259191Skris it clearly returns an error if you try to read the wrong kind of key. 829359191Skris 829459191Skris Added a -pubkey option to the 'x509' utility to output the public key. 829559191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 829659191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 829759191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 829859191Skris that do the same as the EVP_PKEY_assign_*() except they up the 829959191Skris reference count of the added key (they don't "swallow" the 830059191Skris supplied key). 830159191Skris [Steve Henson] 830259191Skris 830359191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 830459191Skris CRLs would fail if the file contained no certificates or no CRLs: 830559191Skris added a new function to read in both types and return the number 830659191Skris read: this means that if none are read it will be an error. The 830759191Skris DER versions of the certificate and CRL reader would always fail 830859191Skris because it isn't possible to mix certificates and CRLs in DER format 830959191Skris without choking one or the other routine. Changed this to just read 831059191Skris a certificate: this is the best we can do. Also modified the code 831159191Skris in apps/verify.c to take notice of return codes: it was previously 831259191Skris attempting to read in certificates from NULL pointers and ignoring 831359191Skris any errors: this is one reason why the cert and CRL reader seemed 831459191Skris to work. It doesn't check return codes from the default certificate 831559191Skris routines: these may well fail if the certificates aren't installed. 831659191Skris [Steve Henson] 831759191Skris 831859191Skris *) Code to support otherName option in GeneralName. 831959191Skris [Steve Henson] 832059191Skris 832159191Skris *) First update to verify code. Change the verify utility 832259191Skris so it warns if it is passed a self signed certificate: 832359191Skris for consistency with the normal behaviour. X509_verify 832459191Skris has been modified to it will now verify a self signed 832559191Skris certificate if *exactly* the same certificate appears 832659191Skris in the store: it was previously impossible to trust a 832759191Skris single self signed certificate. This means that: 832859191Skris openssl verify ss.pem 832959191Skris now gives a warning about a self signed certificate but 833059191Skris openssl verify -CAfile ss.pem ss.pem 833159191Skris is OK. 833259191Skris [Steve Henson] 833359191Skris 833459191Skris *) For servers, store verify_result in SSL_SESSION data structure 833559191Skris (and add it to external session representation). 833659191Skris This is needed when client certificate verifications fails, 833759191Skris but an application-provided verification callback (set by 833859191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 833959191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 834059191Skris but returns 1): When the session is reused, we have to set 834159191Skris ssl->verify_result to the appropriate error code to avoid 834259191Skris security holes. 834359191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 834459191Skris 834559191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 834659191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 834759191Skris didn't contain any existing data because it was being created. 834859191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 834959191Skris 835059191Skris *) Add a salt to the key derivation routines in enc.c. This 835159191Skris forms the first 8 bytes of the encrypted file. Also add a 835259191Skris -S option to allow a salt to be input on the command line. 835359191Skris [Steve Henson] 835459191Skris 835559191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 835659191Skris to compare two certificates. We do this by working out the SHA1 835759191Skris hash and comparing that. X509_cmp() will be needed by the trust 835859191Skris code. 835959191Skris [Steve Henson] 836059191Skris 836159191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 836259191Skris the reference count in the SSL_SESSION returned. 836359191Skris [Geoff Thorpe <geoff@eu.c2.net>] 836459191Skris 836559191Skris *) Fix for 'req': it was adding a null to request attributes. 836659191Skris Also change the X509_LOOKUP and X509_INFO code to handle 836759191Skris certificate auxiliary information. 836859191Skris [Steve Henson] 836959191Skris 837059191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 837159191Skris the 'enc' command. 837259191Skris [Steve Henson] 837359191Skris 837459191Skris *) Add the possibility to add extra information to the memory leak 837559191Skris detecting output, to form tracebacks, showing from where each 837659191Skris allocation was originated: CRYPTO_push_info("constant string") adds 837759191Skris the string plus current file name and line number to a per-thread 837859191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 837959191Skris is like calling CYRPTO_pop_info() until the stack is empty. 838059191Skris Also updated memory leak detection code to be multi-thread-safe. 838159191Skris [Richard Levitte] 838259191Skris 838359191Skris *) Add options -text and -noout to pkcs7 utility and delete the 838459191Skris encryption options which never did anything. Update docs. 838559191Skris [Steve Henson] 838659191Skris 838759191Skris *) Add options to some of the utilities to allow the pass phrase 838859191Skris to be included on either the command line (not recommended on 838959191Skris OSes like Unix) or read from the environment. Update the 839059191Skris manpages and fix a few bugs. 839159191Skris [Steve Henson] 839259191Skris 839359191Skris *) Add a few manpages for some of the openssl commands. 839459191Skris [Steve Henson] 839559191Skris 839659191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 839759191Skris leaking and not finding already revoked certificates. 839859191Skris [Steve Henson] 839959191Skris 840059191Skris *) Extensive changes to support certificate auxiliary information. 840159191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 840259191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 840359191Skris can still read in a certificate file in the usual way but it 840459191Skris will also read in any additional "auxiliary information". By 840559191Skris doing things this way a fair degree of compatibility can be 840659191Skris retained: existing certificates can have this information added 840759191Skris using the new 'x509' options. 840859191Skris 840959191Skris Current auxiliary information includes an "alias" and some trust 841059191Skris settings. The trust settings will ultimately be used in enhanced 841159191Skris certificate chain verification routines: currently a certificate 841259191Skris can only be trusted if it is self signed and then it is trusted 841359191Skris for all purposes. 841459191Skris [Steve Henson] 841559191Skris 841659191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 841759191Skris The problem was that one of the replacement routines had not been working 841859191Skris since SSLeay releases. For now the offending routine has been replaced 841959191Skris with non-optimised assembler. Even so, this now gives around 95% 842059191Skris performance improvement for 1024 bit RSA signs. 842159191Skris [Mark Cox] 842259191Skris 842359191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 842459191Skris handling. Most clients have the effective key size in bits equal to 842559191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 842659191Skris A few however don't do this and instead use the size of the decrypted key 842759191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 842859191Skris the effective key length. In this case the effective key length can still 842959191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 843059191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 843159191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 843259191Skris the key length and effective key length are equal. 843359191Skris [Steve Henson] 843459191Skris 843559191Skris *) Add a bunch of functions that should simplify the creation of 843659191Skris X509_NAME structures. Now you should be able to do: 843759191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 843859191Skris and have it automatically work out the correct field type and fill in 843959191Skris the structures. The more adventurous can try: 844059191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 844159191Skris and it will (hopefully) work out the correct multibyte encoding. 844259191Skris [Steve Henson] 844359191Skris 844459191Skris *) Change the 'req' utility to use the new field handling and multibyte 844559191Skris copy routines. Before the DN field creation was handled in an ad hoc 844659191Skris way in req, ca, and x509 which was rather broken and didn't support 844759191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 844859191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 844959191Skris using the dirstring_type option. See the new comment in the default 845059191Skris openssl.cnf for more info. 845159191Skris [Steve Henson] 845259191Skris 845359191Skris *) Make crypto/rand/md_rand.c more robust: 845459191Skris - Assure unique random numbers after fork(). 845559191Skris - Make sure that concurrent threads access the global counter and 845659191Skris md serializably so that we never lose entropy in them 845759191Skris or use exactly the same state in multiple threads. 845859191Skris Access to the large state is not always serializable because 845959191Skris the additional locking could be a performance killer, and 846059191Skris md should be large enough anyway. 846159191Skris [Bodo Moeller] 846259191Skris 846359191Skris *) New file apps/app_rand.c with commonly needed functionality 846459191Skris for handling the random seed file. 846559191Skris 846659191Skris Use the random seed file in some applications that previously did not: 846759191Skris ca, 846859191Skris dsaparam -genkey (which also ignored its '-rand' option), 846959191Skris s_client, 847059191Skris s_server, 847159191Skris x509 (when signing). 847259191Skris Except on systems with /dev/urandom, it is crucial to have a random 847359191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 847459191Skris for RSA signatures we could do without one. 847559191Skris 847659191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 847759191Skris of each file listed in the '-rand' option. The function as previously 847859191Skris found in genrsa is now in app_rand.c and is used by all programs 847959191Skris that support '-rand'. 848059191Skris [Bodo Moeller] 848159191Skris 848259191Skris *) In RAND_write_file, use mode 0600 for creating files; 848359191Skris don't just chmod when it may be too late. 848459191Skris [Bodo Moeller] 848559191Skris 848659191Skris *) Report an error from X509_STORE_load_locations 848759191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 848859191Skris [Bill Perry] 848959191Skris 849059191Skris *) New function ASN1_mbstring_copy() this copies a string in either 849159191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 849259191Skris into an ASN1_STRING type. A mask of permissible types is passed 849359191Skris and it chooses the "minimal" type to use or an error if not type 849459191Skris is suitable. 849559191Skris [Steve Henson] 849659191Skris 849759191Skris *) Add function equivalents to the various macros in asn1.h. The old 849859191Skris macros are retained with an M_ prefix. Code inside the library can 849959191Skris use the M_ macros. External code (including the openssl utility) 850059191Skris should *NOT* in order to be "shared library friendly". 850159191Skris [Steve Henson] 850259191Skris 850359191Skris *) Add various functions that can check a certificate's extensions 850459191Skris to see if it usable for various purposes such as SSL client, 850559191Skris server or S/MIME and CAs of these types. This is currently 850659191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 850759191Skris verification. Also added a -purpose flag to x509 utility to 850859191Skris print out all the purposes. 850959191Skris [Steve Henson] 851059191Skris 851159191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 851259191Skris functions. 851359191Skris [Steve Henson] 851459191Skris 851559191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 851659191Skris for, obtain and decode and extension and obtain its critical flag. 851759191Skris This allows all the necessary extension code to be handled in a 851859191Skris single function call. 851959191Skris [Steve Henson] 852059191Skris 852159191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 852259191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 852359191Skris [Andy Polyakov] 852459191Skris 852559191Skris *) New -noout option to asn1parse. This causes no output to be produced 852659191Skris its main use is when combined with -strparse and -out to extract data 852759191Skris from a file (which may not be in ASN.1 format). 852859191Skris [Steve Henson] 852959191Skris 853059191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 853159191Skris when producing the local key id. 853259191Skris [Richard Levitte <levitte@stacken.kth.se>] 853359191Skris 853459191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 853559191Skris stated explicitly. If it is not stated then it tries the first server 853659191Skris certificate file. The previous behaviour hard coded the filename 853759191Skris "server.pem". 853859191Skris [Steve Henson] 853959191Skris 854059191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 854159191Skris a public key to be input or output. For example: 854259191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 854359191Skris Also added necessary DSA public key functions to handle this. 854459191Skris [Steve Henson] 854559191Skris 854659191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 854759191Skris in the message. This was handled by allowing 854859191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 854959191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 855059191Skris 855159191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 855259191Skris to the end of the strings whereas this didn't. This would cause problems 855359191Skris if strings read with d2i_ASN1_bytes() were later modified. 855459191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 855559191Skris 855659191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 855759191Skris data and it contains EOF it will end up returning an error. This is 855859191Skris caused by input 46 bytes long. The cause is due to the way base64 855959191Skris BIOs find the start of base64 encoded data. They do this by trying a 856059191Skris trial decode on each line until they find one that works. When they 856159191Skris do a flag is set and it starts again knowing it can pass all the 856259191Skris data directly through the decoder. Unfortunately it doesn't reset 856359191Skris the context it uses. This means that if EOF is reached an attempt 856459191Skris is made to pass two EOFs through the context and this causes the 856559191Skris resulting error. This can also cause other problems as well. As is 856659191Skris usual with these problems it takes *ages* to find and the fix is 856759191Skris trivial: move one line. 856859191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 856959191Skris 857059191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 857159191Skris old code wouldn't work because it needed to select() on sockets and the 857259191Skris tty (for keypresses and to see if data could be written). Win32 only 857359191Skris supports select() on sockets so we select() with a 1s timeout on the 857459191Skris sockets and then see if any characters are waiting to be read, if none 857559191Skris are present then we retry, we also assume we can always write data to 857659191Skris the tty. This isn't nice because the code then blocks until we've 857759191Skris received a complete line of data and it is effectively polling the 857859191Skris keyboard at 1s intervals: however it's quite a bit better than not 857959191Skris working at all :-) A dedicated Windows application might handle this 858059191Skris with an event loop for example. 858159191Skris [Steve Henson] 858259191Skris 858359191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 858459191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 858559191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 858659191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 858759191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 858859191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 858959191Skris This necessitated the support of an extra signature type NID_md5_sha1 859059191Skris for SSL signatures and modifications to the SSL library to use it instead 859159191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 859259191Skris [Steve Henson] 859359191Skris 859459191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 859559191Skris will lookup a CRL issuers certificate and verify the signature in a 859659191Skris similar way to the verify program. Tidy up the crl program so it 859759191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 859859191Skris less strict. It will now permit CRL extensions even if it is not 859959191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 860059191Skris [Steve Henson] 860159191Skris 860259191Skris *) Initialize all non-automatic variables each time one of the openssl 860359191Skris sub-programs is started (this is necessary as they may be started 860459191Skris multiple times from the "OpenSSL>" prompt). 860559191Skris [Lennart Bang, Bodo Moeller] 860659191Skris 860759191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 860859191Skris removing all other RSA functionality (this is what NO_RSA does). This 860959191Skris is so (for example) those in the US can disable those operations covered 861059191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 861159191Skris key generation. 861259191Skris [Steve Henson] 861359191Skris 861459191Skris *) Non-copying interface to BIO pairs. 861559191Skris (still largely untested) 861659191Skris [Bodo Moeller] 861759191Skris 861859191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 861959191Skris ASCII string. This was handled independently in various places before. 862059191Skris [Steve Henson] 862159191Skris 862259191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 862359191Skris UTF8 strings a character at a time. 862459191Skris [Steve Henson] 862559191Skris 862659191Skris *) Use client_version from client hello to select the protocol 862759191Skris (s23_srvr.c) and for RSA client key exchange verification 862859191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 862959191Skris [Bodo Moeller] 863059191Skris 863159191Skris *) Add various utility functions to handle SPKACs, these were previously 863259191Skris handled by poking round in the structure internals. Added new function 863359191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 863459191Skris print, verify and generate SPKACs. Based on an original idea from 863559191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 863659191Skris [Steve Henson] 863759191Skris 863859191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 863959191Skris [Andy Polyakov] 864059191Skris 864159191Skris *) Allow the config file extension section to be overwritten on the 864259191Skris command line. Based on an original idea from Massimiliano Pala 864359191Skris <madwolf@comune.modena.it>. The new option is called -extensions 864459191Skris and can be applied to ca, req and x509. Also -reqexts to override 864559191Skris the request extensions in req and -crlexts to override the crl extensions 864659191Skris in ca. 864759191Skris [Steve Henson] 864859191Skris 864959191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 865059191Skris the same field multiple times by preceding it by "XXXX." for example: 865159191Skris 1.OU="Unit name 1" 865259191Skris 2.OU="Unit name 2" 865359191Skris this is the same syntax as used in the req config file. 865459191Skris [Steve Henson] 865559191Skris 865659191Skris *) Allow certificate extensions to be added to certificate requests. These 865759191Skris are specified in a 'req_extensions' option of the req section of the 865859191Skris config file. They can be printed out with the -text option to req but 865959191Skris are otherwise ignored at present. 866059191Skris [Steve Henson] 866159191Skris 866259191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 866359191Skris data read consists of only the final block it would not decrypted because 866459191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 866559191Skris A misplaced 'break' also meant the decrypted final block might not be 866659191Skris copied until the next read. 866759191Skris [Steve Henson] 866859191Skris 866959191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 867059191Skris a few extra parameters to the DH structure: these will be useful if 867159191Skris for example we want the value of 'q' or implement X9.42 DH. 867259191Skris [Steve Henson] 867359191Skris 867459191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 867559191Skris provides hooks that allow the default DSA functions or functions on a 867659191Skris "per key" basis to be replaced. This allows hardware acceleration and 867759191Skris hardware key storage to be handled without major modification to the 867859191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 867959191Skris associated functions. 868059191Skris [Steve Henson] 868159191Skris 868259191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 868359191Skris as "read only": it can't be written to and the buffer it points to will 868459191Skris not be freed. Reading from a read only BIO is much more efficient than 868559191Skris a normal memory BIO. This was added because there are several times when 868659191Skris an area of memory needs to be read from a BIO. The previous method was 868759191Skris to create a memory BIO and write the data to it, this results in two 868859191Skris copies of the data and an O(n^2) reading algorithm. There is a new 868959191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 869059191Skris an area of memory. Also modified the PKCS#7 routines to use read only 869159191Skris memory BIOs. 869259191Skris [Steve Henson] 869359191Skris 869459191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 869559191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 869659191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 869759191Skris but a retry condition occured while trying to read the rest. 869859191Skris [Bodo Moeller] 869959191Skris 870059191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 870159191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 870259191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 870359191Skris the encrypted data type: this is a more sensible place to put it and it 870459191Skris allows the PKCS#12 code to be tidied up that duplicated this 870559191Skris functionality. 870659191Skris [Steve Henson] 870759191Skris 870859191Skris *) Changed obj_dat.pl script so it takes its input and output files on 870959191Skris the command line. This should avoid shell escape redirection problems 871059191Skris under Win32. 871159191Skris [Steve Henson] 871259191Skris 871359191Skris *) Initial support for certificate extension requests, these are included 871459191Skris in things like Xenroll certificate requests. Included functions to allow 871559191Skris extensions to be obtained and added. 871659191Skris [Steve Henson] 871759191Skris 871859191Skris *) -crlf option to s_client and s_server for sending newlines as 871959191Skris CRLF (as required by many protocols). 872059191Skris [Bodo Moeller] 872159191Skris 872255714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 872355714Skris 872455714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 872555714Skris [Ralf S. Engelschall] 872655714Skris 872755714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 872855714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 872955714Skris 873055714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 873155714Skris program. 873255714Skris [Steve Henson] 873355714Skris 873455714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 873555714Skris DH parameters/keys (q is lost during that conversion, but the resulting 873655714Skris DH parameters contain its length). 873755714Skris 873855714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 873955714Skris much faster than DH_generate_parameters (which creates parameters 874055714Skris where p = 2*q + 1), and also the smaller q makes DH computations 874155714Skris much more efficient (160-bit exponentiation instead of 1024-bit 874255714Skris exponentiation); so this provides a convenient way to support DHE 874355714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 874455714Skris utter importance to use 874555714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 874655714Skris or 874755714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 874855714Skris when such DH parameters are used, because otherwise small subgroup 874955714Skris attacks may become possible! 875055714Skris [Bodo Moeller] 875155714Skris 875255714Skris *) Avoid memory leak in i2d_DHparams. 875355714Skris [Bodo Moeller] 875455714Skris 875555714Skris *) Allow the -k option to be used more than once in the enc program: 875655714Skris this allows the same encrypted message to be read by multiple recipients. 875755714Skris [Steve Henson] 875855714Skris 875955714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 876055714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 876155714Skris it will always use the numerical form of the OID, even if it has a short 876255714Skris or long name. 876355714Skris [Steve Henson] 876455714Skris 876555714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 876655714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 876755714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 876855714Skris no private key components need be present and it might store extra data 876959191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 877059191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 877159191Skris private key operations. 877255714Skris [Steve Henson] 877355714Skris 877455714Skris *) Added support for SPARC Linux. 877555714Skris [Andy Polyakov] 877655714Skris 877755714Skris *) pem_password_cb function type incompatibly changed from 877855714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 877955714Skris to 878055714Skris ....(char *buf, int size, int rwflag, void *userdata); 878155714Skris so that applications can pass data to their callbacks: 878255714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 878355714Skris additional void * argument, which is just handed through whenever 878455714Skris the password callback is called. 878559191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 878655714Skris 878755714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 878855714Skris 878955714Skris Compatibility note: As many C implementations push function arguments 879055714Skris onto the stack in reverse order, the new library version is likely to 879155714Skris interoperate with programs that have been compiled with the old 879255714Skris pem_password_cb definition (PEM_whatever takes some data that 879355714Skris happens to be on the stack as its last argument, and the callback 879455714Skris just ignores this garbage); but there is no guarantee whatsoever that 879555714Skris this will work. 879655714Skris 879755714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 879855714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 879955714Skris problems not only on Windows, but also on some Unix platforms. 880055714Skris To avoid problematic command lines, these definitions are now in an 880155714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 880255714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 880355714Skris [Bodo Moeller] 880455714Skris 880555714Skris *) MIPS III/IV assembler module is reimplemented. 880655714Skris [Andy Polyakov] 880755714Skris 880855714Skris *) More DES library cleanups: remove references to srand/rand and 880955714Skris delete an unused file. 881055714Skris [Ulf M�ller] 881155714Skris 881255714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 881355714Skris since not many people have MASM (ml) and it can be hard to obtain. 881455714Skris This is currently experimental but it seems to work OK and pass all 881555714Skris the tests. Check out INSTALL.W32 for info. 881655714Skris [Steve Henson] 881755714Skris 881855714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 881955714Skris without temporary keys kept an extra copy of the server key, 882055714Skris and connections with temporary keys did not free everything in case 882155714Skris of an error. 882255714Skris [Bodo Moeller] 882355714Skris 882455714Skris *) New function RSA_check_key and new openssl rsa option -check 882555714Skris for verifying the consistency of RSA keys. 882655714Skris [Ulf Moeller, Bodo Moeller] 882755714Skris 882855714Skris *) Various changes to make Win32 compile work: 882955714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 883055714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 883155714Skris comparison" warnings. 883255714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 883355714Skris [Steve Henson] 883455714Skris 883555714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 883655714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 883755714Skris derived keys are printed to stderr. 883855714Skris [Steve Henson] 883955714Skris 884055714Skris *) Copy the flags in ASN1_STRING_dup(). 884155714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 884255714Skris 884355714Skris *) The x509 application mishandled signing requests containing DSA 884455714Skris keys when the signing key was also DSA and the parameters didn't match. 884555714Skris 884655714Skris It was supposed to omit the parameters when they matched the signing key: 884755714Skris the verifying software was then supposed to automatically use the CA's 884855714Skris parameters if they were absent from the end user certificate. 884955714Skris 885055714Skris Omitting parameters is no longer recommended. The test was also 885155714Skris the wrong way round! This was probably due to unusual behaviour in 885255714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 885355714Skris This meant that parameters were omitted when they *didn't* match and 885455714Skris the certificate was useless. Certificates signed with 'ca' didn't have 885555714Skris this bug. 885655714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 885755714Skris 885855714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 885955714Skris The interface is as follows: 886055714Skris Applications can use 886155714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 886255714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 886355714Skris "off" is now the default. 886455714Skris The library internally uses 886555714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 886655714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 886755714Skris to disable memory-checking temporarily. 886855714Skris 886955714Skris Some inconsistent states that previously were possible (and were 887055714Skris even the default) are now avoided. 887155714Skris 887255714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 887355714Skris with each memory chunk allocated; this is occasionally more helpful 887455714Skris than just having a counter. 887555714Skris 887655714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 887755714Skris 887855714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 887955714Skris extensions. 888055714Skris [Bodo Moeller] 888155714Skris 888255714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 888355714Skris which largely parallels "options", but is for changing API behaviour, 888455714Skris whereas "options" are about protocol behaviour. 888555714Skris Initial "mode" flags are: 888655714Skris 888755714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 888855714Skris a single record has been written. 888955714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 889055714Skris retries use the same buffer location. 889155714Skris (But all of the contents must be 889255714Skris copied!) 889355714Skris [Bodo Moeller] 889455714Skris 889579998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 889655714Skris worked. 889755714Skris 889855714Skris *) Fix problems with no-hmac etc. 889955714Skris [Ulf M�ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 890055714Skris 890155714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 890255714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 890355714Skris to mess around with the internals of an RSA structure. 890455714Skris [Steve Henson] 890555714Skris 890655714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 890755714Skris Also really enable memory leak checks in openssl.c and in some 890855714Skris test programs. 890955714Skris [Chad C. Mulligan, Bodo Moeller] 891055714Skris 891155714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 891255714Skris up the length of negative integers. This has now been simplified to just 891355714Skris store the length when it is first determined and use it later, rather 891455714Skris than trying to keep track of where data is copied and updating it to 891555714Skris point to the end. 891655714Skris [Steve Henson, reported by Brien Wheeler 891755714Skris <bwheeler@authentica-security.com>] 891855714Skris 891955714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 892055714Skris of a PKCS#7 signature but with the signing certificate passed to the 892155714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 892255714Skris certificate is present in the PKCS#7 structure. This isn't always the 892355714Skris case: certificates can be omitted from a PKCS#7 structure and be 892455714Skris distributed by "out of band" means (such as a certificate database). 892555714Skris [Steve Henson] 892655714Skris 892755714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 892855714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 892955714Skris necessary function names. 893055714Skris [Steve Henson] 893155714Skris 893255714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 893355714Skris options set by Configure in the top level Makefile, and Configure 893455714Skris was not even able to write more than one option correctly. 893555714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 893655714Skris [Bodo Moeller] 893755714Skris 893855714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 893955714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 894055714Skris for example allow memory BIOs to contain config info. 894155714Skris [Steve Henson] 894255714Skris 894355714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 894455714Skris Whoever hopes to achieve shared-library compatibility across versions 894555714Skris must use this, not the compile-time macro. 894655714Skris (Exercise 0.9.4: Which is the minimum library version required by 894755714Skris such programs?) 894855714Skris Note: All this applies only to multi-threaded programs, others don't 894955714Skris need locks. 895055714Skris [Bodo Moeller] 895155714Skris 895255714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 895355714Skris through a BIO pair triggered the default case, i.e. 895455714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 895555714Skris [Bodo Moeller] 895655714Skris 895755714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 895855714Skris can use the SSL library even if none of the specific BIOs is 895955714Skris appropriate. 896055714Skris [Bodo Moeller] 896155714Skris 896255714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 896355714Skris for the encoded length. 896455714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 896555714Skris 896655714Skris *) Add initial documentation of the X509V3 functions. 896755714Skris [Steve Henson] 896855714Skris 896955714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 897055714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 897155714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 897255714Skris secure PKCS#8 private key format with a high iteration count. 897355714Skris [Steve Henson] 897455714Skris 897555714Skris *) Fix determination of Perl interpreter: A perl or perl5 897655714Skris _directory_ in $PATH was also accepted as the interpreter. 897755714Skris [Ralf S. Engelschall] 897855714Skris 897955714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 898055714Skris wrong with it but it was very old and did things like calling 898155714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 898255714Skris unusual formatting. 898355714Skris [Steve Henson] 898455714Skris 898555714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 898655714Skris to use the new extension code. 898755714Skris [Steve Henson] 898855714Skris 898955714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 899055714Skris with macros. This should make it easier to change their form, add extra 899155714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 899255714Skris constant. 899355714Skris [Steve Henson] 899455714Skris 899555714Skris *) Add to configuration table a new entry that can specify an alternative 899655714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 899755714Skris according to Mark Crispin <MRC@Panda.COM>. 899855714Skris [Bodo Moeller] 899955714Skris 900055714Skris#if 0 900155714Skris *) DES CBC did not update the IV. Weird. 900255714Skris [Ben Laurie] 900355714Skris#else 900455714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 900555714Skris Changing the behaviour of the former might break existing programs -- 900655714Skris where IV updating is needed, des_ncbc_encrypt can be used. 900755714Skris#endif 900855714Skris 900955714Skris *) When bntest is run from "make test" it drives bc to check its 901055714Skris calculations, as well as internally checking them. If an internal check 901155714Skris fails, it needs to cause bc to give a non-zero result or make test carries 901255714Skris on without noticing the failure. Fixed. 901355714Skris [Ben Laurie] 901455714Skris 901555714Skris *) DES library cleanups. 901655714Skris [Ulf M�ller] 901755714Skris 901855714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 901955714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 902055714Skris ciphers. NOTE: although the key derivation function has been verified 902155714Skris against some published test vectors it has not been extensively tested 902255714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 902355714Skris of v2.0. 902455714Skris [Steve Henson] 902555714Skris 902655714Skris *) Instead of "mkdir -p", which is not fully portable, use new 902755714Skris Perl script "util/mkdir-p.pl". 902855714Skris [Bodo Moeller] 902955714Skris 903055714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 903155714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 903255714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 903355714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 903455714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 903555714Skris underlying key generation function so it must do its own ASN1 parsing. 903655714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 903755714Skris 'parameter' argument instead of literal salt and iteration count values 903855714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 903955714Skris [Steve Henson] 904055714Skris 904155714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 904255714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 904355714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 904455714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 904555714Skris value was just used as a "magic string" and not used directly its 904655714Skris value doesn't matter. 904755714Skris [Steve Henson] 904855714Skris 904955714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 905055714Skris support mutable. 905155714Skris [Ben Laurie] 905255714Skris 905355714Skris *) "linux-sparc64" configuration (ultrapenguin). 905455714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 905555714Skris "linux-sparc" configuration. 905655714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 905755714Skris 905855714Skris *) config now generates no-xxx options for missing ciphers. 905955714Skris [Ulf M�ller] 906055714Skris 906155714Skris *) Support the EBCDIC character set (work in progress). 906255714Skris File ebcdic.c not yet included because it has a different license. 906355714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 906455714Skris 906555714Skris *) Support BS2000/OSD-POSIX. 906655714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 906755714Skris 906855714Skris *) Make callbacks for key generation use void * instead of char *. 906955714Skris [Ben Laurie] 907055714Skris 907155714Skris *) Make S/MIME samples compile (not yet tested). 907255714Skris [Ben Laurie] 907355714Skris 907455714Skris *) Additional typesafe stacks. 907555714Skris [Ben Laurie] 907655714Skris 907755714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 907855714Skris [Bodo Moeller] 907955714Skris 908055714Skris 908155714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 908255714Skris 908355714Skris *) New configuration variant "sco5-gcc". 908455714Skris 908555714Skris *) Updated some demos. 908655714Skris [Sean O Riordain, Wade Scholine] 908755714Skris 908855714Skris *) Add missing BIO_free at exit of pkcs12 application. 908955714Skris [Wu Zhigang] 909055714Skris 909155714Skris *) Fix memory leak in conf.c. 909255714Skris [Steve Henson] 909355714Skris 909455714Skris *) Updates for Win32 to assembler version of MD5. 909555714Skris [Steve Henson] 909655714Skris 909755714Skris *) Set #! path to perl in apps/der_chop to where we found it 909855714Skris instead of using a fixed path. 909955714Skris [Bodo Moeller] 910055714Skris 910155714Skris *) SHA library changes for irix64-mips4-cc. 910255714Skris [Andy Polyakov] 910355714Skris 910455714Skris *) Improvements for VMS support. 910555714Skris [Richard Levitte] 910655714Skris 910755714Skris 910855714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 910955714Skris 911055714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 911155714Skris This also avoids the problems with SC4.2 and unpatched SC5. 911255714Skris [Andy Polyakov <appro@fy.chalmers.se>] 911355714Skris 911455714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 911555714Skris These are required because of the typesafe stack would otherwise break 911655714Skris existing code. If old code used a structure member which used to be STACK 911755714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 911855714Skris sk_num or sk_value it would produce an error because the num, data members 911955714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 912055714Skris replaces the old method of assigning a value to sk_value 912155714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 912255714Skris that does this will no longer work (and should use sk_set instead) but 912355714Skris this could be regarded as a "questionable" behaviour anyway. 912455714Skris [Steve Henson] 912555714Skris 912655714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 912755714Skris correctly handle encrypted S/MIME data. 912855714Skris [Steve Henson] 912955714Skris 913055714Skris *) Change type of various DES function arguments from des_cblock 913155714Skris (which means, in function argument declarations, pointer to char) 913255714Skris to des_cblock * (meaning pointer to array with 8 char elements), 913355714Skris which allows the compiler to do more typechecking; it was like 913455714Skris that back in SSLeay, but with lots of ugly casts. 913555714Skris 913655714Skris Introduce new type const_des_cblock. 913755714Skris [Bodo Moeller] 913855714Skris 913955714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 914055714Skris problems: find RecipientInfo structure that matches recipient certificate 914155714Skris and initialise the ASN1 structures properly based on passed cipher. 914255714Skris [Steve Henson] 914355714Skris 914455714Skris *) Belatedly make the BN tests actually check the results. 914555714Skris [Ben Laurie] 914655714Skris 914755714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 914855714Skris to and from BNs: it was completely broken. New compilation option 914955714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 915055714Skris key elements as negative integers. 915155714Skris [Steve Henson] 915255714Skris 915355714Skris *) Reorganize and speed up MD5. 915455714Skris [Andy Polyakov <appro@fy.chalmers.se>] 915555714Skris 915655714Skris *) VMS support. 915755714Skris [Richard Levitte <richard@levitte.org>] 915855714Skris 915955714Skris *) New option -out to asn1parse to allow the parsed structure to be 916055714Skris output to a file. This is most useful when combined with the -strparse 916155714Skris option to examine the output of things like OCTET STRINGS. 916255714Skris [Steve Henson] 916355714Skris 916455714Skris *) Make SSL library a little more fool-proof by not requiring any longer 916555714Skris that SSL_set_{accept,connect}_state be called before 916655714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 916755714Skris in many applications because usually everything *appeared* to work as 916855714Skris intended anyway -- now it really works as intended). 916955714Skris [Bodo Moeller] 917055714Skris 917155714Skris *) Move openssl.cnf out of lib/. 917255714Skris [Ulf M�ller] 917355714Skris 917455714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 917555714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 917655714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 917755714Skris [Ralf S. Engelschall] 917855714Skris 917955714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 918055714Skris handle PKCS#7 enveloped data properly. 918155714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 918255714Skris 918355714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 918455714Skris copying pointers. The cert_st handling is changed by this in 918555714Skris various ways (and thus what used to be known as ctx->default_cert 918655714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 918755714Skris any longer when s->cert does not give us what we need). 918855714Skris ssl_cert_instantiate becomes obsolete by this change. 918955714Skris As soon as we've got the new code right (possibly it already is?), 919055714Skris we have solved a couple of bugs of the earlier code where s->cert 919155714Skris was used as if it could not have been shared with other SSL structures. 919255714Skris 919355714Skris Note that using the SSL API in certain dirty ways now will result 919455714Skris in different behaviour than observed with earlier library versions: 919555714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 919655714Skris does not influence s as it used to. 919755714Skris 919855714Skris In order to clean up things more thoroughly, inside SSL_SESSION 919955714Skris we don't use CERT any longer, but a new structure SESS_CERT 920055714Skris that holds per-session data (if available); currently, this is 920155714Skris the peer's certificate chain and, for clients, the server's certificate 920255714Skris and temporary key. CERT holds only those values that can have 920355714Skris meaningful defaults in an SSL_CTX. 920455714Skris [Bodo Moeller] 920555714Skris 920655714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 920755714Skris from the internal representation. Various PKCS#7 fixes: remove some 920855714Skris evil casts and set the enc_dig_alg field properly based on the signing 920955714Skris key type. 921055714Skris [Steve Henson] 921155714Skris 921255714Skris *) Allow PKCS#12 password to be set from the command line or the 921355714Skris environment. Let 'ca' get its config file name from the environment 921455714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 921555714Skris and 'x509'). 921655714Skris [Steve Henson] 921755714Skris 921855714Skris *) Allow certificate policies extension to use an IA5STRING for the 921955714Skris organization field. This is contrary to the PKIX definition but 922055714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 922155714Skris extension option. 922255714Skris [Steve Henson] 922355714Skris 922455714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 922555714Skris without disallowing inline assembler and the like for non-pedantic builds. 922655714Skris [Ben Laurie] 922755714Skris 922855714Skris *) Support Borland C++ builder. 922955714Skris [Janez Jere <jj@void.si>, modified by Ulf M�ller] 923055714Skris 923155714Skris *) Support Mingw32. 923255714Skris [Ulf M�ller] 923355714Skris 923455714Skris *) SHA-1 cleanups and performance enhancements. 923555714Skris [Andy Polyakov <appro@fy.chalmers.se>] 923655714Skris 923755714Skris *) Sparc v8plus assembler for the bignum library. 923855714Skris [Andy Polyakov <appro@fy.chalmers.se>] 923955714Skris 924055714Skris *) Accept any -xxx and +xxx compiler options in Configure. 924155714Skris [Ulf M�ller] 924255714Skris 924355714Skris *) Update HPUX configuration. 924455714Skris [Anonymous] 924555714Skris 924655714Skris *) Add missing sk_<type>_unshift() function to safestack.h 924755714Skris [Ralf S. Engelschall] 924855714Skris 924955714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 925055714Skris "extra_cert"s in addition to the certificate. (This makes sense 925155714Skris only for "PEM" format files, as chains as a whole are not 925255714Skris DER-encoded.) 925355714Skris [Bodo Moeller] 925455714Skris 925555714Skris *) Support verify_depth from the SSL API. 925655714Skris x509_vfy.c had what can be considered an off-by-one-error: 925755714Skris Its depth (which was not part of the external interface) 925855714Skris was actually counting the number of certificates in a chain; 925955714Skris now it really counts the depth. 926055714Skris [Bodo Moeller] 926155714Skris 926255714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 926355714Skris instead of X509err, which often resulted in confusing error 926455714Skris messages since the error codes are not globally unique 926555714Skris (e.g. an alleged error in ssl3_accept when a certificate 926655714Skris didn't match the private key). 926755714Skris 926855714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 926955714Skris value (so that you don't need SSL_set_session_id_context for each 927055714Skris connection using the SSL_CTX). 927155714Skris [Bodo Moeller] 927255714Skris 927355714Skris *) OAEP decoding bug fix. 927455714Skris [Ulf M�ller] 927555714Skris 927655714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 927755714Skris David Harris. 927855714Skris [Bodo Moeller] 927955714Skris 928055714Skris *) New Configure options "threads" and "no-threads". For systems 928155714Skris where the proper compiler options are known (currently Solaris 928255714Skris and Linux), "threads" is the default. 928355714Skris [Bodo Moeller] 928455714Skris 928555714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 928655714Skris [Bodo Moeller] 928755714Skris 928855714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 928955714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 929055714Skris such as /usr/local/bin. 929155714Skris [Bodo Moeller] 929255714Skris 929355714Skris *) "make linux-shared" to build shared libraries. 929455714Skris [Niels Poppe <niels@netbox.org>] 929555714Skris 929655714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 929755714Skris [Ulf M�ller] 929855714Skris 929955714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 930055714Skris extension adding in x509 utility. 930155714Skris [Steve Henson] 930255714Skris 930355714Skris *) Remove NOPROTO sections and error code comments. 930455714Skris [Ulf M�ller] 930555714Skris 930655714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 930755714Skris prototypes. 930855714Skris [Steve Henson] 930955714Skris 931055714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 931155714Skris [Ulf M�ller] 931255714Skris 931355714Skris *) Complete rewrite of the error code script(s). It is all now handled 931455714Skris by one script at the top level which handles error code gathering, 931555714Skris header rewriting and C source file generation. It should be much better 931655714Skris than the old method: it now uses a modified version of Ulf's parser to 931755714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 931855714Skris aren't needed for error creation any more) and do a better job of 931955714Skris translating function codes into names. The old 'ASN1 error code imbedded 932055714Skris in a comment' is no longer necessary and it doesn't use .err files which 932155714Skris have now been deleted. Also the error code call doesn't have to appear all 932255714Skris on one line (which resulted in some large lines...). 932355714Skris [Steve Henson] 932455714Skris 932555714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 932655714Skris [Bodo Moeller] 932755714Skris 932855714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 932955714Skris 0 (which usually indicates a closed connection), but continue reading. 933055714Skris [Bodo Moeller] 933155714Skris 933255714Skris *) Fix some race conditions. 933355714Skris [Bodo Moeller] 933455714Skris 933555714Skris *) Add support for CRL distribution points extension. Add Certificate 933655714Skris Policies and CRL distribution points documentation. 933755714Skris [Steve Henson] 933855714Skris 933955714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 934055714Skris [Ulf M�ller] 934155714Skris 934255714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 934355714Skris 8 of keying material. Merlin has also confirmed interop with this fix 934455714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 934555714Skris [Merlin Hughes <merlin@baltimore.ie>] 934655714Skris 934755714Skris *) Fix lots of warnings. 934855714Skris [Richard Levitte <levitte@stacken.kth.se>] 934955714Skris 935055714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 935155714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 935255714Skris [Richard Levitte <levitte@stacken.kth.se>] 935355714Skris 935455714Skris *) Fix problems with sizeof(long) == 8. 935555714Skris [Andy Polyakov <appro@fy.chalmers.se>] 935655714Skris 935755714Skris *) Change functions to ANSI C. 935855714Skris [Ulf M�ller] 935955714Skris 936055714Skris *) Fix typos in error codes. 936155714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M�ller] 936255714Skris 936355714Skris *) Remove defunct assembler files from Configure. 936455714Skris [Ulf M�ller] 936555714Skris 936655714Skris *) SPARC v8 assembler BIGNUM implementation. 936755714Skris [Andy Polyakov <appro@fy.chalmers.se>] 936855714Skris 936955714Skris *) Support for Certificate Policies extension: both print and set. 937055714Skris Various additions to support the r2i method this uses. 937155714Skris [Steve Henson] 937255714Skris 937355714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 937455714Skris return a const string when you are expecting an allocated buffer. 937555714Skris [Ben Laurie] 937655714Skris 937755714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 937855714Skris types DirectoryString and DisplayText. 937955714Skris [Steve Henson] 938055714Skris 938155714Skris *) Add code to allow r2i extensions to access the configuration database, 938255714Skris add an LHASH database driver and add several ctx helper functions. 938355714Skris [Steve Henson] 938455714Skris 938555714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 938655714Skris fail when they extended the size of a BIGNUM. 938755714Skris [Steve Henson] 938855714Skris 938955714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 939055714Skris support typesafe stack. 939155714Skris [Steve Henson] 939255714Skris 939355714Skris *) Fix typo in SSL_[gs]et_options(). 939455714Skris [Nils Frostberg <nils@medcom.se>] 939555714Skris 939655714Skris *) Delete various functions and files that belonged to the (now obsolete) 939755714Skris old X509V3 handling code. 939855714Skris [Steve Henson] 939955714Skris 940055714Skris *) New Configure option "rsaref". 940155714Skris [Ulf M�ller] 940255714Skris 940355714Skris *) Don't auto-generate pem.h. 940455714Skris [Bodo Moeller] 940555714Skris 940655714Skris *) Introduce type-safe ASN.1 SETs. 940755714Skris [Ben Laurie] 940855714Skris 940955714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 941055714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 941155714Skris 941255714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 941355714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 941455714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 941555714Skris few STACKed things have been converted already. Feel free to convert more. 941655714Skris In the fullness of time, I'll do away with the STACK type altogether. 941755714Skris [Ben Laurie] 941855714Skris 941955714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 942055714Skris specified in <certfile> by updating the entry in the index.txt file. 942155714Skris This way one no longer has to edit the index.txt file manually for 942255714Skris revoking a certificate. The -revoke option does the gory details now. 942355714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 942455714Skris 942555714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 942655714Skris `-text' option at all and this way the `-noout -text' combination was 942755714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 942855714Skris [Ralf S. Engelschall] 942955714Skris 943055714Skris *) Make sure a corresponding plain text error message exists for the 943155714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 943255714Skris verify callback function determined that a certificate was revoked. 943355714Skris [Ralf S. Engelschall] 943455714Skris 943555714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 943655714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 943755714Skris all available cipers including rc5, which was forgotten until now. 943855714Skris In order to let the testing shell script know which algorithms 943955714Skris are available, a new (up to now undocumented) command 944055714Skris "openssl list-cipher-commands" is used. 944155714Skris [Bodo Moeller] 944255714Skris 944355714Skris *) Bugfix: s_client occasionally would sleep in select() when 944455714Skris it should have checked SSL_pending() first. 944555714Skris [Bodo Moeller] 944655714Skris 944755714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 944855714Skris the raw DSA values prior to ASN.1 encoding. 944955714Skris [Ulf M�ller] 945055714Skris 945155714Skris *) Tweaks to Configure 945255714Skris [Niels Poppe <niels@netbox.org>] 945355714Skris 945455714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 945555714Skris yet... 945655714Skris [Steve Henson] 945755714Skris 945855714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 945955714Skris [Ulf M�ller] 946055714Skris 946155714Skris *) New config option to avoid instructions that are illegal on the 80386. 946255714Skris The default code is faster, but requires at least a 486. 946355714Skris [Ulf M�ller] 946455714Skris 946555714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 946655714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 946755714Skris same as SSL2_VERSION anyway. 946855714Skris [Bodo Moeller] 946955714Skris 947055714Skris *) New "-showcerts" option for s_client. 947155714Skris [Bodo Moeller] 947255714Skris 947355714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 947455714Skris application. Various cleanups and fixes. 947555714Skris [Steve Henson] 947655714Skris 947755714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 947855714Skris modify error routines to work internally. Add error codes and PBE init 947955714Skris to library startup routines. 948055714Skris [Steve Henson] 948155714Skris 948255714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 948355714Skris packing functions to asn1 and evp. Changed function names and error 948455714Skris codes along the way. 948555714Skris [Steve Henson] 948655714Skris 948755714Skris *) PKCS12 integration: and so it begins... First of several patches to 948855714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 948955714Skris objects to objects.h 949055714Skris [Steve Henson] 949155714Skris 949255714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 949355714Skris and display support for Thawte strong extranet extension. 949455714Skris [Steve Henson] 949555714Skris 949655714Skris *) Add LinuxPPC support. 949755714Skris [Jeff Dubrule <igor@pobox.org>] 949855714Skris 949955714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 950055714Skris bn_div_words in alpha.s. 950155714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 950255714Skris 950355714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 950455714Skris OAEP isn't supported when OpenSSL is built with RSAref. 950555714Skris [Ulf Moeller <ulf@fitug.de>] 950655714Skris 950755714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 950855714Skris so they no longer are missing under -DNOPROTO. 950955714Skris [Soren S. Jorvang <soren@t.dk>] 951055714Skris 951155714Skris 951255714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 951355714Skris 951455714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 951555714Skris doesn't work when the session is reused. Coming soon! 951655714Skris [Ben Laurie] 951755714Skris 951855714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 951955714Skris context thus bypassing client cert protection! All software that uses 952055714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 952155714Skris allow session reuse! A fuller solution is in the works. 952255714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 952355714Skris 952455714Skris *) Some more source tree cleanups (removed obsolete files 952555714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 952655714Skris permission on "config" script to be executable) and a fix for the INSTALL 952755714Skris document. 952855714Skris [Ulf Moeller <ulf@fitug.de>] 952955714Skris 953055714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 953155714Skris Malloc, Free. 953255714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 953355714Skris 953455714Skris *) Make rsa_oaep_test return non-zero on error. 953555714Skris [Ulf Moeller <ulf@fitug.de>] 953655714Skris 953755714Skris *) Add support for native Solaris shared libraries. Configure 953855714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 953955714Skris if someone would make that last step automatic. 954055714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 954155714Skris 954255714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 954355714Skris [Ben Laurie] 954455714Skris 954555714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 954655714Skris except NULL ciphers". This means the default cipher list will no longer 954755714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 954855714Skris the string "DEFAULT:eNULL". 954955714Skris [Steve Henson] 955055714Skris 955155714Skris *) Fix to RSA private encryption routines: if p < q then it would 955255714Skris occasionally produce an invalid result. This will only happen with 955355714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 955455714Skris [Steve Henson] 955555714Skris 955655714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 955755714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 955855714Skris because this way one can also use an interpreter named `perl5' (which is 955955714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 956055714Skris installed as `perl'). 956155714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 956255714Skris 956355714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 956455714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 956555714Skris 956655714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 956755714Skris advapi32.lib to Win32 build and change the pem test comparision 956855714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 956955714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 957055714Skris and crypto/des/ede_cbcm_enc.c. 957155714Skris [Steve Henson] 957255714Skris 957355714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 957455714Skris [Ben Laurie] 957555714Skris 957655714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 957755714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 957855714Skris is horrible: I feel ill.... 957955714Skris [Steve Henson] 958055714Skris 958155714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 958255714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 958355714Skris sections: 10 functions were absent from non ANSI section and not exported 958455714Skris from Windows DLLs. Fixed up libeay.num for new functions. 958555714Skris [Steve Henson] 958655714Skris 958755714Skris *) Make `openssl version' output lines consistent. 958855714Skris [Ralf S. Engelschall] 958955714Skris 959055714Skris *) Fix Win32 symbol export lists for BIO functions: Added 959155714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 959255714Skris to ms/libeay{16,32}.def. 959355714Skris [Ralf S. Engelschall] 959455714Skris 959555714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 959655714Skris fine under Unix and passes some trivial tests I've now added. But the 959755714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 959855714Skris added to make sure no one expects that this stuff really works in the 959955714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 960055714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 960155714Skris openssl_bio.xs. 960255714Skris [Ralf S. Engelschall] 960355714Skris 960455714Skris *) Fix the generation of two part addresses in perl. 960555714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 960655714Skris 960755714Skris *) Add config entry for Linux on MIPS. 960855714Skris [John Tobey <jtobey@channel1.com>] 960955714Skris 961055714Skris *) Make links whenever Configure is run, unless we are on Windoze. 961155714Skris [Ben Laurie] 961255714Skris 961355714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 961455714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 961555714Skris in CRLs. 961655714Skris [Steve Henson] 961755714Skris 961855714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 961955714Skris other platforms details on the command line without having to patch the 962055714Skris Configure script everytime: One now can use ``perl Configure 962155714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 962255714Skris to them (seperated by colons). This is treated as there would be a static 962355714Skris pre-configured entry in Configure's %table under key <id> with value 962455714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 962555714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 962655714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 962755714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 962855714Skris [Ralf S. Engelschall] 962955714Skris 963055714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 963155714Skris [Ben Laurie] 963255714Skris 963355714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 963455714Skris on the `perl Configure ...' command line. This way one can compile 963555714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 963655714Skris for linking it into DSOs. 963755714Skris [Ralf S. Engelschall] 963855714Skris 963955714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 964055714Skris Fixed. 964155714Skris [Ben Laurie] 964255714Skris 964355714Skris *) Cleaned up the LICENSE document: The official contact for any license 964455714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 964555714Skris And add a paragraph about the dual-license situation to make sure people 964655714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 964755714Skris to the OpenSSL toolkit. 964855714Skris [Ralf S. Engelschall] 964955714Skris 965055714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 965155714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 965255714Skris Additonally cleaned up the `make links' target: Remove unnecessary 965355714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 965455714Skris to speed processing and no longer clutter the display with confusing 965555714Skris stuff. Instead only the actually done links are displayed. 965655714Skris [Ralf S. Engelschall] 965755714Skris 965855714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 965955714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 966055714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 966155714Skris encryption. 966255714Skris [Ben Laurie] 966355714Skris 966455714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 966555714Skris signed attributes when verifying signatures (this would break them), 966655714Skris the detached data encoding was wrong and public keys obtained using 966755714Skris X509_get_pubkey() weren't freed. 966855714Skris [Steve Henson] 966955714Skris 967055714Skris *) Add text documentation for the BUFFER functions. Also added a work around 967155714Skris to a Win95 console bug. This was triggered by the password read stuff: the 967255714Skris last character typed gets carried over to the next fread(). If you were 967355714Skris generating a new cert request using 'req' for example then the last 967455714Skris character of the passphrase would be CR which would then enter the first 967555714Skris field as blank. 967655714Skris [Steve Henson] 967755714Skris 967855714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 967955714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 968055714Skris button and can be used by applications based on OpenSSL to show the 968155714Skris relationship to the OpenSSL project. 968255714Skris [Ralf S. Engelschall] 968355714Skris 968455714Skris *) Remove confusing variables in function signatures in files 968555714Skris ssl/ssl_lib.c and ssl/ssl.h. 968655714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 968755714Skris 968855714Skris *) Don't install bss_file.c under PREFIX/include/ 968955714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 969055714Skris 969155714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 969255714Skris functions that return function pointers and has support for NT specific 969355714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 969455714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 969555714Skris unsigned to signed types: this was killing the Win32 compile. 969655714Skris [Steve Henson] 969755714Skris 969855714Skris *) Add new certificate file to stack functions, 969955714Skris SSL_add_dir_cert_subjects_to_stack() and 970055714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 970155714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 970255714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 970355714Skris This means that Apache-SSL and similar packages don't have to mess around 970455714Skris to add as many CAs as they want to the preferred list. 970555714Skris [Ben Laurie] 970655714Skris 970755714Skris *) Experiment with doxygen documentation. Currently only partially applied to 970855714Skris ssl/ssl_lib.c. 970955714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 971055714Skris openssl.doxy as the configuration file. 971155714Skris [Ben Laurie] 971255714Skris 971355714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 971455714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 971555714Skris 971655714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 971755714Skris compiled in by default: it has problems with large keys. 971855714Skris [Steve Henson] 971955714Skris 972055714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 972155714Skris DH private keys and/or callback functions which directly correspond to 972255714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 972355714Skris is needed for applications which have to configure certificates on a 972455714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 972555714Skris (e.g. s_server). 972655714Skris For the RSA certificate situation is makes no difference, but 972755714Skris for the DSA certificate situation this fixes the "no shared cipher" 972855714Skris problem where the OpenSSL cipher selection procedure failed because the 972955714Skris temporary keys were not overtaken from the context and the API provided 973055714Skris no way to reconfigure them. 973155714Skris The new functions now let applications reconfigure the stuff and they 973255714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 973355714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 973455714Skris non-public-API function ssl_cert_instantiate() is used as a helper 973555714Skris function and also to reduce code redundancy inside ssl_rsa.c. 973655714Skris [Ralf S. Engelschall] 973755714Skris 973855714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 973955714Skris area because they are useful for the DSA situation and should be 974055714Skris recognized by the users. 974155714Skris [Ralf S. Engelschall] 974255714Skris 974355714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 974455714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 974555714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 974655714Skris already masked variable. 974755714Skris [Richard Levitte <levitte@stacken.kth.se>] 974855714Skris 974955714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 975055714Skris [Richard Levitte <levitte@stacken.kth.se>] 975155714Skris 975255714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 975355714Skris from `int' to `unsigned int' because it's a length and initialized by 975455714Skris EVP_DigestFinal() which expects an `unsigned int *'. 975555714Skris [Richard Levitte <levitte@stacken.kth.se>] 975655714Skris 975755714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 975855714Skris script. Instead use the usual Shell->Perl transition trick. 975955714Skris [Ralf S. Engelschall] 976055714Skris 976155714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 976255714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 976355714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 976455714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 976555714Skris currently the public key is printed (a decision which was already done by 976655714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 976755714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 976855714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 976955714Skris now, too. 977055714Skris [Ralf S. Engelschall] 977155714Skris 977255714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 977355714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 977455714Skris [Arne Ansper <arne@ats.cyber.ee>] 977555714Skris 977655714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 977755714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 977855714Skris config file. 977955714Skris [Steve Henson] 978055714Skris 978155714Skris *) Add cool BIO that does syslog (or event log on NT). 978255714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 978355714Skris 978455714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 978555714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 978655714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 978755714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 978855714Skris [Ben Laurie] 978955714Skris 979055714Skris *) Add preliminary config info for new extension code. 979155714Skris [Steve Henson] 979255714Skris 979355714Skris *) Make RSA_NO_PADDING really use no padding. 979455714Skris [Ulf Moeller <ulf@fitug.de>] 979555714Skris 979655714Skris *) Generate errors when private/public key check is done. 979755714Skris [Ben Laurie] 979855714Skris 979955714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 980055714Skris for some CRL extensions and new objects added. 980155714Skris [Steve Henson] 980255714Skris 980355714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 980455714Skris key usage extension and fuller support for authority key id. 980555714Skris [Steve Henson] 980655714Skris 980755714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 980855714Skris padding method for RSA, which is recommended for new applications in PKCS 980955714Skris #1 v2.0 (RFC 2437, October 1998). 981055714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 981155714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 981255714Skris against Bleichbacher's attack on RSA. 981355714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 981455714Skris Ben Laurie] 981555714Skris 981655714Skris *) Updates to the new SSL compression code 981755714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 981855714Skris 981955714Skris *) Fix so that the version number in the master secret, when passed 982055714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 982155714Skris (because the server will not accept higher), that the version number 982255714Skris is 0x03,0x01, not 0x03,0x00 982355714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 982455714Skris 982555714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 982655714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 982755714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 982855714Skris [Steve Henson] 982955714Skris 983055714Skris *) Support for RAW extensions where an arbitrary extension can be 983155714Skris created by including its DER encoding. See apps/openssl.cnf for 983255714Skris an example. 983355714Skris [Steve Henson] 983455714Skris 983555714Skris *) Make sure latest Perl versions don't interpret some generated C array 983655714Skris code as Perl array code in the crypto/err/err_genc.pl script. 983755714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 983855714Skris 983955714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 984055714Skris not many people have the assembler. Various Win32 compilation fixes and 984155714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 984255714Skris build instructions. 984355714Skris [Steve Henson] 984455714Skris 984555714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 984655714Skris file under Win32 and also build pem.h from pem.org. New script 984755714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 984855714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 984955714Skris [Steve Henson] 985055714Skris 985155714Skris *) Major rework of DES function declarations, in the pursuit of correctness 985255714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 985355714Skris too. You may find this causes warnings in your code. Zapping your evil 985455714Skris casts will probably fix them. Mostly. 985555714Skris [Ben Laurie] 985655714Skris 985755714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 985855714Skris obj_dat.pl. It considered a zero in an object definition to mean 985955714Skris "end of object": none of the objects in objects.h have any zeros 986055714Skris so it wasn't spotted. 986155714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 986255714Skris 986355714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 986455714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 986555714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 986655714Skris vectors if you have them. 986755714Skris [Ben Laurie] 986855714Skris 986955714Skris *) Correct calculation of key length for export ciphers (too much space was 987055714Skris allocated for null ciphers). This has not been tested! 987155714Skris [Ben Laurie] 987255714Skris 987355714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 987455714Skris message is now correct (it understands "crypto" and "ssl" on its 987555714Skris command line). There is also now an "update" option. This will update 987655714Skris the util/ssleay.num and util/libeay.num files with any new functions. 987755714Skris If you do a: 987855714Skris perl util/mkdef.pl crypto ssl update 987955714Skris it will update them. 988055714Skris [Steve Henson] 988155714Skris 988255714Skris *) Overhauled the Perl interface (perl/*): 988355714Skris - ported BN stuff to OpenSSL's different BN library 988455714Skris - made the perl/ source tree CVS-aware 988555714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 988655714Skris their history because I've copied them in the repository) 988755714Skris - removed obsolete files (the test scripts will be replaced 988855714Skris by better Test::Harness variants in the future) 988955714Skris [Ralf S. Engelschall] 989055714Skris 989155714Skris *) First cut for a very conservative source tree cleanup: 989255714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 989355714Skris where we collect the old documents and readme texts. 989455714Skris 2. remove the first part of files where I'm already sure that we no 989555714Skris longer need them because of three reasons: either they are just temporary 989655714Skris files which were left by Eric or they are preserved original files where 989755714Skris I've verified that the diff is also available in the CVS via "cvs diff 989855714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 989955714Skris the crypto/md/ stuff). 990055714Skris [Ralf S. Engelschall] 990155714Skris 990255714Skris *) More extension code. Incomplete support for subject and issuer alt 990355714Skris name, issuer and authority key id. Change the i2v function parameters 990455714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 990555714Skris what that's for :-) Fix to ASN1 macro which messed up 990655714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 990755714Skris [Steve Henson] 990855714Skris 990955714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 991055714Skris INTEGER code. 991155714Skris [Steve Henson] 991255714Skris 991355714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 991455714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 991555714Skris 991655714Skris *) Make sure `make rehash' target really finds the `openssl' program. 991755714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 991855714Skris 991955714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 992055714Skris like to hear about it if this slows down other processors. 992155714Skris [Ben Laurie] 992255714Skris 992355714Skris *) Add CygWin32 platform information to Configure script. 992455714Skris [Alan Batie <batie@aahz.jf.intel.com>] 992555714Skris 992655714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 992755714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 992855714Skris 992955714Skris *) New program nseq to manipulate netscape certificate sequences 993055714Skris [Steve Henson] 993155714Skris 993255714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 993355714Skris few typos. 993455714Skris [Steve Henson] 993555714Skris 993655714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 993755714Skris but the BN code had some problems that would cause failures when 993855714Skris doing certificate verification and some other functions. 993955714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 994055714Skris 994155714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 994255714Skris [Steve Henson] 994355714Skris 994455714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 994555714Skris [Steve Henson] 994655714Skris 994755714Skris *) Add several PKIX and private extended key usage OIDs. 994855714Skris [Steve Henson] 994955714Skris 995055714Skris *) Modify the 'ca' program to handle the new extension code. Modify 995155714Skris openssl.cnf for new extension format, add comments. 995255714Skris [Steve Henson] 995355714Skris 995455714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 995555714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 995655714Skris CA extensions. 995755714Skris [Steve Henson] 995855714Skris 995955714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 996055714Skris error code, add initial support to X509_print() and x509 application. 996155714Skris [Steve Henson] 996255714Skris 996355714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 996455714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 996555714Skris stuff is currently isolated and isn't even compiled yet. 996655714Skris [Steve Henson] 996755714Skris 996855714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 996955714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 997055714Skris Removed the versions check from X509 routines when loading extensions: 997155714Skris this allows certain broken certificates that don't set the version 997255714Skris properly to be processed. 997355714Skris [Steve Henson] 997455714Skris 997555714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 997655714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 997755714Skris can still be regenerated with "make depend". 997855714Skris [Ben Laurie] 997955714Skris 998055714Skris *) Spelling mistake in C version of CAST-128. 998155714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 998255714Skris 998355714Skris *) Changes to the error generation code. The perl script err-code.pl 998455714Skris now reads in the old error codes and retains the old numbers, only 998555714Skris adding new ones if necessary. It also only changes the .err files if new 998655714Skris codes are added. The makefiles have been modified to only insert errors 998755714Skris when needed (to avoid needlessly modifying header files). This is done 998855714Skris by only inserting errors if the .err file is newer than the auto generated 998955714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 999055714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 999155714Skris or delete all the .err files. 999255714Skris [Steve Henson] 999355714Skris 999455714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 999555714Skris been fixed, but is untested. The assembler versions are also fixed, but 999655714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 999755714Skris to regenerate it if needed. 999855714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 999955714Skris Hagino <itojun@kame.net>] 1000055714Skris 1000155714Skris *) File was opened incorrectly in randfile.c. 1000255714Skris [Ulf M�ller <ulf@fitug.de>] 1000355714Skris 1000455714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 1000555714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 1000655714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 1000755714Skris al: it's just almost always a UTCTime. Note this patch adds new error 1000855714Skris codes so do a "make errors" if there are problems. 1000955714Skris [Steve Henson] 1001055714Skris 1001155714Skris *) Correct Linux 1 recognition in config. 1001255714Skris [Ulf M�ller <ulf@fitug.de>] 1001355714Skris 1001455714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 1001555714Skris [Anonymous <nobody@replay.com>] 1001655714Skris 1001755714Skris *) Generate an error if given an empty string as a cert directory. Also 1001855714Skris generate an error if handed NULL (previously returned 0 to indicate an 1001955714Skris error, but didn't set one). 1002055714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 1002155714Skris 1002255714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 1002355714Skris [Ben Laurie] 1002455714Skris 1002555714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 1002655714Skris parameters. This was causing a warning which killed off the Win32 compile. 1002755714Skris [Steve Henson] 1002855714Skris 1002955714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 1003055714Skris [Neil Costigan <neil.costigan@celocom.com>] 1003155714Skris 1003255714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 1003355714Skris based on a text string, looking up short and long names and finally 1003455714Skris "dot" format. The "dot" format stuff didn't work. Added new function 1003555714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 1003655714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 1003755714Skris OID is not part of the table. 1003855714Skris [Steve Henson] 1003955714Skris 1004055714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 1004155714Skris X509_LOOKUP_by_alias(). 1004255714Skris [Ben Laurie] 1004355714Skris 1004455714Skris *) Sort openssl functions by name. 1004555714Skris [Ben Laurie] 1004655714Skris 1004755714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 1004855714Skris encryption from sample DSA keys (in case anyone is interested the password 1004955714Skris was "1234"). 1005055714Skris [Steve Henson] 1005155714Skris 1005255714Skris *) Make _all_ *_free functions accept a NULL pointer. 1005355714Skris [Frans Heymans <fheymans@isaserver.be>] 1005455714Skris 1005555714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 1005655714Skris NULL pointers. 1005755714Skris [Anonymous <nobody@replay.com>] 1005855714Skris 1005955714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 1006055714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 1006155714Skris 1006255714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 1006355714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 1006455714Skris 1006555714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 1006655714Skris [Anonymous <nobody@replay.com>] 1006755714Skris 1006855714Skris *) Add prototype for temp key callback functions 1006955714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 1007055714Skris [Ben Laurie] 1007155714Skris 1007255714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 1007355714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 1007455714Skris [Steve Henson] 1007555714Skris 1007655714Skris *) X509_name_add_entry() freed the wrong thing after an error. 1007755714Skris [Arne Ansper <arne@ats.cyber.ee>] 1007855714Skris 1007955714Skris *) rsa_eay.c would attempt to free a NULL context. 1008055714Skris [Arne Ansper <arne@ats.cyber.ee>] 1008155714Skris 1008255714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 1008355714Skris [Arne Ansper <arne@ats.cyber.ee>] 1008455714Skris 1008555714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 1008655714Skris [Arne Ansper <arne@ats.cyber.ee>] 1008755714Skris 1008855714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 1008955714Skris in X509_STORE_new(), but document the fact that this variable is still 1009055714Skris unused in the certificate verification process. 1009155714Skris [Ralf S. Engelschall] 1009255714Skris 1009355714Skris *) Fix the various library and apps files to free up pkeys obtained from 1009455714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 1009555714Skris [Steve Henson] 1009655714Skris 1009755714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 1009855714Skris demos/maurice/example2.c work, amongst others, probably. 1009955714Skris [Steve Henson and Ben Laurie] 1010055714Skris 1010155714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 1010255714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 1010355714Skris are no longer created. This way we have a single and consistent command 1010455714Skris line interface `openssl <command>', similar to `cvs <command>'. 1010555714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 1010655714Skris 1010755714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 1010855714Skris BIT STRING wrapper always have zero unused bits. 1010955714Skris [Steve Henson] 1011055714Skris 1011155714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 1011255714Skris [Steve Henson] 1011355714Skris 1011455714Skris *) Make the top-level INSTALL documentation easier to understand. 1011555714Skris [Paul Sutton] 1011655714Skris 1011755714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 1011855714Skris make (including if user presses ^C) [Paul Sutton] 1011955714Skris 1012055714Skris *) Make Montgomery context stuff explicit in RSA data structure. 1012155714Skris [Ben Laurie] 1012255714Skris 1012355714Skris *) Fix build order of pem and err to allow for generated pem.h. 1012455714Skris [Ben Laurie] 1012555714Skris 1012655714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 1012755714Skris [Ben Laurie] 1012855714Skris 1012955714Skris *) Enhanced the err-ins.pl script so it makes the error library number 1013055714Skris global and can add a library name. This is needed for external ASN1 and 1013155714Skris other error libraries. 1013255714Skris [Steve Henson] 1013355714Skris 1013455714Skris *) Fixed sk_insert which never worked properly. 1013555714Skris [Steve Henson] 1013655714Skris 1013755714Skris *) Fix ASN1 macros so they can handle indefinite length construted 1013855714Skris EXPLICIT tags. Some non standard certificates use these: they can now 1013955714Skris be read in. 1014055714Skris [Steve Henson] 1014155714Skris 1014255714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 1014355714Skris into a single doc/ssleay.txt bundle. This way the information is still 1014455714Skris preserved but no longer messes up this directory. Now it's new room for 1014555714Skris the new set of documenation files. 1014655714Skris [Ralf S. Engelschall] 1014755714Skris 1014855714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 1014955714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 1015055714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 1015155714Skris number of arguments. 1015255714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 1015355714Skris 1015455714Skris *) Fix test data to work with the above. 1015555714Skris [Ben Laurie] 1015655714Skris 1015755714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 1015855714Skris was already fixed by Eric for 0.9.1 it seems. 1015955714Skris [Ben Laurie - pointed out by Ulf M�ller <ulf@fitug.de>] 1016055714Skris 1016155714Skris *) Autodetect FreeBSD3. 1016255714Skris [Ben Laurie] 1016355714Skris 1016455714Skris *) Fix various bugs in Configure. This affects the following platforms: 1016555714Skris nextstep 1016655714Skris ncr-scde 1016755714Skris unixware-2.0 1016855714Skris unixware-2.0-pentium 1016955714Skris sco5-cc. 1017055714Skris [Ben Laurie] 1017155714Skris 1017255714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 1017355714Skris before they are needed. 1017455714Skris [Ben Laurie] 1017555714Skris 1017655714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 1017755714Skris [Ben Laurie] 1017855714Skris 1017955714Skris 1018055714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 1018155714Skris 1018255714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 1018355714Skris changed SSLeay to OpenSSL in version strings. 1018455714Skris [Ralf S. Engelschall] 1018555714Skris 1018655714Skris *) Some fixups to the top-level documents. 1018755714Skris [Paul Sutton] 1018855714Skris 1018955714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 1019055714Skris because the symlink to include/ was missing. 1019155714Skris [Ralf S. Engelschall] 1019255714Skris 1019355714Skris *) Incorporated the popular no-RSA/DSA-only patches 1019455714Skris which allow to compile a RSA-free SSLeay. 1019555714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 1019655714Skris 1019755714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 1019855714Skris when "ssleay" is still not found. 1019955714Skris [Ralf S. Engelschall] 1020055714Skris 1020155714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 1020255714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 1020355714Skris 1020455714Skris *) Updated the README file. 1020555714Skris [Ralf S. Engelschall] 1020655714Skris 1020755714Skris *) Added various .cvsignore files in the CVS repository subdirs 1020855714Skris to make a "cvs update" really silent. 1020955714Skris [Ralf S. Engelschall] 1021055714Skris 1021155714Skris *) Recompiled the error-definition header files and added 1021255714Skris missing symbols to the Win32 linker tables. 1021355714Skris [Ralf S. Engelschall] 1021455714Skris 1021555714Skris *) Cleaned up the top-level documents; 1021655714Skris o new files: CHANGES and LICENSE 1021755714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 1021855714Skris o merged COPYRIGHT into LICENSE 1021955714Skris o removed obsolete TODO file 1022055714Skris o renamed MICROSOFT to INSTALL.W32 1022155714Skris [Ralf S. Engelschall] 1022255714Skris 1022355714Skris *) Removed dummy files from the 0.9.1b source tree: 1022455714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 1022555714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 1022655714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 1022755714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 1022855714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 1022955714Skris [Ralf S. Engelschall] 1023055714Skris 1023155714Skris *) Added various platform portability fixes. 1023255714Skris [Mark J. Cox] 1023355714Skris 1023455714Skris *) The Genesis of the OpenSSL rpject: 1023555714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 1023655714Skris Young and Tim J. Hudson created while they were working for C2Net until 1023755714Skris summer 1998. 1023855714Skris [The OpenSSL Project] 1023955714Skris 1024055714Skris 1024155714Skris Changes between 0.9.0b and 0.9.1b [not released] 1024255714Skris 1024355714Skris *) Updated a few CA certificates under certs/ 1024455714Skris [Eric A. Young] 1024555714Skris 1024655714Skris *) Changed some BIGNUM api stuff. 1024755714Skris [Eric A. Young] 1024855714Skris 1024955714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 1025055714Skris DGUX x86, Linux Alpha, etc. 1025155714Skris [Eric A. Young] 1025255714Skris 1025355714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 1025455714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 1025555714Skris available). 1025655714Skris [Eric A. Young] 1025755714Skris 1025855714Skris *) Add -strparse option to asn1pars program which parses nested 1025955714Skris binary structures 1026055714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 1026155714Skris 1026255714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 1026355714Skris [Eric A. Young] 1026455714Skris 1026555714Skris *) DSA fix for "ca" program. 1026655714Skris [Eric A. Young] 1026755714Skris 1026855714Skris *) Added "-genkey" option to "dsaparam" program. 1026955714Skris [Eric A. Young] 1027055714Skris 1027155714Skris *) Added RIPE MD160 (rmd160) message digest. 1027255714Skris [Eric A. Young] 1027355714Skris 1027455714Skris *) Added -a (all) option to "ssleay version" command. 1027555714Skris [Eric A. Young] 1027655714Skris 1027755714Skris *) Added PLATFORM define which is the id given to Configure. 1027855714Skris [Eric A. Young] 1027955714Skris 1028055714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 1028155714Skris [Eric A. Young] 1028255714Skris 1028355714Skris *) Extended the ASN.1 parser routines. 1028455714Skris [Eric A. Young] 1028555714Skris 1028655714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 1028755714Skris [Eric A. Young] 1028855714Skris 1028955714Skris *) Added a BN_CTX to the BN library. 1029055714Skris [Eric A. Young] 1029155714Skris 1029255714Skris *) Fixed the weak key values in DES library 1029355714Skris [Eric A. Young] 1029455714Skris 1029555714Skris *) Changed API in EVP library for cipher aliases. 1029655714Skris [Eric A. Young] 1029755714Skris 1029855714Skris *) Added support for RC2/64bit cipher. 1029955714Skris [Eric A. Young] 1030055714Skris 1030155714Skris *) Converted the lhash library to the crypto/mem.c functions. 1030255714Skris [Eric A. Young] 1030355714Skris 1030455714Skris *) Added more recognized ASN.1 object ids. 1030555714Skris [Eric A. Young] 1030655714Skris 1030755714Skris *) Added more RSA padding checks for SSL/TLS. 1030855714Skris [Eric A. Young] 1030955714Skris 1031055714Skris *) Added BIO proxy/filter functionality. 1031155714Skris [Eric A. Young] 1031255714Skris 1031355714Skris *) Added extra_certs to SSL_CTX which can be used 1031455714Skris send extra CA certificates to the client in the CA cert chain sending 1031555714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 1031655714Skris [Eric A. Young] 1031755714Skris 1031855714Skris *) Now Fortezza is denied in the authentication phase because 1031955714Skris this is key exchange mechanism is not supported by SSLeay at all. 1032055714Skris [Eric A. Young] 1032155714Skris 1032255714Skris *) Additional PKCS1 checks. 1032355714Skris [Eric A. Young] 1032455714Skris 1032555714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 1032655714Skris [Eric A. Young] 1032755714Skris 1032855714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 1032955714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 1033055714Skris [Eric A. Young] 1033155714Skris 1033255714Skris *) Fixed a few memory leaks. 1033355714Skris [Eric A. Young] 1033455714Skris 1033555714Skris *) Fixed various code and comment typos. 1033655714Skris [Eric A. Young] 1033755714Skris 1033855714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 1033955714Skris bytes sent in the client random. 1034055714Skris [Edward Bishop <ebishop@spyglass.com>] 1034155714Skris 10342